diff --git a/AUTHORS b/AUTHORS
index 1ca83d05b9f1ac10f19d4e99ebcbdf445803cda3..4811a692e77f512e9a29c4e5fe7b9c281271101c 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -30,6 +30,7 @@ Dan Phung <phung4(at)llnl.gov>
Ashley Pitman <ashley(at)quadrics.com>
Andy Riebs <Andy.Riebs(at)hp.com>
Asier Roa <asier.roa(at)bsc.es>
+Miguel Ros <miguel.ros(at)bsc.es>
Federico Sacerdoti <Federico.Sacerdoti(at)deshaw.com>
Jeff Squyres <jsquyres(at)lam-mpi.org>
Keven Tew <tew1(at)llnl.gov>
diff --git a/BUILD.NOTES b/BUILD.NOTES
index 8e0a08d2dbe591e9719ca4c104e5fb821ad6fd74..273a2d816622b539e78ca7f2c137b447adebe538 100644
--- a/BUILD.NOTES
+++ b/BUILD.NOTES
@@ -10,6 +10,11 @@ Simple build/install on Linux:
make
make install
+To build the files in the contribs directory:
+ make contrib
+ make install-contrib
+ (The RPMs are built by default)
+
If you make changes to any auxdir/* or Makefile.am file, then run
_snowflake_ (where there are recent versions of autoconf, automake
and libtool installed):
@@ -41,16 +46,17 @@ Linux cluster (See BlueGene and AIX specific notes below for some differences).
%_with_sgijob 1 (ON SYSTEMS WITHOUT ELAN SWITCH)
I usually build with using the following syntax:
build -s https://eris.llnl.gov/svn/slurm/tags/slurm-1-2-0-0-pre3
- NOTE: For v1.0 and earlier add: --pre-exec='./autogen.sh'
-4. Move the RPMs to
+4. Remove the RPMs that we don't want:
+ rm -f slurm-perlapi*rpm slurm-torque*rpm
+5. Move the RPMs to
/usr/local/admin/rpms/llnl/RPMS-RHEL4/x86_64 (odevi, or gauss)
/usr/local/admin/rpms/llnl/RPMS-RHEL4/i386/ (adevi)
/usr/local/admin/rpms/llnl/RPMS-RHEL4/ia64/ (tdevi)
send an announcement email (with the latest entry from the NEWS
file) out to linux-admin@lists.llnl.gov.
-5. Copy tagged bzip file (e.g. slurm-0.6.0-0.pre3.bz2) to FTP server
+6. Copy tagged bzip file (e.g. slurm-0.6.0-0.pre3.bz2) to FTP server
for external SLURM users.
-6. Copy bzip file and rpms (including src.rpm) to sourceforge.net:
+7. Copy bzip file and rpms (including src.rpm) to sourceforge.net:
ncftp upload.sf.net
cd upload
put filename
@@ -64,8 +70,7 @@ BlueGene build notes:
%_with_bluegene 1
%with_cflags CFLAGS=-m64
Build on Service Node with using the following syntax
- build -s https://eris.llnl.gov/svn/slurm/tags/slurm-1-2-0-0-pre3
- 4. Copy RPMs to /usr/admin/sles/llnl/RPMS-SLES9
+ rpmbuild -ta slurm-...bz2
To build and run on AIX:
0. svn co https://eris.llnl.gov/svn/slurm/trunk slurm
@@ -114,8 +119,10 @@ To build and run on AIX:
There will be a log file create named /tmp/mplog.<jobid>.<taskid>
7. If you update proctrack, be sure to run "slibclean" to clear cached
version.
-8. Install the rpms int /usr/admin/inst.images/slurm/aix5.3 on an OCF AIX
- machine (pdev is a good choice).
+8. Remove the RPMs that we don't want:
+ rm -f slurm-perlapi*rpm slurm-torque*rpm
+ and install the other RPMs into /usr/admin/inst.images/slurm/aix5.3 on an
+ OCF AIX machine (pdev is a good choice).
AIX/Federation switch window problems
To clean switch windows: ntblclean =w 8 -a sni0
diff --git a/META b/META
index 51846978ecdf72fc28169e677f2aa335012f2eb0..8723309f560d0e1b1e016f37a22f9105a64a57be 100644
--- a/META
+++ b/META
@@ -3,8 +3,9 @@
Api_revision: 0
Major: 1
Meta: 1
- Micro: 19
+ Micro: 20
Minor: 2
Name: slurm
Release: 1
- Version: 1.2.19
+ Release_tags: dist
+ Version: 1.2.20
diff --git a/NEWS b/NEWS
index 113cb3f80bb6442d41d1300bef5903e249dec75b..964b4f1c3e0d968e3e71e1d9bc4927ab6f08bbf1 100644
--- a/NEWS
+++ b/NEWS
@@ -1,6 +1,37 @@
This file describes changes in recent versions of SLURM. It primarily
documents those changes that are of interest to users and admins.
+* Changes in SLURM 1.2.20
+=========================
+ -- In switch/federation, fix small memory leak effecting slurmd.
+ -- Add PMI_FANOUT_OFF_HOST environment variable to control how message
+ forwarding is done for PMI (MPICH2). See "man srun" for details.
+ -- From sbatch set SLURM_NTASKS_PER_NODE when --ntasks-per-node option is
+ specified.
+ -- BLUEGENE: Documented the prefix should always be lower case and the 3
+ digit suffix should be uppercase if any letters are used as digits.
+ -- In sched/wiki and sched/wiki2, add support for --cpus-per-task option.
+ From Miguel Ros, BSC.
+ -- In sched/wiki2, prevent invalid memory pointer (and likely seg fault)
+ for job associated with a partition that has since been deleted.
+ -- In sched/wiki2 plus select/cons_res, prevent invalid memory pointer
+ (and likely seg fault) when a job is requeued.
+ -- In sched/wiki, add support for job suspend, resume, and modify.
+ -- In sched/wiki, add suppport for processor allocation (not just node allocation)
+ with layout control.
+ -- Prevent re-sending job termination RPC to a node that has already completed
+ the job. Only send it to specific nodes which have not reported completion.
+ -- Support larger environment variables 64K instead of BUFSIZ (8k on some
+ systems).
+ -- If a job is being requeued, job step create requests will print a
+ warning and repeatedly retry rather than aborting.
+ -- Add optional mode value to srun and sbatch --get-user-env option.
+ -- Print error message and retry job submit commands when MaxJobCount
+ is reached. From Don Albert, Bull.
+ -- Treat invalid begin time specification as a fatal error in sbatch and
+ srun. From Don Albert, Bull.
+ -- Validate begin time specification to avoid hours >24, minutes >59, etc.
+
* Changes in SLURM 1.2.19
=========================
*** NOTE IMPORTANT CHANGE IN RPM BUILD BELOW ****
@@ -2707,4 +2738,4 @@ documents those changes that are of interest to users and admins.
-- Change directory to /tmp in slurmd if daemonizing.
-- Logfiles are reopened on reconfigure.
-$Id: NEWS 12599 2007-11-01 21:19:34Z jette $
+$Id: NEWS 12712 2007-11-29 00:12:08Z jette $
diff --git a/doc/html/bluegene.shtml b/doc/html/bluegene.shtml
index 9121d3ce10df4a07c79aa107f8825de5ab87d002..8a4d01db2f3301ea0f98ae8eb645d9781c346e28 100644
--- a/doc/html/bluegene.shtml
+++ b/doc/html/bluegene.shtml
@@ -83,7 +83,7 @@ date
<h3><a name="naming">Naming Convensions</a></h3>
<p>The naming of base partitions includes a three-digit suffix representing the its
-coordinates in the X, Y and Z dimensions with a zero origin.
+coordinates in the X, Y and Z dimensions with a zero origin.
For example, "bg012" represents the base partition whose coordinate is at X=0, Y=1 and Z=2. In a system
configured with <i>small blocks</i> (any block less than a full base partition) there will be divisions
into the base partition notation. For example, if there were 64 psets in the
@@ -100,6 +100,24 @@ For example, "bg[620x731]" is used to represent the eight base partitions enclos
with endpoints bg620 and bg731 (bg620, bg621, bg630, bg631, bg720, bg721,
bg730 and bg731).</p></a>
+<p>
+<b>IMPORTANT:</b> As of SLURM version 1.2 SLURM can handle a bluegene
+system of size 36x36x36. To try to keep with the 'three-digit suffix
+representing the its coordinates in the X, Y and Z dimensions with a
+zero origin', we now support A-Z as valid numbers. This makes it so
+the prefix <b>must always be lower case</b>, and any letters in the
+three-digit suffix <b> must always be upper case</b>. This schema
+should be used in your slurm.conf file and in your bluegene.conf file
+if you put a prefix there even though it is not necessary there. This
+schema should also be used to specify midplanes or locations in
+configure mode of smap.
+
+<br>
+valid: bgl[000xC44] bgl000 bglZZZ
+<br>
+invalid: BGL[000xC44] BglC00 bglb00 Bglzzz
+</p>
+
<p>One new tool provided is <i>smap</i>.
As of SLURM verison 1.2, <i>sview</i> is
another new tool offering even more viewing and configuring options.
diff --git a/doc/html/team.shtml b/doc/html/team.shtml
index 33695062f6da22eb6fcf4e998f4af53824159cdb..a2b16f1ff8e312350081d9480c225b2905598e87 100644
--- a/doc/html/team.shtml
+++ b/doc/html/team.shtml
@@ -49,6 +49,7 @@ Networking, Italy)</li>
<li>Ashley Pittman (Quadrics)</li>
<li>Andy Riebs (HP)</li>
<li>Asier Roa (Barcelona Supercomputer Center, Spain)<li>
+<li>Miguel Ros (Barcelona Supercomputer Center, Spain)<li>
<li>Federico Sacerdoti (D.E. Shaw)<li>
<li>Jeff Squyres (LAM MPI)</li>
<li>Prashanth Tamraparni (HP, India)</li>
diff --git a/doc/man/man1/salloc.1 b/doc/man/man1/salloc.1
index 490c9c252ad78ec5acbd8642a77d19041fa40a5d..3fe996045934959cae130c86a170e777dc7a56a1 100644
--- a/doc/man/man1/salloc.1
+++ b/doc/man/man1/salloc.1
@@ -24,10 +24,9 @@ a specific time of day (seconds are optional).
You may also specify \fImidnight\fR, \fInoon\fR, or
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
-You can also say what day the job will be run, by giving
-a date in the form \fImonth\-name\fR day with an optional year,
-or giving a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
-or \fIDD.MM.YY\fR. You can also
+You can also say what day the job will be run, by specifying
+a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
+or \fIMM.DD.YY\fR. You can also
give times like \fInow + count time\-units\fR, where the time\-units
can be \fIminutes\fR, \fIhours\fR, \fIdays\fR, or \fIweeks\fR
and you can tell SLURM to run the job today with the keyword
@@ -61,7 +60,7 @@ If no nodes have the requested features, then the job will be rejected
by the slurm job manager.
.TP
-\fB\-\-comment\fR
+\fB\-\-comment\fR=<\fIstring\fR>
An arbitrary comment.
.TP
diff --git a/doc/man/man1/sbatch.1 b/doc/man/man1/sbatch.1
index 532d4d8e6e7b1357f43c77876b86b97a1d218c1b..e317a291817843db5fc868272c1750d92fd97f11 100644
--- a/doc/man/man1/sbatch.1
+++ b/doc/man/man1/sbatch.1
@@ -34,10 +34,9 @@ a specific time of day (seconds are optional).
You may also specify \fImidnight\fR, \fInoon\fR, or
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
-You can also say what day the job will be run, by giving
-a date in the form \fImonth\-name\fR day with an optional year,
-or giving a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
-or \fIDD.MM.YY\fR. You can also
+You can also say what day the job will be run, by specifying
+a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
+or \fIMM.DD.YY\fR. You can also
give times like \fInow + count time\-units\fR, where the time\-units
can be \fIminutes\fR, \fIhours\fR, \fIdays\fR, or \fIweeks\fR
and you can tell SLURM to run the job today with the keyword
@@ -78,7 +77,7 @@ the \-\-cpus\-per\-task=3 options, the controller knows that each task requires
of 4 nodes, one for each of the 4 tasks.
.TP
-\fB\-\-comment\fR
+\fB\-\-comment\fR=<\fIstring\fR>
An arbitrary comment.
.TP
@@ -120,15 +119,25 @@ The order of the node names in the list is not important; the node names
will be sorted my SLURM.
.TP
-\fB\-\-get\-user\-env\fR[=\fItimeout\fR]
+\fB\-\-get\-user\-env\fR[=\fItimeout\fR][\fImode\fR]
This option will tell sbatch to retrieve the
-login environment variables for the user specified in the \-\-uid option.
-The environment variables are retrieved by running "su - <username> -c
-/usr/bin/env" and parsing the output. Be aware that any environment
-variables already set in sbatch's environment will take precedence over any
-environment variables in the user's login environment.
-Optional timeout value is in seconds. Default value is 8 seconds.
-NOTE: This option only works if the caller has an effective uid of "root".
+login environment variables for the user specified in the \fB\-\-uid\fR option.
+The environment variables are retrieved by running something of this sort
+"su - <username> -c /usr/bin/env" and parsing the output.
+Be aware that any environment variables already set in sbatch's environment
+will take precedence over any environment variables in the user's
+login environment.
+The optional \fItimeout\fR value is in seconds. Default value is 8 seconds.
+The optional \fImode\fR value control the "su" options.
+With a \fImode\fR value of "S", "su" is executed without the "\-" option.
+With a \fImode\fR value of "L", "su" is executed with the "\-" option,
+replicating the login environment.
+If \fImode\fR not specified, the mode established at SLURM build time
+is used.
+Example of use include "\-\-get\-user\-env", "\-\-get\-user\-env=10"
+"\-\-get\-user\-env=10L", and "\-\-get\-user\-env=S".
+NOTE: This option only works if the caller has an
+effective uid of "root".
This option was originally created for use by Moab.
.TP
diff --git a/doc/man/man1/scontrol.1 b/doc/man/man1/scontrol.1
index 13255266ca8dd84c31a00c325b15458af6d73ef2..74ad6370fa2ce841af14ff2a4bae12553a76b49f 100644
--- a/doc/man/man1/scontrol.1
+++ b/doc/man/man1/scontrol.1
@@ -348,10 +348,9 @@ a specific time of day (seconds are optional).
You may also specify \fImidnight\fR, \fInoon\fR, or
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
-You can also say what day the job will be run, by giving
-a date in the form \fImonth\-name\fR day with an optional year,
-or giving a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
-or \fIDD.MM.YY\fR. You can also
+You can also say what day the job will be run, by specifying
+a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
+or \fIMM.DD.YY\fR. You can also
give times like \fInow + count time\-units\fR, where the time\-units
can be \fIminutes\fR, \fIhours\fR, \fIdays\fR, or \fIweeks\fR
and you can tell SLURM to run the job today with the keyword
diff --git a/doc/man/man1/srun.1 b/doc/man/man1/srun.1
index bcbc3cb435205636644f00727385f1c492d8bc42..c2ab45e664d27cd6840c0ae8dbb32d77ec125938 100644
--- a/doc/man/man1/srun.1
+++ b/doc/man/man1/srun.1
@@ -1,4 +1,4 @@
-\." $Id: srun.1 12574 2007-10-26 17:00:52Z jette $
+\." $Id: srun.1 12697 2007-11-27 22:02:29Z jette $
.\"
.TH SRUN "1" "July 2007" "srun 1.2" "slurm components"
@@ -99,10 +99,9 @@ a specific time of day (seconds are optional).
You may also specify \fImidnight\fR, \fInoon\fR, or
\fIteatime\fR (4pm) and you can have a time\-of\-day suffixed
with \fIAM\fR or \fIPM\fR for running in the morning or the evening.
-You can also say what day the job will be run, by giving
-a date in the form \fImonth\-name\fR day with an optional year,
-or giving a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
-or \fIDD.MM.YY\fR. You can also
+You can also say what day the job will be run, by specifying
+a date of the form \fIMMDDYY\fR or \fIMM/DD/YY\fR
+or \fIMM.DD.YY\fR. You can also
give times like \fInow + count time\-units\fR, where the time\-units
can be \fIseconds\fR (default), \fIminutes\fR, \fIhours\fR,
\fIdays\fR, or \fIweeks\fR and you can tell SLURM to run
@@ -221,7 +220,7 @@ tasks will be allocated per node as possible while satisfying
the \fB\-c\fR restriction.
.TP
-\fB\-\-comment\fR
+\fB\-\-comment\fR=<\fIstring\fR>
An arbitrary comment.
.TP
@@ -282,16 +281,26 @@ even if consumable resources are enabled
(e.g. \fBSelectType=select/cons_res\fR).
.TP
-\fB\-\-get\-user\-env\fR[=\fItimeout\fR]
+\fB\-\-get\-user\-env\fR[=\fItimeout\fR][\fImode\fR]
For a batch script submission, this option will tell srun to retrieve the
-login environment variables for the user specified in the \-\-uid option.
-The environment variables are retrieved by running "su - <username> -c
-/usr/bin/env" and parsing the output. Be aware that any environment
-variables already set in srun's environment will take precedence over any
-environment variables in the user's login environment.
-Optional timeout value is in seconds. Default value is 8 seconds.
+login environment variables for the user specified in the \fB\-\-uid\fR option.
+The environment variables are retrieved by running something of this sort
+"su - <username> -c /usr/bin/env" and parsing the output.
+Be aware that any environment variables already set in srun's environment
+will take precedence over any environment variables in the user's
+login environment.
+The optional \fItimeout\fR value is in seconds. Default value is 8 seconds.
+The optional \fImode\fR value control the "su" options.
+With a \fImode\fR value of "S", "su" is executed without the "\-" option.
+With a \fImode\fR value of "L", "su" is executed with the "\-" option,
+replicating the login environment.
+If \fImode\fR not specified, the mode established at SLURM build time
+is used.
+Example of use include "\-\-get\-user\-env", "\-\-get\-user\-env=10"
+"\-\-get\-user\-env=10L", and "\-\-get\-user\-env=S".
NOTE: This option only works if the caller has an
-effective uid of "root", and only takes effect in batch mode (\-b/\-\-batch).
+effective uid of "root", and only takes effect in batch mode
+(\fB\-b\fR/\fB\-\-batch\fR).
This option was originally created for use by Moab.
.TP
@@ -622,6 +631,8 @@ comma\-separated and case insensitive types are recongnized:
\fBBULK_XFER\fR and adapter names (e.g. \fBSNI0\fR and \fBSNI1\fR).
For more information, on IBM systems see \fIpoe\fR documenation on
the environment variables \fBMP_EUIDEVICE\fR and \fBMP_USE_BULK_XFER\fR.
+Note that only four jobs steps may be active at once on a node with the
+\fBBULK_XFER\fR option due to limitations in the Federation switch driver.
.TP
\fB\-\-nice\fR[=\fIadjustment]\fR
@@ -1185,6 +1196,19 @@ offload work from the srun command to the applications and
likely increase the vulernability to failures.
The default value is 32.
.TP
+\fBPMI_FANOUT_OFF_HOST\fR
+This is used exclusively with PMI (MPICH2 and MVAPICH2) and
+controls the fanout of data communications. The srun command
+sends messages to application programs (via the PMI library)
+and those applications may be called upon to forward that
+data to additional tasks. By default, srun sends one message
+per host and one task on that host forwards the data to other
+tasks on that host up to \fBPMI_FANOUT\fR.
+If \fBPMI_FANOUT_OFF_HOST\fR is defined, the user task
+may be required to forward the data to tasks on other hosts.
+Setting \fBPMI_FANOUT_OFF_HOST\fR may increase performance
+and vulernability to failures.
+.TP
\fBPMI_TIME\fR
This is used exclusively with PMI (MPICH2 and MVAPICH2) and
controls how much the communications from the tasks to the
diff --git a/doc/man/man5/slurm.conf.5 b/doc/man/man5/slurm.conf.5
index 2883fa5cae2b6697e04368be7c288859006d3307..2ff61088c9bc4ac46756626a3c51b52218250942 100644
--- a/doc/man/man5/slurm.conf.5
+++ b/doc/man/man5/slurm.conf.5
@@ -234,6 +234,12 @@ If the job fails to terminate gracefully
in the interval specified, it will be forcably terminated.
The default value is 30 seconds.
May not exceed 65533.
+
+.TP
+\fBMailProg\fR
+Fully qualified pathname to the program used to send email per user request.
+The default value is "/bin/mail".
+
.TP
\fBMaxJobCount\fR
The maximum number of jobs SLURM can have in its active database
@@ -1035,10 +1041,6 @@ users from directly using those resources.
Possible values are "YES" and "NO".
The default value is "NO".
.TP
-\fBMailProg\fR
-Fully qualified pathname to the program used to send email per user request.
-The default value is "/bin/mail".
-.TP
\fBMaxNodes\fR
Maximum count of nodes (or base partitions for BlueGene systems) which
may be allocated to any single job.
diff --git a/slurm.spec b/slurm.spec
index d3c1fbc892d4d574e6f307660f3aba8041594520..33e2a50cbbecc5997057cfc7e57abc68d127a69d 100644
--- a/slurm.spec
+++ b/slurm.spec
@@ -1,4 +1,4 @@
-# $Id: slurm.spec 12605 2007-11-01 23:12:43Z jette $
+# $Id: slurm.spec 12730 2007-11-29 19:51:09Z jette $
#
# Note that this package is not relocatable
@@ -34,7 +34,6 @@
%slurm_without_opt bluegene
%slurm_without_opt auth_none
%slurm_without_opt debug
-%slurm_without_opt sgijob
# Build with munge by default on all platforms (disable with --without munge)
%slurm_with_opt munge
@@ -48,20 +47,27 @@
%endif
# Define with_aix on AIX systems (for proctrack)
-%ifos aix
+%ifos aix5.3
%slurm_with_opt aix
%endif
+# Build with sgijob on CHAOS systems
+# (add elan too when it is available)
+%if %{?chaos}0
+%slurm_with_opt sgijob
+%else
+%slurm_without_opt sgijob
+%endif
Name: slurm
-Version: 1.2.19
+Version: 1.2.20
Release: 1%{?dist}
Summary: Simple Linux Utility for Resource Management
License: GPL
Group: System Environment/Base
-Source: slurm-1.2.19.tar.bz2
+Source: slurm-1.2.20.tar.bz2
BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}
URL: http://www.llnl.gov/linux/slurm
BuildRequires: openssl-devel >= 0.9.6 openssl >= 0.9.6
@@ -205,7 +211,7 @@ SLURM process tracking plugin for SGI job containers.
#############################################################################
%prep
-%setup -n slurm-1.2.19
+%setup -n slurm-1.2.20
%build
%configure --program-prefix=%{?_program_prefix:%{_program_prefix}} \
diff --git a/src/api/job_info.c b/src/api/job_info.c
index b3a70a6afa737bf0c40273b3170f8bc8d240ff89..829c6e32905889437dde9578dbebb8070964da8e 100644
--- a/src/api/job_info.c
+++ b/src/api/job_info.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* job_info.c - get/print the job state information of slurm
- * $Id: job_info.c 12249 2007-09-11 00:48:52Z jette $
+ * $Id: job_info.c 12627 2007-11-06 19:48:55Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -87,10 +87,11 @@ slurm_print_job_info_msg ( FILE* out, job_info_msg_t *jinfo, int one_liner )
static void _sprint_range(char *str, uint16_t lower, uint16_t upper)
{
- char tmp[128];
- convert_num_unit((float)lower, str, UNIT_NONE);
+ /* Note: We don't have the size of str here */
+ convert_num_unit((float)lower, str, 16, UNIT_NONE);
if (upper > 0) {
- convert_num_unit((float)upper, tmp, UNIT_NONE);
+ char tmp[128];
+ convert_num_unit((float)upper, tmp, sizeof(tmp), UNIT_NONE);
strcat(str, "-");
strcat(str, tmp);
}
@@ -252,7 +253,8 @@ slurm_sprint_job_info ( job_info_t * job_ptr, int one_liner )
/****** Line 6a (optional) ******/
#if 0
/* mainly for debugging */
- convert_num_unit((float)job_ptr->num_cpu_groups, tmp1, UNIT_NONE);
+ convert_num_unit((float)job_ptr->num_cpu_groups, tmp1, sizeof(tmp1),
+ UNIT_NONE);
snprintf(tmp_line, sizeof(tmp_line),
"NumCPUGroups=%s ",
tmp1);
@@ -300,9 +302,11 @@ slurm_sprint_job_info ( job_info_t * job_ptr, int one_liner )
}
/****** Line 7 ******/
- convert_num_unit((float)job_ptr->num_procs, tmp1, UNIT_NONE);
+ convert_num_unit((float)job_ptr->num_procs, tmp1, sizeof(tmp1),
+ UNIT_NONE);
#ifdef HAVE_BG
- convert_num_unit((float)job_ptr->num_nodes, tmp2, UNIT_NONE);
+ convert_num_unit((float)job_ptr->num_nodes, tmp2, sizeof(tmp2),
+ UNIT_NONE);
sprintf(tmp_line, "ReqProcs=%s MinBPs=%s ", tmp1, tmp2);
#else
_sprint_range(tmp2, job_ptr->num_nodes, job_ptr->max_nodes);
@@ -356,8 +360,10 @@ slurm_sprint_job_info ( job_info_t * job_ptr, int one_liner )
xstrcat(out, "\n ");
/****** Line 10 ******/
- convert_num_unit((float)job_ptr->job_min_memory, tmp1, UNIT_NONE);
- convert_num_unit((float)job_ptr->job_min_tmp_disk, tmp2, UNIT_NONE);
+ convert_num_unit((float)job_ptr->job_min_memory, tmp1, sizeof(tmp1),
+ UNIT_NONE);
+ convert_num_unit((float)job_ptr->job_min_tmp_disk, tmp2, sizeof(tmp2),
+ UNIT_NONE);
snprintf(tmp_line, sizeof(tmp_line),
"MinMemory=%s MinTmpDisk=%s Features=%s",
tmp1, tmp2, job_ptr->features);
diff --git a/src/api/partition_info.c b/src/api/partition_info.c
index 02135e3be50180b99be56da04102d6808600d94c..5729074ce9cccd7ec1d7dff609ccc79498eb2a2a 100644
--- a/src/api/partition_info.c
+++ b/src/api/partition_info.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* partition_info.c - get/print the partition state information of slurm
- * $Id: partition_info.c 11781 2007-07-02 23:00:56Z jette $
+ * $Id: partition_info.c 12627 2007-11-06 19:48:55Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -105,17 +105,19 @@ char *slurm_sprint_partition_info ( partition_info_t * part_ptr,
int one_liner )
{
int j;
- char tmp1[7], tmp2[7];
+ char tmp1[16], tmp2[16];
char tmp_line[MAXHOSTRANGELEN];
char *out = NULL;
/****** Line 1 ******/
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->total_nodes, tmp1, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_nodes, tmp1, sizeof(tmp1),
+ UNIT_NONE);
#else
- sprintf(tmp1, "%u", part_ptr->total_nodes);
+ snprintf(tmp1, sizeof(tmp1), "%u", part_ptr->total_nodes);
#endif
- convert_num_unit((float)part_ptr->total_cpus, tmp2, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_cpus, tmp2, sizeof(tmp2),
+ UNIT_NONE);
snprintf(tmp_line, sizeof(tmp_line),
"PartitionName=%s TotalNodes=%s TotalCPUs=%s ",
part_ptr->name, tmp1, tmp2);
@@ -169,9 +171,10 @@ char *slurm_sprint_partition_info ( partition_info_t * part_ptr,
/****** Line 3 ******/
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->min_nodes, tmp1, UNIT_NONE);
+ convert_num_unit((float)part_ptr->min_nodes, tmp1, sizeof(tmp1),
+ UNIT_NONE);
#else
- sprintf(tmp1, "%u", part_ptr->min_nodes);
+ snprintf(tmp1, sizeof(tmp1), "%u", part_ptr->min_nodes);
#endif
sprintf(tmp_line, "MinNodes=%s ", tmp1);
xstrcat(out, tmp_line);
@@ -180,9 +183,10 @@ char *slurm_sprint_partition_info ( partition_info_t * part_ptr,
sprintf(tmp_line, "MaxNodes=UNLIMITED ");
else {
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->max_nodes, tmp1, UNIT_NONE);
+ convert_num_unit((float)part_ptr->max_nodes, tmp1, sizeof(tmp1),
+ UNIT_NONE);
#else
- sprintf(tmp1, "%u", part_ptr->max_nodes);
+ snprintf(tmp1, sizeof(tmp1),"%u", part_ptr->max_nodes);
#endif
sprintf(tmp_line, "MaxNodes=%s ", tmp1);
}
diff --git a/src/api/pmi_server.c b/src/api/pmi_server.c
index d0385508917ad14dda2e37c58c76e74fb238d962..406da910d5b099fda0796573897a1cf0a01a17c0 100644
--- a/src/api/pmi_server.c
+++ b/src/api/pmi_server.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* pmi.c - Global PMI data as maintained within srun
- * $Id: pmi_server.c 12315 2007-09-13 23:56:02Z jette $
+ * $Id: pmi_server.c 12620 2007-11-05 19:00:45Z jette $
*****************************************************************************
* Copyright (C) 2005-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -182,7 +182,7 @@ static void *_agent(void *x)
struct kvs_hosts *kvs_host_list;
int i, j, kvs_set_cnt = 0, host_cnt, pmi_fanout = 32;
int msg_sent = 0, max_forward = 0;
- char *tmp;
+ char *tmp, *fanout_off_host;
pthread_t msg_id;
pthread_attr_t attr;
DEF_TIMERS;
@@ -193,6 +193,7 @@ static void *_agent(void *x)
if (pmi_fanout < 1)
pmi_fanout = 32;
}
+ fanout_off_host = getenv("PMI_FANOUT_OFF_HOST");
/* only send one message to each host,
* build table of the ports on each host */
@@ -213,7 +214,8 @@ static void *_agent(void *x)
for (j=(i+1); j<args->barrier_xmit_cnt; j++) {
if (args->barrier_xmit_ptr[j].port == 0)
continue; /* already sent message */
- if (strcmp(args->barrier_xmit_ptr[i].hostname,
+ if ((fanout_off_host == NULL) &&
+ strcmp(args->barrier_xmit_ptr[i].hostname,
args->barrier_xmit_ptr[j].hostname))
continue; /* another host */
kvs_host_list[host_cnt].task_id = 0; /* not avail */
diff --git a/src/api/signal.c b/src/api/signal.c
index c46aa41d997a0d1ea855fbb3fba0e6cdf59dc68e..3091fa6de220ccc1b2856ec3b8c996ca7e7fc784 100644
--- a/src/api/signal.c
+++ b/src/api/signal.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* signal.c - Send a signal to a slurm job or job step
- * $Id: signal.c 10574 2006-12-15 23:38:29Z jette $
+ * $Id: signal.c 12647 2007-11-12 17:09:47Z da $
*****************************************************************************
* Copyright (C) 2005 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -172,7 +172,7 @@ _local_send_recv_rc_msgs(const char *nodelist, slurm_msg_type_t type,
msg->msg_type = type;
msg->data = data;
- if((ret_list = slurm_send_recv_msgs(nodelist, msg, 10000))) {
+ if((ret_list = slurm_send_recv_msgs(nodelist, msg, 0))) {
while((ret_data_info = list_pop(ret_list))) {
temp_rc = slurm_get_return_code(ret_data_info->type,
ret_data_info->data);
@@ -397,7 +397,7 @@ static int _terminate_batch_script_step(
return -1;
}
free(name);
- i = slurm_send_recv_rc_msg_only_one(&msg, &rc, 10000);
+ i = slurm_send_recv_rc_msg_only_one(&msg, &rc, 0);
if (i != 0)
rc = i;
diff --git a/src/common/daemonize.c b/src/common/daemonize.c
index f62ef9221fd14f71c81d485008ef9bfa860fdd4b..528aaaa145511b81ca7da01a1706fc53368dd99d 100644
--- a/src/common/daemonize.c
+++ b/src/common/daemonize.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* daemonize.c - daemonization routine
- * $Id: daemonize.c 10927 2007-02-05 18:43:52Z jette $
+ * $Id: daemonize.c 12723 2007-11-29 18:55:48Z jette $
*****************************************************************************
* Copyright (C) 2002 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -139,12 +139,14 @@ read_pidfile(const char *pidfile, int *pidfd)
if (fscanf(fp, "%lu", &pid) < 1) {
error ("Possible corrupt pidfile `%s'", pidfile);
+ (void) close(fd);
return ((pid_t) 0);
}
if ((lpid = fd_is_read_lock_blocked(fd)) == (pid_t) 0) {
verbose ("pidfile not locked, assuming no running daemon");
- return (lpid);
+ (void) close(fd);
+ return ((pid_t) 0);
}
if (lpid != (pid_t) pid)
@@ -154,7 +156,7 @@ read_pidfile(const char *pidfile, int *pidfd)
if (pidfd != NULL)
*pidfd = fd;
else
- (void) close(fd); /* Ignore errors */
+ (void) close(fd);
return (lpid);
}
diff --git a/src/common/env.c b/src/common/env.c
index 6502fd56b9e808142545e96d172ae613e1446c91..ac32d5254bc407784c5faa6bffee0023e5584675 100644
--- a/src/common/env.c
+++ b/src/common/env.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* src/common/env.c - add an environment variable to environment vector
- * $Id: env.c 12592 2007-10-31 21:12:49Z jette $
+ * $Id: env.c 12697 2007-11-27 22:02:29Z jette $
*****************************************************************************
* Copyright (C) 2002-2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -77,6 +77,8 @@ strong_alias(env_array_overwrite_fmt, slurm_env_array_overwrite_fmt);
#define SU_WAIT_MSEC 8000 /* 8000 msec for /bin/su to return user
* env vars for --get-user-env option */
+#define ENV_BUFSIZE (64 * 1024)
+
/*
* Return pointer to `name' entry in environment if found, or
* pointer to the last entry (i.e. NULL) if `name' is not
@@ -153,12 +155,12 @@ int
setenvfs(const char *fmt, ...)
{
va_list ap;
- char buf[BUFSIZ];
+ char buf[ENV_BUFSIZE];
char *bufcpy;
int rc;
va_start(ap, fmt);
- vsnprintf(buf, BUFSIZ, fmt, ap);
+ vsnprintf(buf, ENV_BUFSIZE, fmt, ap);
va_end(ap);
bufcpy = xstrdup(buf);
@@ -169,7 +171,7 @@ setenvfs(const char *fmt, ...)
int
setenvf(char ***envp, const char *name, const char *fmt, ...)
{
- char buf[BUFSIZ];
+ char buf[ENV_BUFSIZE];
char **ep = NULL;
char *str = NULL;
va_list ap;
@@ -177,7 +179,7 @@ setenvf(char ***envp, const char *name, const char *fmt, ...)
char *bufcpy;
va_start(ap, fmt);
- vsnprintf (buf, BUFSIZ, fmt, ap);
+ vsnprintf (buf, ENV_BUFSIZE, fmt, ap);
va_end(ap);
bufcpy = xstrdup(buf);
@@ -944,7 +946,7 @@ char **env_array_create(void)
int env_array_append_fmt(char ***array_ptr, const char *name,
const char *value_fmt, ...)
{
- char buf[BUFSIZ];
+ char buf[ENV_BUFSIZE];
char **ep = NULL;
char *str = NULL;
va_list ap;
@@ -959,7 +961,7 @@ int env_array_append_fmt(char ***array_ptr, const char *name,
}
va_start(ap, value_fmt);
- vsnprintf (buf, BUFSIZ, value_fmt, ap);
+ vsnprintf (buf, ENV_BUFSIZE, value_fmt, ap);
va_end(ap);
ep = _find_name_in_env(*array_ptr, name);
@@ -1020,7 +1022,7 @@ int env_array_append(char ***array_ptr, const char *name,
int env_array_overwrite_fmt(char ***array_ptr, const char *name,
const char *value_fmt, ...)
{
- char buf[BUFSIZ];
+ char buf[ENV_BUFSIZE];
char **ep = NULL;
char *str = NULL;
va_list ap;
@@ -1035,7 +1037,7 @@ int env_array_overwrite_fmt(char ***array_ptr, const char *name,
}
va_start(ap, value_fmt);
- vsnprintf (buf, BUFSIZ, value_fmt, ap);
+ vsnprintf (buf, ENV_BUFSIZE, value_fmt, ap);
va_end(ap);
xstrfmtcat (str, "%s=%s", name, buf);
@@ -1154,10 +1156,11 @@ static int _env_array_entry_splitter(const char *entry,
*/
static int _env_array_putenv(const char *string)
{
- char name[BUFSIZ];
- char value[BUFSIZ];
+ char name[ENV_BUFSIZE];
+ char value[ENV_BUFSIZE];
- if (!_env_array_entry_splitter(string, name, BUFSIZ, value, BUFSIZ))
+ if (!_env_array_entry_splitter(string, name, ENV_BUFSIZE, value,
+ ENV_BUFSIZE))
return 0;
if (setenv(name, value, 1) == -1)
return 0;
@@ -1189,14 +1192,15 @@ void env_array_set_environment(char **env_array)
void env_array_merge(char ***dest_array, const char **src_array)
{
char **ptr;
- char name[BUFSIZ];
- char value[BUFSIZ];
+ char name[ENV_BUFSIZE];
+ char value[ENV_BUFSIZE];
if (src_array == NULL)
return;
for (ptr = (char **)src_array; *ptr != NULL; ptr++) {
- _env_array_entry_splitter(*ptr, name, BUFSIZ, value, BUFSIZ);
+ _env_array_entry_splitter(*ptr, name, ENV_BUFSIZE, value,
+ ENV_BUFSIZE);
env_array_overwrite(dest_array, name, value);
}
}
@@ -1224,8 +1228,8 @@ static void _strip_cr_nl(char *line)
*/
char **_load_env_cache(const char *username)
{
- char *state_save_loc, fname[BUFSIZ];
- char line[BUFSIZ], name[BUFSIZ], value[BUFSIZ];
+ char *state_save_loc, fname[ENV_BUFSIZE];
+ char line[ENV_BUFSIZE], name[ENV_BUFSIZE], value[ENV_BUFSIZE];
char **env = NULL;
FILE *fp;
int i;
@@ -1247,10 +1251,11 @@ char **_load_env_cache(const char *username)
info("Getting cached environment variables at %s", fname);
env = env_array_create();
while (1) {
- if (!fgets(line, BUFSIZ, fp))
+ if (!fgets(line, ENV_BUFSIZE, fp))
break;
_strip_cr_nl(line);
- _env_array_entry_splitter(line, name, BUFSIZ, value, BUFSIZ);
+ _env_array_entry_splitter(line, name, ENV_BUFSIZE, value,
+ ENV_BUFSIZE);
env_array_overwrite(&env, name, value);
}
fclose(fp);
@@ -1267,17 +1272,18 @@ char **_load_env_cache(const char *username)
* in the event that option 1 times out.
*
* timeout value is in seconds or zero for default (8 secs)
+ * mode is 1 for short ("su <user>"), 2 for long ("su - <user>")
* On error, returns NULL.
*
* NOTE: The calling process must have an effective uid of root for
* this function to succeed.
*/
-char **env_array_user_default(const char *username, int timeout)
+char **env_array_user_default(const char *username, int timeout, int mode)
{
FILE *su;
- char line[BUFSIZ];
- char name[BUFSIZ];
- char value[BUFSIZ];
+ char line[ENV_BUFSIZE];
+ char name[ENV_BUFSIZE];
+ char value[ENV_BUFSIZE];
char **env = NULL;
char *starttoken = "XXXXSLURMSTARTPARSINGHEREXXXX";
char *stoptoken = "XXXXSLURMSTOPPARSINGHEREXXXXX";
@@ -1311,11 +1317,17 @@ char **env_array_user_default(const char *username, int timeout)
snprintf(cmdstr, sizeof(cmdstr),
"echo; echo; echo; echo %s; env; echo %s",
starttoken, stoptoken);
+ if (mode == 1)
+ execl("/bin/su", "su", username, "-c", cmdstr, NULL);
+ else if (mode == 2)
+ execl("/bin/su", "su", "-", username, "-c", cmdstr, NULL);
+ else { /* Default system configuration */
#ifdef LOAD_ENV_NO_LOGIN
- execl("/bin/su", "su", username, "-c", cmdstr, NULL);
+ execl("/bin/su", "su", username, "-c", cmdstr, NULL);
#else
- execl("/bin/su", "su", "-", username, "-c", cmdstr, NULL);
+ execl("/bin/su", "su", "-", username, "-c", cmdstr, NULL);
#endif
+ }
exit(1);
}
@@ -1353,7 +1365,7 @@ char **env_array_user_default(const char *username, int timeout)
}
if (!(ufds.revents & POLLIN))
break;
- while (fgets(line, BUFSIZ, su)) {
+ while (fgets(line, ENV_BUFSIZE, su)) {
if (!strncmp(line, starttoken, len)) {
found = 1;
break;
@@ -1393,14 +1405,15 @@ char **env_array_user_default(const char *username, int timeout)
/* stop at the line containing the stoptoken string */
if (!(ufds.revents & POLLIN))
break;
- if ((fgets(line, BUFSIZ, su) == 0) ||
+ if ((fgets(line, ENV_BUFSIZE, su) == 0) ||
(!strncmp(line, stoptoken, len))) {
found = 1;
break;
}
_strip_cr_nl(line);
- _env_array_entry_splitter(line, name, BUFSIZ, value, BUFSIZ);
+ _env_array_entry_splitter(line, name, ENV_BUFSIZE, value,
+ ENV_BUFSIZE);
env_array_overwrite(&env, name, value);
}
close(fildes[0]);
diff --git a/src/common/env.h b/src/common/env.h
index 01ed6c770402abd97e1eee63a4176f151ca9337a..94000ed24f202354845e431ba3c42b5d84fa4925 100644
--- a/src/common/env.h
+++ b/src/common/env.h
@@ -238,15 +238,20 @@ void env_array_set_environment(char **env_array);
/*
* Return an array of strings representing the specified user's default
- * environment variables, as determined by calling (more-or-less)
- * "/bin/su - <username> -c /usr/bin/env".
+ * environment variables following a two-prongged approach.
+ * 1. Execute (more or less): "/bin/su - <username> -c /usr/bin/env"
+ * Depending upon the user's login scripts, this may take a very
+ * long time to complete or possibly never return
+ * 2. Load the user environment from a cache file. This is used
+ * in the event that option 1 times out.
*
* timeout value is in seconds or zero for default (8 secs)
+ * mode is 1 for short ("su <user>"), 2 for long ("su - <user>")
* On error, returns NULL.
*
* NOTE: The calling process must have an effective uid of root for
* this function to succeed.
*/
-char **env_array_user_default(const char *username, int timeout);
+char **env_array_user_default(const char *username, int timeout, int mode);
#endif
diff --git a/src/common/hostlist.c b/src/common/hostlist.c
index 61fee0aba43f00a07e3fb4b653e29af47ac5158a..36b489af1844a10cf32b1cd9987a308d8a555c11 100644
--- a/src/common/hostlist.c
+++ b/src/common/hostlist.c
@@ -1,5 +1,5 @@
/*****************************************************************************\
- * $Id: hostlist.c 12538 2007-10-23 17:11:04Z jette $
+ * $Id: hostlist.c 12632 2007-11-06 23:27:07Z da $
*****************************************************************************
* $LSDId: hostlist.c,v 1.14 2003/10/14 20:11:54 grondo Exp $
*****************************************************************************
@@ -1569,8 +1569,6 @@ static int _parse_box_range(char *str, struct _range *ranges, int len, int *coun
for(i = 0; i<3; i++) {
if ((str[i] >= '0') && (str[i] <= '9'))
a[i] = str[i] - '0';
- else if ((str[i] >= 'a') && (str[i] <= 'z'))
- a[i] = str[i] - 'a' + 10;
else if ((str[i] >= 'A') && (str[i] <= 'Z'))
a[i] = str[i] - 'A' + 10;
else
@@ -1578,8 +1576,6 @@ static int _parse_box_range(char *str, struct _range *ranges, int len, int *coun
if ((str[i+4] >= '0') && (str[i+4] <= '9'))
b[i] = str[i+4] - '0';
- else if ((str[i+4] >= 'a') && (str[i+4] <= 'z'))
- b[i] = str[i+4] - 'a' + 10;
else if ((str[i+4] >= 'A') && (str[i+4] <= 'Z'))
b[i] = str[i+4] - 'A' + 10;
else
diff --git a/src/common/node_select.c b/src/common/node_select.c
index e6020814d53f63b5ca57fb39660dc9bb8abca705..2f9a57d0cb56856cd0d57f96ce6d211bb85fb2d5 100644
--- a/src/common/node_select.c
+++ b/src/common/node_select.c
@@ -9,7 +9,7 @@
* the plugin. This is because functions required by the plugin can not be
* resolved on the front-end nodes, so we can't load the plugins there.
*
- * $Id: node_select.c 11590 2007-05-25 18:52:33Z da $
+ * $Id: node_select.c 12627 2007-11-06 19:48:55Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -1071,7 +1071,7 @@ extern char *select_g_sprint_jobinfo(select_jobinfo_t jobinfo,
{
uint16_t geometry[SYSTEM_DIMENSIONS];
int i;
- char max_procs_char[7], start_char[32];
+ char max_procs_char[8], start_char[32];
char *tmp_image = "default";
if (buf == NULL) {
@@ -1108,7 +1108,8 @@ extern char *select_g_sprint_jobinfo(select_jobinfo_t jobinfo,
sprintf(max_procs_char, "None");
else
convert_num_unit((float)jobinfo->max_procs,
- max_procs_char, UNIT_NONE);
+ max_procs_char, sizeof(max_procs_char),
+ UNIT_NONE);
if (jobinfo->start[0] == (uint16_t) NO_VAL)
sprintf(start_char, "None");
else {
@@ -1130,7 +1131,8 @@ extern char *select_g_sprint_jobinfo(select_jobinfo_t jobinfo,
sprintf(max_procs_char, "None");
else
convert_num_unit((float)jobinfo->max_procs,
- max_procs_char, UNIT_NONE);
+ max_procs_char, sizeof(max_procs_char),
+ UNIT_NONE);
if (jobinfo->start[0] == (uint16_t) NO_VAL)
sprintf(start_char, "None");
else {
@@ -1181,7 +1183,8 @@ extern char *select_g_sprint_jobinfo(select_jobinfo_t jobinfo,
sprintf(max_procs_char, "None");
else
convert_num_unit((float)jobinfo->max_procs,
- max_procs_char, UNIT_NONE);
+ max_procs_char, sizeof(max_procs_char),
+ UNIT_NONE);
snprintf(buf, size, "%s", max_procs_char);
break;
diff --git a/src/common/parse_time.c b/src/common/parse_time.c
index e70fe5f29a9cdc99cf80018f28ae256806b8a6b5..ea7882fdf06af30e856994359c9ff92afe4894a3 100644
--- a/src/common/parse_time.c
+++ b/src/common/parse_time.c
@@ -117,6 +117,10 @@ _get_time(char *time_str, int *pos, int *hour, int *minute, int * second)
if ((time_str[offset] < '0') || (time_str[offset] > '9'))
goto prob;
hr = (hr * 10) + time_str[offset++] - '0';
+ if (hr > 23) {
+ offset -= 2;
+ goto prob;
+ }
if (time_str[offset] != ':')
goto prob;
offset++;
@@ -128,6 +132,10 @@ _get_time(char *time_str, int *pos, int *hour, int *minute, int * second)
if ((time_str[offset] < '0') || (time_str[offset] > '9'))
goto prob;
min = (min * 10) + time_str[offset++] - '0';
+ if (min > 59) {
+ offset -= 2;
+ goto prob;
+ }
/* get optional second */
if (time_str[offset] == ':') {
@@ -138,6 +146,10 @@ _get_time(char *time_str, int *pos, int *hour, int *minute, int * second)
if ((time_str[offset] < '0') || (time_str[offset] > '9'))
goto prob;
sec = (sec * 10) + time_str[offset++] - '0';
+ if (sec > 59) {
+ offset -= 2;
+ goto prob;
+ }
} else
sec = 0;
@@ -146,8 +158,20 @@ _get_time(char *time_str, int *pos, int *hour, int *minute, int * second)
}
if (strncasecmp(time_str+offset, "pm", 2)== 0) {
hr += 12;
+ if (hr > 23) {
+ if (hr == 24)
+ hr = 12;
+ else
+ goto prob;
+ }
offset += 2;
} else if (strncasecmp(time_str+offset, "am", 2) == 0) {
+ if (hr > 11) {
+ if (hr == 12)
+ hr = 0;
+ else
+ goto prob;
+ }
offset += 2;
}
diff --git a/src/common/slurm_protocol_api.c b/src/common/slurm_protocol_api.c
index 2da16712c941a54cf2d47c6181d8d27faa7f2031..9d1da7691417de41310a076e5f1eb2d10842caf9 100644
--- a/src/common/slurm_protocol_api.c
+++ b/src/common/slurm_protocol_api.c
@@ -2282,13 +2282,13 @@ extern int nodelist_find(const char *nodelist, const char *name)
return id;
}
-extern void convert_num_unit(float num, char *buf, int orig_type)
+extern void convert_num_unit(float num, char *buf, int buf_size, int orig_type)
{
char *unit = "\0KMGP?";
int i = (int)num % 512;
if(i > 0 || (int)num == 0) {
- sprintf(buf, "%d%c", (int)num, unit[orig_type]);
+ snprintf(buf, buf_size, "%d%c", (int)num, unit[orig_type]);
return;
}
@@ -2301,9 +2301,9 @@ extern void convert_num_unit(float num, char *buf, int orig_type)
orig_type = UNIT_UNKNOWN;
i = (int)num;
if(i == num)
- sprintf(buf, "%d%c", i, unit[orig_type]);
+ snprintf(buf, buf_size, "%d%c", i, unit[orig_type]);
else
- sprintf(buf, "%.2f%c", num, unit[orig_type]);
+ snprintf(buf, buf_size, "%.2f%c", num, unit[orig_type]);
}
extern int revert_num_unit(const char *buf)
diff --git a/src/common/slurm_protocol_api.h b/src/common/slurm_protocol_api.h
index 2efb4854c42a40ecab2fb866538bc12c73ab0a6a..bd5614c56b4f7efd8d8edcc009b3c1710be746fd 100644
--- a/src/common/slurm_protocol_api.h
+++ b/src/common/slurm_protocol_api.h
@@ -716,7 +716,7 @@ extern void slurm_free_msg(slurm_msg_t * msg);
/* must free this memory with free not xfree */
extern char *nodelist_nth_host(const char *nodelist, int inx);
extern int nodelist_find(const char *nodelist, const char *name);
-extern void convert_num_unit(float num, char *buf, int orig_type);
+extern void convert_num_unit(float num, char *buf, int buf_size, int orig_type);
extern int revert_num_unit(const char *buf);
/*
diff --git a/src/plugins/sched/wiki/Makefile.am b/src/plugins/sched/wiki/Makefile.am
index b164be29e2395cd93184ee34347551f56f71777c..ba054203a77b030fadf5a9e26c88fe68e9cfa78a 100644
--- a/src/plugins/sched/wiki/Makefile.am
+++ b/src/plugins/sched/wiki/Makefile.am
@@ -16,8 +16,12 @@ sched_wiki_la_SOURCES = \
crypto.h \
get_jobs.c \
get_nodes.c \
+ hostlist.c \
+ job_modify.c \
msg.c \
msg.h \
+ resume_job.c \
sched_wiki.c \
- start_job.c
+ start_job.c \
+ suspend_job.c
sched_wiki_la_LDFLAGS = $(SO_LDFLAGS) $(PLUGIN_FLAGS)
diff --git a/src/plugins/sched/wiki/Makefile.in b/src/plugins/sched/wiki/Makefile.in
index 55320a07219392c7b898954b84fd49e21d7c3f9e..17ec4ae8413bfebc5baf1db7ba7e934685ea2c7f 100644
--- a/src/plugins/sched/wiki/Makefile.in
+++ b/src/plugins/sched/wiki/Makefile.in
@@ -75,7 +75,8 @@ pkglibLTLIBRARIES_INSTALL = $(INSTALL)
LTLIBRARIES = $(pkglib_LTLIBRARIES)
sched_wiki_la_LIBADD =
am_sched_wiki_la_OBJECTS = cancel_job.lo crypto.lo get_jobs.lo \
- get_nodes.lo msg.lo sched_wiki.lo start_job.lo
+ get_nodes.lo hostlist.lo job_modify.lo msg.lo resume_job.lo \
+ sched_wiki.lo start_job.lo suspend_job.lo
sched_wiki_la_OBJECTS = $(am_sched_wiki_la_OBJECTS)
sched_wiki_la_LINK = $(LIBTOOL) --tag=CC $(AM_LIBTOOLFLAGS) \
$(LIBTOOLFLAGS) --mode=link $(CCLD) $(AM_CFLAGS) $(CFLAGS) \
@@ -263,10 +264,14 @@ sched_wiki_la_SOURCES = \
crypto.h \
get_jobs.c \
get_nodes.c \
+ hostlist.c \
+ job_modify.c \
msg.c \
msg.h \
+ resume_job.c \
sched_wiki.c \
- start_job.c
+ start_job.c \
+ suspend_job.c
sched_wiki_la_LDFLAGS = $(SO_LDFLAGS) $(PLUGIN_FLAGS)
all: all-am
@@ -342,9 +347,13 @@ distclean-compile:
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/crypto.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/get_jobs.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/get_nodes.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/hostlist.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/job_modify.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/msg.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/resume_job.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/sched_wiki.Plo@am__quote@
@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/start_job.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/suspend_job.Plo@am__quote@
.c.o:
@am__fastdepCC_TRUE@ $(COMPILE) -MT $@ -MD -MP -MF $(DEPDIR)/$*.Tpo -c -o $@ $<
diff --git a/src/plugins/sched/wiki/get_jobs.c b/src/plugins/sched/wiki/get_jobs.c
index df478987d4c531109eea445934cad7ab26d293e0..e021e05a847044ba2b4dc4e0705421af9ed85a68 100644
--- a/src/plugins/sched/wiki/get_jobs.c
+++ b/src/plugins/sched/wiki/get_jobs.c
@@ -39,6 +39,7 @@
#include <sys/types.h>
#include "./msg.h"
+#include "src/common/hostlist.h"
#include "src/common/list.h"
#include "src/common/uid.h"
#include "src/slurmctld/locks.h"
@@ -47,6 +48,7 @@
static char * _dump_all_jobs(int *job_cnt, int state_info);
static char * _dump_job(struct job_record *job_ptr, int state_info);
static char * _get_group_name(gid_t gid);
+static uint16_t _get_job_cpus_per_task(struct job_record *job_ptr);
static uint32_t _get_job_end_time(struct job_record *job_ptr);
static uint32_t _get_job_min_disk(struct job_record *job_ptr);
static uint32_t _get_job_min_mem(struct job_record *job_ptr);
@@ -56,6 +58,8 @@ static uint32_t _get_job_submit_time(struct job_record *job_ptr);
static uint32_t _get_job_suspend_time(struct job_record *job_ptr);
static uint32_t _get_job_tasks(struct job_record *job_ptr);
static uint32_t _get_job_time_limit(struct job_record *job_ptr);
+static char * _task_list(struct job_record *job_ptr);
+
#define SLURM_INFO_ALL 0
#define SLURM_INFO_VOLITILE 1
@@ -209,8 +213,7 @@ static char * _dump_job(struct job_record *job_ptr, int state_info)
xstrcat(buf, tmp);
xfree(hosts);
} else if (!IS_JOB_FINISHED(job_ptr)) {
- char *hosts = bitmap2wiki_node_name(
- job_ptr->node_bitmap);
+ char *hosts = _task_list(job_ptr);
snprintf(tmp, sizeof(tmp),
"TASKLIST=%s;", hosts);
xstrcat(buf, tmp);
@@ -230,15 +233,18 @@ static char * _dump_job(struct job_record *job_ptr, int state_info)
(uint32_t) _get_job_time_limit(job_ptr));
xstrcat(buf, tmp);
- if (job_ptr->job_state == JOB_PENDING) {
- /* Don't report actual tasks or nodes allocated since
- * this can impact requeue on heterogenous clusters */
- snprintf(tmp, sizeof(tmp),
- "TASKS=%u;NODES=%u;",
- _get_job_tasks(job_ptr),
- _get_job_min_nodes(job_ptr));
- xstrcat(buf, tmp);
- }
+ /* Don't report actual tasks or nodes allocated since
+ * this can impact requeue on heterogenous clusters */
+ snprintf(tmp, sizeof(tmp),
+ "TASKS=%u;NODES=%u;",
+ _get_job_tasks(job_ptr),
+ _get_job_min_nodes(job_ptr));
+ xstrcat(buf, tmp);
+
+ snprintf(tmp, sizeof(tmp),
+ "DPROCS=%u;",
+ _get_job_cpus_per_task(job_ptr));
+ xstrcat(buf, tmp);
snprintf(tmp, sizeof(tmp),
"QUEUETIME=%u;STARTTIME=%u;PARTITIONMASK=%s;",
@@ -267,6 +273,18 @@ static char * _dump_job(struct job_record *job_ptr, int state_info)
xstrcat(buf, tmp);
}
+ if (job_ptr->account) {
+ snprintf(tmp, sizeof(tmp),
+ "ACCOUNT=%s;", job_ptr->account);
+ xstrcat(buf, tmp);
+ }
+
+ if (job_ptr->comment && job_ptr->comment[0]) {
+ snprintf(tmp,sizeof(tmp),
+ "COMMENT=%s;", job_ptr->comment);
+ xstrcat(buf,tmp);
+ }
+
if (state_info == SLURM_INFO_VOLITILE)
return buf;
@@ -280,6 +298,15 @@ static char * _dump_job(struct job_record *job_ptr, int state_info)
return buf;
}
+static uint16_t _get_job_cpus_per_task(struct job_record *job_ptr)
+{
+ uint16_t cpus_per_task = 1;
+
+ if (job_ptr->details && job_ptr->details->cpus_per_task)
+ cpus_per_task = job_ptr->details->cpus_per_task;
+ return cpus_per_task;
+}
+
static uint32_t _get_job_min_mem(struct job_record *job_ptr)
{
if (job_ptr->details)
@@ -337,7 +364,7 @@ static uint32_t _get_job_tasks(struct job_record *job_ptr)
job_ptr->details->ntasks_per_node));
}
- return task_cnt;
+ return task_cnt / _get_job_cpus_per_task(job_ptr);
}
static uint32_t _get_job_time_limit(struct job_record *job_ptr)
@@ -427,3 +454,37 @@ extern char * bitmap2wiki_node_name(bitstr_t *bitmap)
}
return buf;
}
+
+
+/* Return task list in Maui format: tux0:tux0:tux1:tux1:tux2 */
+static char * _task_list(struct job_record *job_ptr)
+{
+ int i, j, task_cnt;
+ char *buf = NULL, *host;
+ hostlist_t hl = hostlist_create(job_ptr->nodes);
+
+ buf = xstrdup("");
+ if (hl == NULL)
+ return buf;
+
+ for (i=0; i<job_ptr->alloc_lps_cnt; i++) {
+ host = hostlist_shift(hl);
+ if (host == NULL) {
+ error("bad alloc_lps_cnt for job %u (%s, %d)",
+ job_ptr->job_id, job_ptr->nodes,
+ job_ptr->alloc_lps_cnt);
+ break;
+ }
+ task_cnt = job_ptr->alloc_lps[i];
+ if (job_ptr->details && job_ptr->details->cpus_per_task)
+ task_cnt /= job_ptr->details->cpus_per_task;
+ for (j=0; j<task_cnt; j++) {
+ if (buf)
+ xstrcat(buf, ":");
+ xstrcat(buf, host);
+ }
+ free(host);
+ }
+ hostlist_destroy(hl);
+ return buf;
+}
diff --git a/src/plugins/sched/wiki/hostlist.c b/src/plugins/sched/wiki/hostlist.c
new file mode 100644
index 0000000000000000000000000000000000000000..c31dc6bf89207b154d55f354e6f062353055becc
--- /dev/null
+++ b/src/plugins/sched/wiki/hostlist.c
@@ -0,0 +1,291 @@
+/*****************************************************************************\
+ * hostlist.c - Convert hostlist expressions between Slurm and Moab formats
+ *****************************************************************************
+ * Copyright (C) 2007 The Regents of the University of California.
+ * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ * Written by Morris Jette <jette1@llnl.gov>
+ * UCRL-CODE-226842.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * In addition, as a special exception, the copyright holders give permission
+ * to link the code of portions of this program with the OpenSSL library under
+ * certain conditions as described in each individual source file, and
+ * distribute linked combinations including the two. You must obey the GNU
+ * General Public License in all respects for all of the code used other than
+ * OpenSSL. If you modify file(s) with this exception, you may extend this
+ * exception to your version of the file(s), but you are not obligated to do
+ * so. If you do not wish to do so, delete this exception statement from your
+ * version. If you delete this exception statement from all source files in
+ * the program, then also delete it here.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+\*****************************************************************************/
+
+#if HAVE_CONFIG_H
+# include "config.h"
+# if HAVE_INTTYPES_H
+# include <inttypes.h>
+# else
+# if HAVE_STDINT_H
+# include <stdint.h>
+# endif
+# endif /* HAVE_INTTYPES_H */
+#else /* !HAVE_CONFIG_H */
+# include <inttypes.h>
+#endif /* HAVE_CONFIG_H */
+
+#include <stdlib.h>
+#include <string.h>
+
+#include "./msg.h"
+#include "src/common/hostlist.h"
+#include "src/common/node_select.h"
+#include "src/common/xmalloc.h"
+#include "src/common/xstring.h"
+
+static void _append_hl_buf(char **buf, hostlist_t *hl_tmp, int *reps);
+static char * _task_list(struct job_record *job_ptr);
+static char * _task_list_exp(struct job_record *job_ptr);
+
+/*
+ * Convert Moab supplied TASKLIST expression into a SLURM hostlist expression
+ *
+ * Moab format 1: tux0:tux0:tux1:tux1:tux2 (list host for each cpu)
+ * Moab format 2: tux[0-1]*2:tux2 (list cpu count after host name)
+ *
+ * SLURM format: tux0,tux0,tux1,tux1,tux2 (if consumable resources enabled)
+ * SLURM format: tux0,tux1,tux2 (if consumable resources disabled)
+ *
+ * NOTE: returned string must be released with xfree()
+ */
+extern char * moab2slurm_task_list(char *moab_tasklist, int *task_cnt)
+{
+ char *slurm_tasklist, *host, *tmp1, *tmp2, *tok, *tok_p;
+ int i, reps;
+ hostlist_t hl;
+ static uint32_t cr_test = 0, cr_enabled = 0;
+
+ if (cr_test == 0) {
+ select_g_get_info_from_plugin(SELECT_CR_PLUGIN,
+ &cr_enabled);
+ cr_test = 1;
+ }
+
+ *task_cnt = 0;
+
+ /* Moab format 2 if string contains '*' or '[' */
+ tmp1 = strchr(moab_tasklist, (int) '*');
+ if (tmp1 == NULL)
+ tmp1 = strchr(moab_tasklist, (int) '[');
+
+ if (tmp1 == NULL) { /* Moab format 1 */
+ slurm_tasklist = xstrdup(moab_tasklist);
+ if (moab_tasklist[0])
+ *task_cnt = 1;
+ for (i=0; slurm_tasklist[i]!='\0'; i++) {
+ if (slurm_tasklist[i] == ':') {
+ slurm_tasklist[i] = ',';
+ (*task_cnt)++;
+ } else if (slurm_tasklist[i] == ',')
+ (*task_cnt)++;
+ }
+ return slurm_tasklist;
+ }
+
+ /* Moab format 2 */
+ slurm_tasklist = xstrdup("");
+ tmp1 = xstrdup(moab_tasklist);
+ tok = strtok_r(tmp1, ":", &tok_p);
+ while (tok) {
+ /* find task count, assume 1 if no "*" */
+ tmp2 = strchr(tok, (int) '*');
+ if (tmp2) {
+ reps = atoi(tmp2 + 1);
+ tmp2[0] = '\0';
+ } else
+ reps = 1;
+
+ /* find host expression */
+ hl = hostlist_create(tok);
+ while ((host = hostlist_shift(hl))) {
+ for (i=0; i<reps; i++) {
+ if (slurm_tasklist[0])
+ xstrcat(slurm_tasklist, ",");
+ xstrcat(slurm_tasklist, host);
+ if (!cr_enabled)
+ break;
+ }
+ free(host);
+ (*task_cnt) += reps;
+ }
+ hostlist_destroy(hl);
+
+ /* get next token */
+ tok = strtok_r(NULL, ":", &tok_p);
+ }
+ xfree(tmp1);
+ return slurm_tasklist;
+}
+
+/*
+ * Report a job's tasks a a MOAB TASKLIST expression
+ *
+ * Moab format 1: tux0:tux0:tux1:tux1:tux2 (list host for each cpu)
+ * Moab format 2: tux[0-1]*2:tux2 (list cpu count after host name)
+ *
+ * NOTE: returned string must be released with xfree()
+ */
+extern char * slurm_job2moab_task_list(struct job_record *job_ptr)
+{
+ if (use_host_exp)
+ return _task_list_exp(job_ptr);
+ else
+ return _task_list(job_ptr);
+}
+
+/* Return task list in Moab format 1: tux0:tux0:tux1:tux1:tux2 */
+static char * _task_list(struct job_record *job_ptr)
+{
+ int i, j;
+ char *buf = NULL, *host;
+ hostlist_t hl = hostlist_create(job_ptr->nodes);
+
+ if (hl == NULL) {
+ error("hostlist_create error for job %u, %s",
+ job_ptr->job_id, job_ptr->nodes);
+ return buf;
+ }
+
+ for (i=0; i<job_ptr->alloc_lps_cnt; i++) {
+ host = hostlist_shift(hl);
+ if (host == NULL) {
+ error("bad alloc_lps_cnt for job %u (%s, %d)",
+ job_ptr->job_id, job_ptr->nodes,
+ job_ptr->alloc_lps_cnt);
+ break;
+ }
+ for (j=0; j<job_ptr->alloc_lps[i]; j++) {
+ if (buf)
+ xstrcat(buf, ":");
+ xstrcat(buf, host);
+ }
+ free(host);
+ }
+ hostlist_destroy(hl);
+ return buf;
+}
+
+/* Append to buf a compact tasklist expression (e.g. "tux[0-1]*2")
+ * Prepend ":" to expression as needed */
+static void _append_hl_buf(char **buf, hostlist_t *hl_tmp, int *reps)
+{
+ int host_str_len = 4096;
+ char *host_str, tmp_str[64];
+ char *tok, *sep;
+ int i, in_bracket = 0, fini = 0;
+
+ host_str = xmalloc(host_str_len);
+ hostlist_uniq(*hl_tmp);
+ while (hostlist_ranged_string(*hl_tmp, host_str_len, host_str) < 0) {
+ host_str_len *= 2;
+ xrealloc(*host_str, host_str_len);
+ }
+
+ /* Note that host_str may be of this form "alpha,beta". We want
+ * to record this as "alpha*#:beta*#" and NOT "alpha,beta*#".
+ * NOTE: Do not break up command within brackets (e.g. "tux[1,2-4]") */
+ if (*buf)
+ sep = ":";
+ else
+ sep = "";
+ tok = host_str;
+ for (i=0; (fini == 0) ; i++) {
+ switch (tok[i]) {
+ case '[':
+ in_bracket = 1;
+ break;
+ case ']':
+ in_bracket = 0;
+ break;
+ case '\0':
+ fini = 1;
+ if (in_bracket)
+ error("badly formed hostlist %s", tok);
+ case ',':
+ if (in_bracket)
+ break;
+ tok[i] = '\0';
+ snprintf(tmp_str, sizeof(tmp_str), "%s%s*%d",
+ sep, tok, *reps);
+ xstrcat(*buf, tmp_str);
+ sep = ":";
+ tok += (i + 1);
+ i = -1;
+ break;
+ }
+ }
+ xfree(host_str);
+ hostlist_destroy(*hl_tmp);
+ *hl_tmp = (hostlist_t) NULL;
+ *reps = 0;
+}
+
+/* Return task list in Moab format 2: tux[0-1]*2:tux2 */
+static char * _task_list_exp(struct job_record *job_ptr)
+{
+ int i, reps = -1;
+ char *buf = NULL, *host;
+ hostlist_t hl = hostlist_create(job_ptr->nodes);
+ hostlist_t hl_tmp = (hostlist_t) NULL;
+
+ if (hl == NULL) {
+ error("hostlist_create error for job %u, %s",
+ job_ptr->job_id, job_ptr->nodes);
+ return buf;
+ }
+
+ for (i=0; i<job_ptr->alloc_lps_cnt; i++) {
+ host = hostlist_shift(hl);
+ if (host == NULL) {
+ error("bad alloc_lps_cnt for job %u (%s, %d)",
+ job_ptr->job_id, job_ptr->nodes,
+ job_ptr->alloc_lps_cnt);
+ break;
+ }
+
+ if (reps == job_ptr->alloc_lps[i]) {
+ /* append to existing hostlist record */
+ if (hostlist_push(hl_tmp, host) == 0)
+ error("hostlist_push failure");
+ } else {
+ if (hl_tmp)
+ _append_hl_buf(&buf, &hl_tmp, &reps);
+
+ /* start new hostlist record */
+ hl_tmp = hostlist_create(host);
+ if (hl_tmp)
+ reps = job_ptr->alloc_lps[i];
+ else
+ error("hostlist_create failure");
+ }
+ free(host);
+ }
+ hostlist_destroy(hl);
+ if (hl_tmp)
+ _append_hl_buf(&buf, &hl_tmp, &reps);
+ return buf;
+}
diff --git a/src/plugins/sched/wiki/job_modify.c b/src/plugins/sched/wiki/job_modify.c
new file mode 100644
index 0000000000000000000000000000000000000000..e3a62654951e8e3241d248157ca37ac629aa2c60
--- /dev/null
+++ b/src/plugins/sched/wiki/job_modify.c
@@ -0,0 +1,303 @@
+/*****************************************************************************\
+ * job_modify.c - Process Wiki job modify request
+ *****************************************************************************
+ * Copyright (C) 2006-2007 The Regents of the University of California.
+ * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ * Written by Morris Jette <jette1@llnl.gov>
+ * UCRL-CODE-226842.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * In addition, as a special exception, the copyright holders give permission
+ * to link the code of portions of this program with the OpenSSL library under
+ * certain conditions as described in each individual source file, and
+ * distribute linked combinations including the two. You must obey the GNU
+ * General Public License in all respects for all of the code used other than
+ * OpenSSL. If you modify file(s) with this exception, you may extend this
+ * exception to your version of the file(s), but you are not obligated to do
+ * so. If you do not wish to do so, delete this exception statement from your
+ * version. If you delete this exception statement from all source files in
+ * the program, then also delete it here.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+\*****************************************************************************/
+
+#include "./msg.h"
+#include <strings.h>
+#include "src/slurmctld/locks.h"
+#include "src/slurmctld/slurmctld.h"
+
+static void _null_term(char *str)
+{
+ char *tmp_ptr;
+ for (tmp_ptr=str; ; tmp_ptr++) {
+ if (tmp_ptr[0] == '\0')
+ break;
+ if (isspace(tmp_ptr[0])) {
+ tmp_ptr[0] = '\0';
+ break;
+ }
+ }
+}
+
+/* return -1 on error */
+static int32_t _get_depend_id(char *str)
+{
+ /* stand-alone job_id */
+ if (isdigit(str[0]))
+ return (int32_t) atol(str);
+
+ if (strncasecmp(str, "afterany:", 9) != 0) /* invalid spec */
+ return (int32_t) -1;
+
+ str += 9;
+ if (!isdigit(str[0]))
+ return (int32_t) -1;
+ return (int32_t) atol(str);
+}
+
+static int _job_modify(uint32_t jobid, char *bank_ptr,
+ int32_t depend_id, char *new_hostlist,
+ uint32_t new_node_cnt, char *part_name_ptr,
+ uint32_t new_time_limit)
+{
+ struct job_record *job_ptr;
+
+ job_ptr = find_job_record(jobid);
+ if (job_ptr == NULL) {
+ error("wiki: MODIFYJOB has invalid jobid %u", jobid);
+ return ESLURM_INVALID_JOB_ID;
+ }
+ if (IS_JOB_FINISHED(job_ptr)) {
+ error("wiki: MODIFYJOB jobid %u is finished", jobid);
+ return ESLURM_DISABLED;
+ }
+
+ if (depend_id != -1) {
+ info("wiki: changing job dependency to %d", depend_id);
+ job_ptr->dependency = depend_id;
+ }
+
+ if (new_time_limit) {
+ time_t old_time = job_ptr->time_limit;
+ job_ptr->time_limit = new_time_limit;
+ info("wiki: change job %u time_limit to %u",
+ jobid, new_time_limit);
+ /* Update end_time based upon change
+ * to preserve suspend time info */
+ job_ptr->end_time = job_ptr->end_time +
+ ((job_ptr->time_limit -
+ old_time) * 60);
+ last_job_update = time(NULL);
+ }
+ if (bank_ptr) {
+ info("wiki: change job %u bank %s", jobid, bank_ptr);
+ xfree(job_ptr->account);
+ job_ptr->account = xstrdup(bank_ptr);
+ last_job_update = time(NULL);
+ }
+
+ if (new_hostlist) {
+ int i, rc = 0, task_cnt;
+ hostlist_t hl;
+ char *tasklist;
+
+ if (!job_ptr->details) {
+ /* Job is done, nothing to reset */
+ if (new_hostlist == '\0')
+ goto host_fini;
+ error("wiki: MODIFYJOB tasklist of non-pending "
+ "job %u", jobid);
+ return ESLURM_DISABLED;
+ }
+
+ xfree(job_ptr->details->req_nodes);
+ FREE_NULL_BITMAP(job_ptr->details->req_node_bitmap);
+ if (new_hostlist == '\0')
+ goto host_fini;
+
+ tasklist = moab2slurm_task_list(new_hostlist, &task_cnt);
+ if (tasklist == NULL) {
+ rc = 1;
+ goto host_fini;
+ }
+ hl = hostlist_create(tasklist);
+ if (hl == 0) {
+ rc = 1;
+ goto host_fini;
+ }
+ hostlist_uniq(hl);
+ hostlist_sort(hl);
+ i = strlen(new_hostlist) + 16;
+ job_ptr->details->req_nodes = xmalloc(i);
+ i = hostlist_ranged_string(hl, i, job_ptr->details->req_nodes);
+ hostlist_destroy(hl);
+ if (i < 0) {
+ rc = 1;
+ goto host_fini;
+ }
+ if (node_name2bitmap(job_ptr->details->req_nodes, false,
+ &job_ptr->details->req_node_bitmap)) {
+ rc = 1;
+ goto host_fini;
+ }
+
+host_fini: if (rc) {
+ info("wiki: change job %u invalid hostlist %s", jobid, new_hostlist);
+ xfree(job_ptr->details->req_nodes);
+ return EINVAL;
+ } else {
+ info("wiki: change job %u hostlist %s", jobid, new_hostlist);
+ }
+ }
+
+ if (part_name_ptr) {
+ struct part_record *part_ptr;
+ part_ptr = find_part_record(part_name_ptr);
+ if (part_ptr == NULL) {
+ error("wiki: MODIFYJOB has invalid partition %s",
+ part_name_ptr);
+ return ESLURM_INVALID_PARTITION_NAME;
+ }
+ info("wiki: change job %u partition %s",
+ jobid, part_name_ptr);
+ strncpy(job_ptr->partition, part_name_ptr, MAX_SLURM_NAME);
+ job_ptr->part_ptr = part_ptr;
+ last_job_update = time(NULL);
+ }
+
+ if (new_node_cnt) {
+ if (IS_JOB_PENDING(job_ptr) && job_ptr->details) {
+ job_ptr->details->min_nodes = new_node_cnt;
+ if (job_ptr->details->max_nodes
+ && (job_ptr->details->max_nodes < new_node_cnt))
+ job_ptr->details->max_nodes = new_node_cnt;
+ info("wiki: change job %u min_nodes to %u",
+ jobid, new_node_cnt);
+ last_job_update = time(NULL);
+ } else {
+ error("wiki: MODIFYJOB node count of non-pending "
+ "job %u", jobid);
+ return ESLURM_DISABLED;
+ }
+ }
+
+ return SLURM_SUCCESS;
+}
+
+/* Modify a job:
+ * CMD=MODIFYJOB ARG=<jobid> PARTITION=<name> NODES=<number>
+ * DEPEND=afterany:<jobid> TIMELIMT=<seconds> BANK=<name>
+ * RET 0 on success, -1 on failure */
+extern int job_modify_wiki(char *cmd_ptr, int *err_code, char **err_msg)
+{
+ char *arg_ptr, *bank_ptr, *depend_ptr, *nodes_ptr;
+ char *host_ptr, *part_ptr, *time_ptr, *tmp_char;
+ int slurm_rc;
+ int depend_id = -1;
+ uint32_t jobid, new_node_cnt = 0, new_time_limit = 0;
+ static char reply_msg[128];
+ /* Locks: write job, read node and partition info */
+ slurmctld_lock_t job_write_lock = {
+ NO_LOCK, WRITE_LOCK, READ_LOCK, READ_LOCK };
+
+ arg_ptr = strstr(cmd_ptr, "ARG=");
+ if (arg_ptr == NULL) {
+ *err_code = -300;
+ *err_msg = "MODIFYJOB lacks ARG=";
+ error("wiki: MODIFYJOB lacks ARG=");
+ return -1;
+ }
+ /* Change all parsed "=" to ":" then search for remaining "="
+ * and report results as unrecognized options */
+ arg_ptr[3] = ':';
+ arg_ptr += 4;
+ jobid = strtoul(arg_ptr, &tmp_char, 10);
+ if ((tmp_char[0] != '\0') && (!isspace(tmp_char[0]))) {
+ *err_code = -300;
+ *err_msg = "Invalid ARG value";
+ error("wiki: MODIFYJOB has invalid jobid");
+ return -1;
+ }
+ bank_ptr = strstr(cmd_ptr, "BANK=");
+ depend_ptr = strstr(cmd_ptr, "DEPEND=");
+ host_ptr = strstr(cmd_ptr, "HOSTLIST=");
+ nodes_ptr = strstr(cmd_ptr, "NODES=");
+ part_ptr = strstr(cmd_ptr, "PARTITION=");
+ time_ptr = strstr(cmd_ptr, "TIMELIMIT=");
+ if (bank_ptr) {
+ bank_ptr[4] = ':';
+ bank_ptr += 5;
+ _null_term(bank_ptr);
+ }
+ if (depend_ptr) {
+ depend_ptr[6] = ':';
+ depend_ptr += 7;
+ depend_id = _get_depend_id(depend_ptr);
+ if (depend_id == -1) {
+ *err_code = -300;
+ *err_msg = "MODIFYJOB has invalid DEPEND specificiation";
+ error("wiki: MODIFYJOB has invalid DEPEND spec: %s",
+ depend_ptr);
+ return -1;
+ }
+ }
+ if (host_ptr) {
+ host_ptr[8] = ':';
+ host_ptr += 9;
+ _null_term(bank_ptr);
+ }
+ if (nodes_ptr) {
+ nodes_ptr[5] = ':';
+ nodes_ptr += 6;
+ new_node_cnt = strtoul(nodes_ptr, NULL, 10);
+ }
+ if (part_ptr) {
+ part_ptr[9] = ':';
+ part_ptr += 10;
+ _null_term(part_ptr);
+ }
+ if (time_ptr) {
+ time_ptr[9] = ':';
+ time_ptr += 10;
+ new_time_limit = strtoul(time_ptr, NULL, 10);
+ }
+
+ /* Look for any un-parsed "=" */
+ tmp_char = strchr(cmd_ptr, '=');
+ if (tmp_char) {
+ tmp_char[0] = '\0';
+ while (tmp_char[-1] && (!isspace(tmp_char[-1])))
+ tmp_char--;
+ error("wiki: Invalid MODIFYJOB option %s", tmp_char);
+ }
+
+ lock_slurmctld(job_write_lock);
+ slurm_rc = _job_modify(jobid, bank_ptr, depend_id, host_ptr,
+ new_node_cnt, part_ptr, new_time_limit);
+ unlock_slurmctld(job_write_lock);
+ if (slurm_rc != SLURM_SUCCESS) {
+ *err_code = -700;
+ *err_msg = slurm_strerror(slurm_rc);
+ error("wiki: Failed to modify job %u (%m)", jobid);
+ return -1;
+ }
+
+ snprintf(reply_msg, sizeof(reply_msg),
+ "job %u modified successfully", jobid);
+ *err_msg = reply_msg;
+ return 0;
+}
diff --git a/src/plugins/sched/wiki/msg.c b/src/plugins/sched/wiki/msg.c
index 0f9b9e9b303a5b348d20bdfe30c86a1fb01f5acc..20cedd7e3ee3f0ef00e95f35d3f5056d79be02e4 100644
--- a/src/plugins/sched/wiki/msg.c
+++ b/src/plugins/sched/wiki/msg.c
@@ -505,12 +505,14 @@ static int _parse_msg(char *msg, char **req)
\*****************************************************************************/
static void _proc_msg(slurm_fd new_fd, char *msg)
{
- char *req, *cmd_ptr;
+ DEF_TIMERS;
+ char *req, *cmd_ptr, *msg_type = NULL;
char response[128];
if (new_fd < 0)
return;
+ START_TIMER;
if (!msg) {
err_code = -300;
err_msg = "NULL request message";
@@ -531,20 +533,34 @@ static void _proc_msg(slurm_fd new_fd, char *msg)
cmd_ptr +=4;
err_code = 0;
if (strncmp(cmd_ptr, "GETJOBS", 7) == 0) {
+ msg_type = "wiki:GETJOBS";
if (!get_jobs(cmd_ptr, &err_code, &err_msg))
goto free_resp_msg;
} else if (strncmp(cmd_ptr, "GETNODES", 8) == 0) {
+ msg_type = "wiki:GETNODES";
if (!get_nodes(cmd_ptr, &err_code, &err_msg))
goto free_resp_msg;
} else if (strncmp(cmd_ptr, "STARTJOB", 8) == 0) {
+ msg_type = "wiki:STARTJOB";
start_job(cmd_ptr, &err_code, &err_msg);
} else if (strncmp(cmd_ptr, "CANCELJOB", 9) == 0) {
+ msg_type = "wiki:CANCELJOB";
cancel_job(cmd_ptr, &err_code, &err_msg);
+ } else if (strncmp(cmd_ptr, "SUSPENDJOB", 10) == 0) {
+ msg_type = "wiki:SUSPENDJOB";
+ suspend_job(cmd_ptr, &err_code, &err_msg);
+ } else if (strncmp(cmd_ptr, "RESUMEJOB", 9) == 0) {
+ msg_type = "wiki:RESUMEJOB";
+ resume_job(cmd_ptr, &err_code, &err_msg);
+ } else if (strncmp(cmd_ptr, "MODIFYJOB", 9) == 0) {
+ msg_type = "wiki:MODIFYJOB";
+ job_modify_wiki(cmd_ptr, &err_code, &err_msg);
} else {
err_code = -300;
err_msg = "unsupported request type";
error("wiki: unrecognized request type: %s", req);
}
+ END_TIMER2(msg_type);
resp_msg:
snprintf(response, sizeof(response),
diff --git a/src/plugins/sched/wiki/msg.h b/src/plugins/sched/wiki/msg.h
index 76458521a45c5bac36e9b5eb5a8a03fb6946eaa4..e19e45b1f08e449f47100c7101465e89fc5f65a4 100644
--- a/src/plugins/sched/wiki/msg.h
+++ b/src/plugins/sched/wiki/msg.h
@@ -114,6 +114,7 @@ extern int job_release_task(char *cmd_ptr, int *err_code, char **err_msg);
extern int job_requeue_wiki(char *cmd_ptr, int *err_code, char **err_msg);
extern int job_signal_wiki(char *cmd_ptr, int *err_code, char **err_msg);
extern int job_will_run(char *cmd_ptr, int *err_code, char **err_msg);
+extern char * moab2slurm_task_list(char *moab_tasklist, int *task_cnt);
extern int parse_wiki_config(void);
extern int start_job(char *cmd_ptr, int *err_code, char **err_msg);
extern int suspend_job(char *cmd_ptr, int *err_code, char **err_msg);
diff --git a/src/plugins/sched/wiki/resume_job.c b/src/plugins/sched/wiki/resume_job.c
new file mode 100644
index 0000000000000000000000000000000000000000..319219244b55f6128c64c10797901e298a3b2f0a
--- /dev/null
+++ b/src/plugins/sched/wiki/resume_job.c
@@ -0,0 +1,85 @@
+/*****************************************************************************\
+ * resume_job.c - Process Wiki resume job request
+ *****************************************************************************
+ * Copyright (C) 2006 The Regents of the University of California.
+ * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ * Written by Morris Jette <jette1@llnl.gov>
+ * UCRL-CODE-226842.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * In addition, as a special exception, the copyright holders give permission
+ * to link the code of portions of this program with the OpenSSL library under
+ * certain conditions as described in each individual source file, and
+ * distribute linked combinations including the two. You must obey the GNU
+ * General Public License in all respects for all of the code used other than
+ * OpenSSL. If you modify file(s) with this exception, you may extend this
+ * exception to your version of the file(s), but you are not obligated to do
+ * so. If you do not wish to do so, delete this exception statement from your
+ * version. If you delete this exception statement from all source files in
+ * the program, then also delete it here.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+\*****************************************************************************/
+
+#include "./msg.h"
+#include "src/slurmctld/locks.h"
+#include "src/slurmctld/slurmctld.h"
+
+/* RET 0 on success, -1 on failure */
+extern int resume_job(char *cmd_ptr, int *err_code, char **err_msg)
+{
+ char *arg_ptr, *tmp_char;
+ int slurm_rc;
+ suspend_msg_t msg;
+ uint32_t jobid;
+ static char reply_msg[128];
+ /* Locks: write job and node info */
+ slurmctld_lock_t job_write_lock = {
+ NO_LOCK, WRITE_LOCK, WRITE_LOCK, NO_LOCK };
+
+ arg_ptr = strstr(cmd_ptr, "ARG=");
+ if (arg_ptr == NULL) {
+ *err_code = -300;
+ *err_msg = "RESUMEJOB lacks ARG";
+ error("wiki: RESUMEJOB lacks ARG");
+ return -1;
+ }
+ jobid = strtoul(arg_ptr+4, &tmp_char, 10);
+ if ((tmp_char[0] != '\0') && (!isspace(tmp_char[0]))) {
+ *err_code = -300;
+ *err_msg = "Invalid ARG value";
+ error("wiki: RESUMEJOB has invalid jobid");
+ return -1;
+ }
+
+ msg.job_id = jobid;
+ msg.op = RESUME_JOB;
+ lock_slurmctld(job_write_lock);
+ slurm_rc = job_suspend(&msg, 0, -1);
+ unlock_slurmctld(job_write_lock);
+ if (slurm_rc != SLURM_SUCCESS) {
+ *err_code = -700;
+ *err_msg = slurm_strerror(slurm_rc);
+ error("wiki: Failed to resume job %u (%m)", jobid);
+ return -1;
+ }
+
+ snprintf(reply_msg, sizeof(reply_msg),
+ "job %u resumed successfully", jobid);
+ *err_msg = reply_msg;
+ return 0;
+}
diff --git a/src/plugins/sched/wiki/start_job.c b/src/plugins/sched/wiki/start_job.c
index b581aef83502cda104be4331570ef8e03865e5f4..f7cdd46d189ad8b080c780beb3ac794e7a9ca0ee 100644
--- a/src/plugins/sched/wiki/start_job.c
+++ b/src/plugins/sched/wiki/start_job.c
@@ -36,21 +36,23 @@
\*****************************************************************************/
#include "./msg.h"
+#include "src/common/node_select.h"
+#include "src/common/slurm_protocol_defs.h"
#include "src/common/xstring.h"
#include "src/slurmctld/locks.h"
#include "src/slurmctld/slurmctld.h"
#include "src/slurmctld/state_save.h"
-static int _start_job(uint32_t jobid, char *hostlist,
- int *err_code, char **err_msg);
+static int _start_job(uint32_t jobid, int task_cnt, char *hostlist,
+ char *tasklist, int *err_code, char **err_msg);
/* RET 0 on success, -1 on failure */
extern int start_job(char *cmd_ptr, int *err_code, char **err_msg)
{
- char *arg_ptr, *task_ptr, *node_ptr, *tmp_char;
- int i;
+ char *arg_ptr, *task_ptr, *tasklist, *tmp_char;
+ int i, rc, task_cnt;
uint32_t jobid;
- hostlist_t hl;
+ hostlist_t hl = (hostlist_t) NULL;
char host_string[MAXHOSTRANGELEN];
static char reply_msg[128];
@@ -76,20 +78,19 @@ extern int start_job(char *cmd_ptr, int *err_code, char **err_msg)
error("wiki: STARTJOB lacks TASKLIST");
return -1;
}
- node_ptr = task_ptr + 9;
- for (i=0; node_ptr[i]!='\0'; i++) {
- if (node_ptr[i] == ':')
- node_ptr[i] = ',';
- }
- hl = hostlist_create(node_ptr);
- if (hl == NULL) {
+ task_ptr += 9; /* skip over "TASKLIST=" */
+ tasklist = moab2slurm_task_list(task_ptr, &task_cnt);
+ if (tasklist)
+ hl = hostlist_create(tasklist);
+ if ((tasklist == NULL) || (hl == NULL)) {
*err_code = -300;
*err_msg = "STARTJOB TASKLIST is invalid";
error("wiki: STARTJOB TASKLIST is invalid: %s",
- node_ptr);
+ task_ptr);
+ xfree(tasklist);
return -1;
}
- hostlist_uniq(hl); /* for now, don't worry about task layout */
+ hostlist_uniq(hl);
hostlist_sort(hl);
i = hostlist_ranged_string(hl, sizeof(host_string), host_string);
hostlist_destroy(hl);
@@ -98,28 +99,54 @@ extern int start_job(char *cmd_ptr, int *err_code, char **err_msg)
*err_msg = "STARTJOB has invalid TASKLIST";
error("wiki: STARTJOB has invalid TASKLIST: %s",
host_string);
+ xfree(tasklist);
return -1;
}
- if (_start_job(jobid, host_string, err_code, err_msg) != 0)
- return -1;
- snprintf(reply_msg, sizeof(reply_msg),
- "job %u started successfully", jobid);
- *err_msg = reply_msg;
- return 0;
+ rc = _start_job(jobid, task_cnt, host_string, tasklist,
+ err_code, err_msg);
+ xfree(tasklist);
+ if (rc == 0) {
+ snprintf(reply_msg, sizeof(reply_msg),
+ "job %u started successfully", jobid);
+ *err_msg = reply_msg;
+ }
+ return rc;
}
-static int _start_job(uint32_t jobid, char *hostlist,
- int *err_code, char **err_msg)
+/*
+ * Attempt to start a job
+ * jobid (IN) - job id
+ * task_cnt (IN) - total count of tasks to start
+ * hostlist (IN) - SLURM hostlist expression with no repeated hostnames
+ * tasklist (IN/OUT) - comma separated list of hosts with tasks to be started,
+ * list hostname once per task to start
+ * err_code (OUT) - Moab error code
+ * err_msg (OUT) - Moab error message
+ */
+static int _start_job(uint32_t jobid, int task_cnt, char *hostlist,
+ char *tasklist, int *err_code, char **err_msg)
{
- int rc = 0;
+ int rc = 0, old_task_cnt = 1;
struct job_record *job_ptr;
/* Write lock on job info, read lock on node info */
slurmctld_lock_t job_write_lock = {
NO_LOCK, WRITE_LOCK, READ_LOCK, NO_LOCK };
- char *new_node_list, *save_req_nodes = NULL;
+ char *new_node_list = NULL;
static char tmp_msg[128];
- bitstr_t *new_bitmap, *save_req_bitmap = (bitstr_t *) NULL;
+ bitstr_t *new_bitmap = (bitstr_t *) NULL;
+ bitstr_t *save_req_bitmap = (bitstr_t *) NULL;
+ bitoff_t i, bsize;
+ int ll; /* layout info index */
+ char *node_name, *node_idx, *node_cur, *save_req_nodes = NULL;
+ size_t node_name_len;
+ static uint32_t cr_test = 0, cr_enabled = 0;
+
+ if (cr_test == 0) {
+ select_g_get_info_from_plugin(SELECT_CR_PLUGIN,
+ &cr_enabled);
+ cr_test = 1;
+ }
lock_slurmctld(job_write_lock);
job_ptr = find_job_record(jobid);
@@ -134,105 +161,135 @@ static int _start_job(uint32_t jobid, char *hostlist,
if ((job_ptr->details == NULL)
|| (job_ptr->job_state != JOB_PENDING)) {
*err_code = -700;
- *err_msg = "Job not pending, can't update";
- error("wiki: Attempt to start non-pending job %u",
- jobid);
+ *err_msg = "Job not pending, can't start";
+ error("wiki: Attempt to start job %u in state %s",
+ jobid, job_state_string(job_ptr->job_state));
rc = -1;
goto fini;
}
- new_node_list = xstrdup(hostlist);
- if (hostlist && (new_node_list == NULL)) {
- *err_code = -700;
- *err_msg = "Invalid TASKLIST";
- error("wiki: Attempt to set invalid node list for job %u, %s",
- jobid, hostlist);
- rc = -1;
- goto fini;
- }
+ if (task_cnt) {
+ new_node_list = xstrdup(hostlist);
+ if (node_name2bitmap(new_node_list, false, &new_bitmap) != 0) {
+ *err_code = -700;
+ *err_msg = "Invalid TASKLIST";
+ error("wiki: Attempt to set invalid node list for "
+ "job %u, %s",
+ jobid, hostlist);
+ xfree(new_node_list);
+ rc = -1;
+ goto fini;
+ }
- if (node_name2bitmap(new_node_list, false, &new_bitmap) != 0) {
- *err_code = -700;
- *err_msg = "Invalid TASKLIST";
- error("wiki: Attempt to set invalid node list for job %u, %s",
- jobid, hostlist);
- xfree(new_node_list);
- rc = -1;
- goto fini;
+ /* User excluded node list incompatable with Wiki
+ * Exclude all nodes not explicitly requested */
+ FREE_NULL_BITMAP(job_ptr->details->exc_node_bitmap);
+ job_ptr->details->exc_node_bitmap = bit_copy(new_bitmap);
+ bit_not(job_ptr->details->exc_node_bitmap);
}
- /* Remove any excluded nodes, incompatable with Wiki */
- if (job_ptr->details->exc_nodes) {
- error("wiki: clearing exc_nodes for job %u", jobid);
- xfree(job_ptr->details->exc_nodes);
- if (job_ptr->details->exc_node_bitmap)
- FREE_NULL_BITMAP(job_ptr->details->exc_node_bitmap);
+ /* Build layout information from tasklist (assuming that Moab
+ * sends a non-bracketed list of nodes, repeated as many times
+ * as cpus should be used per node); at this point, node names
+ * are comma-separated. This is _not_ a fast algorithm as it
+ * performs many string compares. */
+ xfree(job_ptr->details->req_node_layout);
+ if (task_cnt && cr_enabled) {
+ job_ptr->details->req_node_layout = (uint16_t *)
+ xmalloc(bit_set_count(new_bitmap) * sizeof(uint16_t));
+ bsize = bit_size(new_bitmap);
+ for (i = 0, ll = -1; i < bsize; i++) {
+ if (!bit_test(new_bitmap, i))
+ continue;
+ ll++;
+ node_name = node_record_table_ptr[i].name;
+ node_name_len = strlen(node_name);
+ if (node_name_len == 0)
+ continue;
+ node_cur = tasklist;
+ while (*node_cur) {
+ if ((node_idx = strstr(node_cur, node_name))) {
+ if ((node_idx[node_name_len] == ',') ||
+ (node_idx[node_name_len] == '\0')) {
+ job_ptr->details->
+ req_node_layout[ll]++;
+ }
+ node_cur = strchr(node_idx, ',');
+ if (node_cur)
+ continue;
+ }
+ break;
+ }
+ }
}
- /* start it now */
+ /* save and update job state to start now */
save_req_nodes = job_ptr->details->req_nodes;
job_ptr->details->req_nodes = new_node_list;
save_req_bitmap = job_ptr->details->req_node_bitmap;
job_ptr->details->req_node_bitmap = new_bitmap;
+ old_task_cnt = job_ptr->num_procs;
+ job_ptr->num_procs = MAX(task_cnt, old_task_cnt);
job_ptr->priority = 100000000;
fini: unlock_slurmctld(job_write_lock);
- if (rc == 0) { /* New job to start ASAP */
- (void) schedule(); /* provides own locking */
- /* Check to insure the job was actually started */
- lock_slurmctld(job_write_lock);
- /* job_ptr = find_job_record(jobid); don't bother */
- if ((job_ptr->job_id == jobid) && job_ptr->details &&
- (job_ptr->job_state == JOB_RUNNING)) {
- /* Restore required node list */
- xfree(job_ptr->details->req_nodes);
- job_ptr->details->req_nodes = save_req_nodes;
- FREE_NULL_BITMAP(job_ptr->details->req_node_bitmap);
- job_ptr->details->req_node_bitmap = save_req_bitmap;
- } else {
- xfree(save_req_nodes);
- FREE_NULL_BITMAP(save_req_bitmap);
- }
+ if (rc)
+ return rc;
- if ((job_ptr->job_id == jobid)
- && (job_ptr->job_state != JOB_RUNNING)) {
- uint16_t wait_reason = 0;
- char *wait_string;
-
- /* restore job state */
- job_ptr->priority = 0;
- if (job_ptr->details) {
- /* Details get cleared on job abort; happens
- * if the request is sufficiently messed up.
- * This happens when Moab tries to start a
- * a job on invalid nodes (wrong partition). */
- xfree(job_ptr->details->req_nodes);
- FREE_NULL_BITMAP(job_ptr->details->
- req_node_bitmap);
- }
- if (job_ptr->job_state == JOB_FAILED)
- wait_string = "Invalid request, job aborted";
- else {
- wait_reason = job_ptr->state_reason;
- if (wait_reason == WAIT_HELD) {
- /* some job is completing, slurmctld did
- * not even try to schedule this job */
- wait_reason = WAIT_RESOURCES;
- }
- wait_string = job_reason_string(wait_reason);
- job_ptr->state_reason = WAIT_HELD;
+ /* No errors so far */
+ (void) schedule(); /* provides own locking */
+
+ /* Check to insure the job was actually started */
+ lock_slurmctld(job_write_lock);
+ if (job_ptr->job_id != jobid)
+ job_ptr = find_job_record(jobid);
+
+ if (job_ptr && (job_ptr->job_id == jobid)
+ && (job_ptr->job_state != JOB_RUNNING)) {
+ uint16_t wait_reason = 0;
+ char *wait_string;
+
+ if (job_ptr->job_state == JOB_FAILED)
+ wait_string = "Invalid request, job aborted";
+ else {
+ wait_reason = job_ptr->state_reason;
+ if (wait_reason == WAIT_HELD) {
+ /* some job is completing, slurmctld did
+ * not even try to schedule this job */
+ wait_reason = WAIT_RESOURCES;
}
- *err_code = -910 - wait_reason;
- snprintf(tmp_msg, sizeof(tmp_msg),
- "Could not start job %u: %s",
- jobid, wait_string);
- *err_msg = tmp_msg;
- error("wiki: %s", tmp_msg);
- rc = -1;
+ wait_string = job_reason_string(wait_reason);
+ job_ptr->state_reason = WAIT_HELD;
}
- unlock_slurmctld(job_write_lock);
- schedule_node_save(); /* provides own locking */
- schedule_job_save(); /* provides own locking */
+ *err_code = -910 - wait_reason;
+ snprintf(tmp_msg, sizeof(tmp_msg),
+ "Could not start job %u(%s): %s",
+ jobid, new_node_list, wait_string);
+ *err_msg = tmp_msg;
+ error("wiki: %s", tmp_msg);
+
+ /* restore some of job state */
+ job_ptr->priority = 0;
+ job_ptr->num_procs = old_task_cnt;
+ rc = -1;
}
+
+ if (job_ptr && (job_ptr->job_id == jobid) && job_ptr->details) {
+ /* Restore required node list in case job requeued */
+ xfree(job_ptr->details->req_nodes);
+ job_ptr->details->req_nodes = save_req_nodes;
+ FREE_NULL_BITMAP(job_ptr->details->req_node_bitmap);
+ job_ptr->details->req_node_bitmap = save_req_bitmap;
+ FREE_NULL_BITMAP(job_ptr->details->exc_node_bitmap);
+ xfree(job_ptr->details->req_node_layout);
+ } else {
+ error("wiki: start_job(%u) job missing", jobid);
+ xfree(save_req_nodes);
+ FREE_NULL_BITMAP(save_req_bitmap);
+ }
+
+ unlock_slurmctld(job_write_lock);
+ schedule_node_save(); /* provides own locking */
+ schedule_job_save(); /* provides own locking */
return rc;
}
diff --git a/src/plugins/sched/wiki/suspend_job.c b/src/plugins/sched/wiki/suspend_job.c
new file mode 100644
index 0000000000000000000000000000000000000000..058d2dce3f271bb4f7775a8a3aa5701c80b39bfe
--- /dev/null
+++ b/src/plugins/sched/wiki/suspend_job.c
@@ -0,0 +1,85 @@
+/*****************************************************************************\
+ * suspend_job.c - Process Wiki suspend job request
+ *****************************************************************************
+ * Copyright (C) 2006 The Regents of the University of California.
+ * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ * Written by Morris Jette <jette1@llnl.gov>
+ * UCRL-CODE-226842.
+ *
+ * This file is part of SLURM, a resource management program.
+ * For details, see <http://www.llnl.gov/linux/slurm/>.
+ *
+ * SLURM is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * In addition, as a special exception, the copyright holders give permission
+ * to link the code of portions of this program with the OpenSSL library under
+ * certain conditions as described in each individual source file, and
+ * distribute linked combinations including the two. You must obey the GNU
+ * General Public License in all respects for all of the code used other than
+ * OpenSSL. If you modify file(s) with this exception, you may extend this
+ * exception to your version of the file(s), but you are not obligated to do
+ * so. If you do not wish to do so, delete this exception statement from your
+ * version. If you delete this exception statement from all source files in
+ * the program, then also delete it here.
+ *
+ * SLURM is distributed in the hope that it will be useful, but WITHOUT ANY
+ * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with SLURM; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+\*****************************************************************************/
+
+#include "./msg.h"
+#include "src/slurmctld/locks.h"
+#include "src/slurmctld/slurmctld.h"
+
+/* RET 0 on success, -1 on failure */
+extern int suspend_job(char *cmd_ptr, int *err_code, char **err_msg)
+{
+ char *arg_ptr, *tmp_char;
+ int slurm_rc;
+ suspend_msg_t msg;
+ uint32_t jobid;
+ static char reply_msg[128];
+ /* Locks: write job and node info */
+ slurmctld_lock_t job_write_lock = {
+ NO_LOCK, WRITE_LOCK, WRITE_LOCK, NO_LOCK };
+
+ arg_ptr = strstr(cmd_ptr, "ARG=");
+ if (arg_ptr == NULL) {
+ *err_code = -300;
+ *err_msg = "SUSPENDJOB lacks ARG";
+ error("wiki: SUSPENDJOB lacks ARG");
+ return -1;
+ }
+ jobid = strtoul(arg_ptr+4, &tmp_char, 10);
+ if ((tmp_char[0] != '\0') && (!isspace(tmp_char[0]))) {
+ *err_code = -300;
+ *err_msg = "Invalid ARG value";
+ error("wiki: SUSPENDJOB has invalid jobid");
+ return -1;
+ }
+
+ msg.job_id = jobid;
+ msg.op = SUSPEND_JOB;
+ lock_slurmctld(job_write_lock);
+ slurm_rc = job_suspend(&msg, 0, -1);
+ unlock_slurmctld(job_write_lock);
+ if (slurm_rc != SLURM_SUCCESS) {
+ *err_code = -700;
+ *err_msg = slurm_strerror(slurm_rc);
+ error("wiki: Failed to suspend job %u (%m)", jobid);
+ return -1;
+ }
+
+ snprintf(reply_msg, sizeof(reply_msg),
+ "job %u suspended successfully", jobid);
+ *err_msg = reply_msg;
+ return 0;
+}
diff --git a/src/plugins/sched/wiki2/get_jobs.c b/src/plugins/sched/wiki2/get_jobs.c
index e8688b3d5919a5d994f108a3d9741cefb8c17178..ea2afa63f7a5f7f8f8227a59c49472fd14304355 100644
--- a/src/plugins/sched/wiki2/get_jobs.c
+++ b/src/plugins/sched/wiki2/get_jobs.c
@@ -50,6 +50,7 @@ static char * _dump_job(struct job_record *job_ptr, int state_info);
static char * _get_group_name(gid_t gid);
static void _get_job_comment(struct job_record *job_ptr,
char *buffer, int buf_size);
+static uint16_t _get_job_cpus_per_task(struct job_record *job_ptr);
static uint32_t _get_job_end_time(struct job_record *job_ptr);
static char * _get_job_features(struct job_record *job_ptr);
static uint32_t _get_job_min_disk(struct job_record *job_ptr);
@@ -274,6 +275,11 @@ static char * _dump_job(struct job_record *job_ptr, int state_info)
_get_job_min_nodes(job_ptr));
xstrcat(buf, tmp);
+ snprintf(tmp, sizeof(tmp),
+ "DPROCS=%u;",
+ _get_job_cpus_per_task(job_ptr));
+ xstrcat(buf, tmp);
+
snprintf(tmp, sizeof(tmp),
"QUEUETIME=%u;STARTTIME=%u;RCLASS=%s;",
_get_job_submit_time(job_ptr),
@@ -346,7 +352,8 @@ static void _get_job_comment(struct job_record *job_ptr,
/* SHARED NODES */
if (cr_enabled) { /* consumable resources */
- if (job_ptr->part_ptr->shared == SHARED_EXCLUSIVE)
+ if (job_ptr->part_ptr &&
+ (job_ptr->part_ptr->shared == SHARED_EXCLUSIVE))
sharing = 0;
else if (job_ptr->details && (job_ptr->details->shared != 0))
sharing = 1;
@@ -383,6 +390,15 @@ static void _get_job_comment(struct job_record *job_ptr,
size += snprintf((buffer + size), (buf_size - size), "\";");
}
+static uint16_t _get_job_cpus_per_task(struct job_record *job_ptr)
+{
+ uint16_t cpus_per_task = 1;
+
+ if (job_ptr->details && job_ptr->details->cpus_per_task)
+ cpus_per_task = job_ptr->details->cpus_per_task;
+ return cpus_per_task;
+}
+
static uint32_t _get_job_min_mem(struct job_record *job_ptr)
{
if (job_ptr->details)
@@ -440,7 +456,7 @@ static uint32_t _get_job_tasks(struct job_record *job_ptr)
job_ptr->details->ntasks_per_node));
}
- return task_cnt;
+ return task_cnt / _get_job_cpus_per_task(job_ptr);
}
static uint32_t _get_job_time_limit(struct job_record *job_ptr)
diff --git a/src/plugins/sched/wiki2/hostlist.c b/src/plugins/sched/wiki2/hostlist.c
index c31dc6bf89207b154d55f354e6f062353055becc..7a96bd30cc5956bd22ab0b1c1826373e33ec285a 100644
--- a/src/plugins/sched/wiki2/hostlist.c
+++ b/src/plugins/sched/wiki2/hostlist.c
@@ -160,7 +160,7 @@ extern char * slurm_job2moab_task_list(struct job_record *job_ptr)
/* Return task list in Moab format 1: tux0:tux0:tux1:tux1:tux2 */
static char * _task_list(struct job_record *job_ptr)
{
- int i, j;
+ int i, j, task_cnt;
char *buf = NULL, *host;
hostlist_t hl = hostlist_create(job_ptr->nodes);
@@ -178,7 +178,10 @@ static char * _task_list(struct job_record *job_ptr)
job_ptr->alloc_lps_cnt);
break;
}
- for (j=0; j<job_ptr->alloc_lps[i]; j++) {
+ task_cnt = job_ptr->alloc_lps[i];
+ if (job_ptr->details && job_ptr->details->cpus_per_task)
+ task_cnt /= job_ptr->details->cpus_per_task;
+ for (j=0; j<task_cnt; j++) {
if (buf)
xstrcat(buf, ":");
xstrcat(buf, host);
@@ -247,7 +250,7 @@ static void _append_hl_buf(char **buf, hostlist_t *hl_tmp, int *reps)
/* Return task list in Moab format 2: tux[0-1]*2:tux2 */
static char * _task_list_exp(struct job_record *job_ptr)
{
- int i, reps = -1;
+ int i, reps = -1, task_cnt;
char *buf = NULL, *host;
hostlist_t hl = hostlist_create(job_ptr->nodes);
hostlist_t hl_tmp = (hostlist_t) NULL;
@@ -267,7 +270,10 @@ static char * _task_list_exp(struct job_record *job_ptr)
break;
}
- if (reps == job_ptr->alloc_lps[i]) {
+ task_cnt = job_ptr->alloc_lps[i];
+ if (job_ptr->details && job_ptr->details->cpus_per_task)
+ task_cnt /= job_ptr->details->cpus_per_task;
+ if (reps == task_cnt) {
/* append to existing hostlist record */
if (hostlist_push(hl_tmp, host) == 0)
error("hostlist_push failure");
@@ -278,7 +284,7 @@ static char * _task_list_exp(struct job_record *job_ptr)
/* start new hostlist record */
hl_tmp = hostlist_create(host);
if (hl_tmp)
- reps = job_ptr->alloc_lps[i];
+ reps = task_cnt;
else
error("hostlist_create failure");
}
diff --git a/src/plugins/select/bluegene/plugin/bg_job_place.c b/src/plugins/select/bluegene/plugin/bg_job_place.c
index 8fc6c6b8a83828d393ea3425a071be8d5102a669..f03eae075ad2d68af028d2bc7f778946e34fdfe1 100644
--- a/src/plugins/select/bluegene/plugin/bg_job_place.c
+++ b/src/plugins/select/bluegene/plugin/bg_job_place.c
@@ -2,7 +2,7 @@
* bg_job_place.c - blue gene job placement (e.g. base block selection)
* functions.
*
- * $Id: bg_job_place.c 12533 2007-10-22 23:19:23Z jette $
+ * $Id: bg_job_place.c 12627 2007-11-06 19:48:55Z jette $
*****************************************************************************
* Copyright (C) 2004-2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -474,7 +474,8 @@ try_again:
/* We use the proccessor count per partition here
mostly to see if we can run on a smaller partition.
*/
- convert_num_unit((float)proc_cnt, tmp_char, UNIT_NONE);
+ convert_num_unit((float)proc_cnt, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
debug("block %s CPU count (%s) not suitable",
record->bg_block_id,
tmp_char);
@@ -490,7 +491,7 @@ try_again:
|| (req_nodes != 0 && record->bp_count > req_nodes)
|| (record->bp_count < target_size)) {
convert_num_unit((float)record->node_cnt, tmp_char,
- UNIT_NONE);
+ sizeof(tmp_char), UNIT_NONE);
debug("block %s node count (%s) not suitable",
record->bg_block_id,
tmp_char);
diff --git a/src/plugins/select/cons_res/select_cons_res.c b/src/plugins/select/cons_res/select_cons_res.c
index 6ee686efd8247dc22d6cca202142cb84e6a6f495..ecb3aa3c68a8a99d506c59ab7e738f5afeb6e86f 100644
--- a/src/plugins/select/cons_res/select_cons_res.c
+++ b/src/plugins/select/cons_res/select_cons_res.c
@@ -2,7 +2,7 @@
* select_cons_res.c - node selection plugin supporting consumable
* resources policies.
*
- * $Id: select_cons_res.c 12452 2007-10-05 19:07:07Z da $
+ * $Id: select_cons_res.c 12649 2007-11-15 18:02:35Z da $
*****************************************************************************\
*
* The following example below illustrates how four jobs are allocated
@@ -545,7 +545,7 @@ static void _append_to_job_list(struct select_cr_job *new_job)
/*
* _count_cpus - report how many cpus are available with the identified nodes
*/
-static void _count_cpus(unsigned *bitmap, uint16_t *sum)
+static void _count_cpus(bitstr_t *bitmap, uint16_t *sum)
{
int i, allocated_lps;
*sum = 0;
diff --git a/src/plugins/switch/federation/federation.c b/src/plugins/switch/federation/federation.c
index e67383afb5d4f92d9b04d98eb4c9489b6b8dbe2e..aa0a8d37c8796a418950f38435e3c6d9c1276c3a 100644
--- a/src/plugins/switch/federation/federation.c
+++ b/src/plugins/switch/federation/federation.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
** federation.c - Library routines for initiating jobs on IBM Federation
- ** $Id: federation.c 11600 2007-05-31 23:59:57Z morrone $
+ ** $Id: federation.c 12618 2007-11-02 22:31:39Z jette $
*****************************************************************************
* Copyright (C) 2004 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -2250,14 +2250,20 @@ _check_rdma_job_count(char *adapter)
{
unsigned int job_count;
unsigned int *job_keys;
- int z;
+ int rc, z;
+
+ rc = ntbl_rdma_jobs(NTBL_VERSION, adapter,
+ &job_count, &job_keys);
+ if (rc != NTBL_SUCESSS) {
+ error("ntbl_rdma_jobs(): %d", rc);
+ return SLURM_ERROR;
+ }
- ntbl_rdma_jobs(NTBL_VERSION, adapter,
- &job_count, &job_keys);
debug3("Adapter %s, RDMA job_count = %u",
adapter, job_count);
for (z = 0; z < job_count; z++)
debug3(" job key = %u", job_keys[z]);
+ free(job_keys);
if (job_count >= 4) {
error("RDMA job_count is too high: %u", job_count);
return SLURM_ERROR;
diff --git a/src/sacct/print.c b/src/sacct/print.c
index 3de3f88d2e9e110a20bd3a6bcc7b285d97fc62cb..dd7172657ee1856eaa9eafa6d70578ce362aca7d 100644
--- a/src/sacct/print.c
+++ b/src/sacct/print.c
@@ -238,7 +238,8 @@ void print_idrss(type_t type, void *object)
rusage = step->rusage;
break;
}
- convert_num_unit((float)rusage.ru_idrss, outbuf, UNIT_NONE);
+ convert_num_unit((float)rusage.ru_idrss, outbuf, sizeof(outbuf),
+ UNIT_NONE);
printf("%8s", outbuf);
}
@@ -726,13 +727,14 @@ void print_pages(type_t type, void *object)
sacct = job->sacct;
nodes = job->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_pages, buf1, UNIT_NONE);
+ convert_num_unit((float)sacct.max_pages,
+ buf1, sizeof(buf1), UNIT_NONE);
if(job->track_steps)
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/- - -", buf1);
else {
convert_num_unit((float)sacct.ave_pages,
- buf2, UNIT_NONE);
+ buf2, sizeof(buf2), UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
@@ -746,8 +748,10 @@ void print_pages(type_t type, void *object)
sacct = step->sacct;
nodes = step->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_pages, buf1, UNIT_NONE);
- convert_num_unit((float)sacct.ave_pages, buf2, UNIT_NONE);
+ convert_num_unit((float)sacct.max_pages, buf1, sizeof(buf1),
+ UNIT_NONE);
+ convert_num_unit((float)sacct.ave_pages, buf2, sizeof(buf2),
+ UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
@@ -782,13 +786,14 @@ void print_rss(type_t type, void *object)
sacct = job->sacct;
nodes = job->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_rss, buf1, UNIT_NONE);
+ convert_num_unit((float)sacct.max_rss, buf1, sizeof(buf1),
+ UNIT_NONE);
if(job->track_steps)
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/- - -", buf1);
else {
convert_num_unit((float)sacct.ave_rss,
- buf2, UNIT_NONE);
+ buf2, sizeof(buf2), UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
@@ -802,8 +807,10 @@ void print_rss(type_t type, void *object)
sacct = step->sacct;
nodes = step->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_rss, buf1, UNIT_NONE);
- convert_num_unit((float)sacct.ave_rss, buf2, UNIT_NONE);
+ convert_num_unit((float)sacct.max_rss, buf1, sizeof(buf1),
+ UNIT_NONE);
+ convert_num_unit((float)sacct.ave_rss, buf2, sizeof(buf2),
+ UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
@@ -1059,12 +1066,13 @@ void print_vsize(type_t type, void *object)
sacct = job->sacct;
nodes = job->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_vsize, buf1, UNIT_NONE);
+ convert_num_unit((float)sacct.max_vsize,
+ buf1, sizeof(buf1),UNIT_NONE);
if(job->track_steps)
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/- - -", buf1);
else {
convert_num_unit((float)sacct.ave_vsize,
- buf2, UNIT_NONE);
+ buf2, sizeof(buf2), UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
@@ -1078,8 +1086,10 @@ void print_vsize(type_t type, void *object)
sacct = step->sacct;
nodes = step->nodes;
pos = sacct.min_cpu_id.nodeid;
- convert_num_unit((float)sacct.max_vsize, buf1, UNIT_NONE);
- convert_num_unit((float)sacct.ave_vsize, buf2, UNIT_NONE);
+ convert_num_unit((float)sacct.max_vsize, buf1, sizeof(buf1),
+ UNIT_NONE);
+ convert_num_unit((float)sacct.ave_vsize, buf2, sizeof(buf2),
+ UNIT_NONE);
find_hostname(pos, nodes, buf3);
snprintf(outbuf, FORMAT_STRING_SIZE, "%s/%s:%u - %s",
buf1,
diff --git a/src/salloc/opt.c b/src/salloc/opt.c
index c6d4e84950a456de2c55199cf355acd188f5a84b..80b5435bcbf52f36d907a8906961f208a9c70672 100644
--- a/src/salloc/opt.c
+++ b/src/salloc/opt.c
@@ -907,9 +907,8 @@ void set_options(const int argc, char **argv)
case LONG_OPT_BEGIN:
opt.begin = parse_time(optarg);
if (opt.begin == 0) {
- error("Invalid time specification %s",
+ fatal("Invalid time specification %s",
optarg);
- exit(1);
}
break;
case LONG_OPT_MAIL_TYPE:
diff --git a/src/salloc/salloc.c b/src/salloc/salloc.c
index 6d9aa5a3fd17355bfabae0cc5f9346ebc5f5a2bb..66981c6a12ad7f41039babb3d5d59683f2b31017 100644
--- a/src/salloc/salloc.c
+++ b/src/salloc/salloc.c
@@ -2,7 +2,7 @@
* salloc.c - Request a SLURM job allocation and
* launch a user-specified command.
*
- * $Id: salloc.c 12536 2007-10-22 23:57:53Z jette $
+ * $Id: salloc.c 12700 2007-11-27 23:39:24Z jette $
*****************************************************************************
* Copyright (C) 2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -51,6 +51,8 @@
#include "src/salloc/opt.h"
#include "src/salloc/msg.h"
+#define MAX_RETRIES 3
+
char **command_argv;
int command_argc;
pid_t command_pid = -1;
@@ -79,9 +81,11 @@ int main(int argc, char *argv[])
char **env = NULL;
int status = 0;
int errnum = 0;
+ int retries = 0;
pid_t pid;
pid_t rc_pid = 0;
int rc = 0;
+ static char *msg = "Slurm job queue full, sleeping and retrying.";
log_init(xbasename(argv[0]), logopt, 0, NULL);
if (initialize_and_process_args(argc, argv) < 0) {
@@ -116,8 +120,18 @@ int main(int argc, char *argv[])
xsignal(SIGUSR2, _signal_while_allocating);
before = time(NULL);
- alloc = slurm_allocate_resources_blocking(&desc, opt.max_wait,
- _pending_callback);
+ while ((alloc = slurm_allocate_resources_blocking(&desc, opt.max_wait,
+ _pending_callback)) == NULL) {
+ if ((errno != ESLURM_ERROR_ON_DESC_TO_RECORD_COPY) ||
+ (retries >= MAX_RETRIES))
+ break;
+ if (retries == 0)
+ error(msg);
+ else
+ debug(msg);
+ sleep (++retries);
+ }
+
if (alloc == NULL) {
if (allocation_interrupted) {
/* cancelled by signal */
diff --git a/src/sbatch/opt.c b/src/sbatch/opt.c
index 24427f8f66d08be25bfb344d27b1e757dce79153..ecdffd4538709179825a5620c8f165e0f30f87aa 100644
--- a/src/sbatch/opt.c
+++ b/src/sbatch/opt.c
@@ -141,6 +141,7 @@ static void _opt_pbs_batch_script(const void *body, int size);
/* set options based upon env vars */
static void _opt_env(void);
+static void _proc_get_user_env(char *optarg);
/* list known options and their settings */
static void _opt_list(void);
@@ -446,7 +447,9 @@ static void _opt_default()
opt.ifname = xstrdup("/dev/null");
opt.ofname = NULL;
opt.efname = NULL;
- opt.get_user_env = -1;
+
+ opt.get_user_env_time = -1;
+ opt.get_user_env_mode = -1;
}
/*---[ env var processing ]-----------------------------------------------*/
@@ -616,7 +619,7 @@ static struct option long_options[] = {
{"immediate", no_argument, 0, 'I'},
{"job-name", required_argument, 0, 'J'},
{"no-kill", no_argument, 0, 'k'},
- {"tasks", required_argument, 0, 'n'},
+ {"ntasks", required_argument, 0, 'n'},
{"nodes", required_argument, 0, 'N'},
{"output", required_argument, 0, 'o'},
{"overcommit", no_argument, 0, 'O'},
@@ -1227,6 +1230,10 @@ static void _set_options(int argc, char **argv)
break;
case LONG_OPT_BEGIN:
opt.begin = parse_time(optarg);
+ if (opt.begin == 0) {
+ fatal("Invalid time specification %s",
+ optarg);
+ }
break;
case LONG_OPT_MAIL_TYPE:
opt.mail_type |= _parse_mail_type(optarg);
@@ -1285,15 +1292,17 @@ static void _set_options(int argc, char **argv)
break;
case LONG_OPT_TASKSPERNODE:
opt.tasks_per_node = _get_int(optarg, "ntasks-per-node");
+ setenvf(NULL, "SLURM_NTASKS_PER_NODE", "%d",
+ opt.tasks_per_node);
break;
case LONG_OPT_WRAP:
/* handled in process_options_first_pass() */
break;
case LONG_OPT_GET_USER_ENV:
if (optarg)
- opt.get_user_env = strtol(optarg, NULL, 10);
+ _proc_get_user_env(optarg);
else
- opt.get_user_env = 0;
+ opt.get_user_env_time = 0;
break;
default:
fatal("Unrecognized command line parameter %c",
@@ -1306,6 +1315,25 @@ static void _set_options(int argc, char **argv)
}
}
+static void _proc_get_user_env(char *optarg)
+{
+ char *end_ptr;
+
+ if ((optarg[0] >= '0') && (optarg[0] <= '9'))
+ opt.get_user_env_time = strtol(optarg, &end_ptr, 10);
+ else {
+ opt.get_user_env_time = 0;
+ end_ptr = optarg;
+ }
+
+ if ((end_ptr == NULL) || (end_ptr[0] == '\0'))
+ return;
+ if ((end_ptr[0] == 's') || (end_ptr[0] == 'S'))
+ opt.get_user_env_mode = 1;
+ else if ((end_ptr[0] == 'l') || (end_ptr[0] == 'L'))
+ opt.get_user_env_mode = 2;
+}
+
static void _set_pbs_options(int argc, char **argv)
{
int opt_char, option_index = 0;
diff --git a/src/sbatch/opt.h b/src/sbatch/opt.h
index a14602553c88ae4d7250109c8f8a75e83e60176c..9591a83869a1b5c93bf4616e169d3d4690230f4d 100644
--- a/src/sbatch/opt.h
+++ b/src/sbatch/opt.h
@@ -1,6 +1,6 @@
/*****************************************************************************\
* opt.h - definitions for srun option processing
- * $Id: opt.h 12574 2007-10-26 17:00:52Z jette $
+ * $Id: opt.h 12697 2007-11-27 22:02:29Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -116,7 +116,8 @@ typedef struct sbatch_options {
char *ifname; /* input file name */
char *ofname; /* output file name */
char *efname; /* error file name */
- int get_user_env; /* --get-user-env[=timeout] */
+ int get_user_env_time; /* --get-user-env[=timeout] */
+ int get_user_env_mode; /* --get-user-env=[S|L] */
} opt_t;
extern opt_t opt;
diff --git a/src/sbatch/sbatch.c b/src/sbatch/sbatch.c
index 7e2e11049b7525c37c683039ff27af68ce736b6a..b5b340cda91387498e3eeee7afcb114b8dc61562 100644
--- a/src/sbatch/sbatch.c
+++ b/src/sbatch/sbatch.c
@@ -1,7 +1,7 @@
/*****************************************************************************\
* sbatch.c - Submit a SLURM batch script.
*
- * $Id: sbatch.c 12574 2007-10-26 17:00:52Z jette $
+ * $Id: sbatch.c 12700 2007-11-27 23:39:24Z jette $
*****************************************************************************
* Copyright (C) 2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -46,6 +46,8 @@
#include "src/sbatch/opt.h"
+#define MAX_RETRIES 3
+
static int fill_job_desc_from_opts(job_desc_msg_t *desc);
static void *get_script_buffer(const char *filename, int *size);
static void set_prio_process_env(void);
@@ -60,6 +62,7 @@ int main(int argc, char *argv[])
char *script_name;
void *script_body;
int script_size = 0;
+ int retries = 0;
log_init(xbasename(argv[0]), logopt, 0, NULL);
script_name = process_options_first_pass(argc, argv);
@@ -93,10 +96,21 @@ int main(int argc, char *argv[])
desc.script = (char *)script_body;
- if (slurm_submit_batch_job(&desc, &resp) == -1) {
- error("Batch job submission failed: %m");
- exit(3);
- }
+ while (slurm_submit_batch_job(&desc, &resp) < 0) {
+ static char *msg = "Slurm job queue full, sleeping and retrying.";
+
+ if ((errno != ESLURM_ERROR_ON_DESC_TO_RECORD_COPY) ||
+ (retries >= MAX_RETRIES)) {
+ error("Batch job submission failed: %m");
+ exit(3);
+ }
+
+ if (retries)
+ debug(msg);
+ else
+ error(msg);
+ sleep (++retries);
+ }
info("Submitted batch job %d", resp->job_id);
xfree(desc.script);
@@ -194,12 +208,13 @@ static int fill_job_desc_from_opts(job_desc_msg_t *desc)
desc->shared = opt.shared;
desc->environment = NULL;
- if (opt.get_user_env >= 0) {
+ if (opt.get_user_env_time >= 0) {
struct passwd *pw = NULL;
pw = getpwuid(opt.uid);
if (pw != NULL) {
desc->environment = env_array_user_default(pw->pw_name,
- opt.get_user_env);
+ opt.get_user_env_time,
+ opt.get_user_env_mode);
/* FIXME - should we abort if j->environment
* is NULL? */
}
diff --git a/src/sinfo/print.c b/src/sinfo/print.c
index 9a235a17627ba10e43023c9d567ea0dd83854ef0..6e4704c245d9d170673d4e6aa984c58dfe7b3cdb 100644
--- a/src/sinfo/print.c
+++ b/src/sinfo/print.c
@@ -172,10 +172,10 @@ static int _print_secs(long time, int width, bool right, bool cut_output)
static int
_build_min_max_16_string(char *buffer, int buf_size, uint16_t min, uint16_t max, bool range)
{
- char tmp_min[7];
- char tmp_max[7];
- convert_num_unit((float)min, tmp_min, UNIT_NONE);
- convert_num_unit((float)max, tmp_max, UNIT_NONE);
+ char tmp_min[8];
+ char tmp_max[8];
+ convert_num_unit((float)min, tmp_min, sizeof(tmp_min), UNIT_NONE);
+ convert_num_unit((float)max, tmp_max, sizeof(tmp_max), UNIT_NONE);
if (max == min)
return snprintf(buffer, buf_size, "%s", tmp_max);
@@ -193,10 +193,10 @@ _build_min_max_16_string(char *buffer, int buf_size, uint16_t min, uint16_t max,
static int
_build_min_max_32_string(char *buffer, int buf_size, uint32_t min, uint32_t max, bool range)
{
- char tmp_min[7];
- char tmp_max[7];
- convert_num_unit((float)min, tmp_min, UNIT_NONE);
- convert_num_unit((float)max, tmp_max, UNIT_NONE);
+ char tmp_min[8];
+ char tmp_max[8];
+ convert_num_unit((float)min, tmp_min, sizeof(tmp_min), UNIT_NONE);
+ convert_num_unit((float)max, tmp_max, sizeof(tmp_max), UNIT_NONE);
if (max == min)
return snprintf(buffer, buf_size, "%s", tmp_max);
@@ -305,20 +305,20 @@ int _print_cpus_aiot(sinfo_data_t * sinfo_data, int width,
bool right_justify, char *suffix)
{
char id[FORMAT_STRING_SIZE];
- char tmpa[7];
- char tmpi[7];
- char tmpo[7];
- char tmpt[7];
+ char tmpa[8];
+ char tmpi[8];
+ char tmpo[8];
+ char tmpt[8];
if (sinfo_data) {
#ifdef HAVE_BG
convert_num_unit((float)sinfo_data->cpus_alloc,
- tmpa, UNIT_NONE);
+ tmpa, sizeof(tmpa), UNIT_NONE);
convert_num_unit((float)sinfo_data->cpus_idle,
- tmpi, UNIT_NONE);
+ tmpi, sizeof(tmpi), UNIT_NONE);
convert_num_unit((float)sinfo_data->cpus_other,
- tmpo, UNIT_NONE);
+ tmpo, sizeof(tmpo), UNIT_NONE);
convert_num_unit((float)sinfo_data->cpus_total,
- tmpt, UNIT_NONE);
+ tmpt, sizeof(tmpt), UNIT_NONE);
#else
sprintf(tmpa, "%u", sinfo_data->cpus_alloc);
sprintf(tmpi, "%u", sinfo_data->cpus_idle);
@@ -516,13 +516,13 @@ int _print_nodes_t(sinfo_data_t * sinfo_data, int width,
bool right_justify, char *suffix)
{
char id[FORMAT_STRING_SIZE];
- char tmp[7];
+ char tmp[8];
if (sinfo_data) {
#ifdef HAVE_BG
convert_num_unit((float)sinfo_data->nodes_total, tmp,
- UNIT_NONE);
+ sizeof(tmp), UNIT_NONE);
#else
- sprintf(tmp, "%d", sinfo_data->nodes_total);
+ snprintf(tmp, sizeof(tmp), "%d", sinfo_data->nodes_total);
#endif
snprintf(id, FORMAT_STRING_SIZE, "%s", tmp);
_print_str(id, width, right_justify, true);
@@ -538,17 +538,17 @@ int _print_nodes_ai(sinfo_data_t * sinfo_data, int width,
bool right_justify, char *suffix)
{
char id[FORMAT_STRING_SIZE];
- char tmpa[7];
- char tmpi[7];
+ char tmpa[8];
+ char tmpi[8];
if (sinfo_data) {
#ifdef HAVE_BG
convert_num_unit((float)sinfo_data->nodes_alloc,
- tmpa, UNIT_NONE);
+ tmpa, sizeof(tmpa), UNIT_NONE);
convert_num_unit((float)sinfo_data->nodes_idle,
- tmpi, UNIT_NONE);
+ tmpi, sizeof(tmpi), UNIT_NONE);
#else
- sprintf(tmpa, "%d", sinfo_data->nodes_alloc);
- sprintf(tmpi, "%d", sinfo_data->nodes_idle);
+ snprintf(tmpa, sizeof(tmpa), "%d", sinfo_data->nodes_alloc);
+ snprintf(tmpi, sizeof(tmpi), "%d", sinfo_data->nodes_idle);
#endif
snprintf(id, FORMAT_STRING_SIZE, "%s/%s",
tmpa, tmpi);
@@ -565,25 +565,25 @@ int _print_nodes_aiot(sinfo_data_t * sinfo_data, int width,
bool right_justify, char *suffix)
{
char id[FORMAT_STRING_SIZE];
- char tmpa[7];
- char tmpi[7];
- char tmpo[7];
- char tmpt[7];
+ char tmpa[8];
+ char tmpi[8];
+ char tmpo[8];
+ char tmpt[8];
if (sinfo_data) {
#ifdef HAVE_BG
convert_num_unit((float)sinfo_data->nodes_alloc,
- tmpa, UNIT_NONE);
+ tmpa, sizeof(tmpa), UNIT_NONE);
convert_num_unit((float)sinfo_data->nodes_idle,
- tmpi, UNIT_NONE);
+ tmpi, sizeof(tmpi), UNIT_NONE);
convert_num_unit((float)sinfo_data->nodes_other,
- tmpo, UNIT_NONE);
+ tmpo, sizeof(tmpo), UNIT_NONE);
convert_num_unit((float)sinfo_data->nodes_total,
- tmpt, UNIT_NONE);
+ tmpt, sizeof(tmpt), UNIT_NONE);
#else
- sprintf(tmpa, "%u", sinfo_data->nodes_alloc);
- sprintf(tmpi, "%u", sinfo_data->nodes_idle);
- sprintf(tmpo, "%u", sinfo_data->nodes_other);
- sprintf(tmpt, "%u", sinfo_data->nodes_total);
+ snprintf(tmpa, sizeof(tmpa), "%u", sinfo_data->nodes_alloc);
+ snprintf(tmpi, sizeof(tmpi), "%u", sinfo_data->nodes_idle);
+ snprintf(tmpo, sizeof(tmpo), "%u", sinfo_data->nodes_other);
+ snprintf(tmpt, sizeof(tmpt), "%u", sinfo_data->nodes_total);
#endif
snprintf(id, FORMAT_STRING_SIZE, "%s/%s/%s/%s",
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index 521700b5e27b2b86b016cd49cc4b3d68568d519d..7d589f168eeae287921eed73cfb5f503de756b82 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -3,7 +3,7 @@
* Note: there is a global job list (job_list), time stamp
* (last_job_update), and hash table (job_hash)
*
- * $Id: job_mgr.c 12460 2007-10-05 23:50:48Z jette $
+ * $Id: job_mgr.c 12655 2007-11-20 21:02:43Z jette $
*****************************************************************************
* Copyright (C) 2002-2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -3090,15 +3090,15 @@ void reset_job_bitmaps(void)
job_ptr->part_ptr = part_ptr;
FREE_NULL_BITMAP(job_ptr->node_bitmap);
- if ((job_ptr->nodes_completing)
- && (node_name2bitmap(job_ptr->nodes_completing,
- false, &job_ptr->node_bitmap))) {
+ if ((job_ptr->nodes_completing) &&
+ (node_name2bitmap(job_ptr->nodes_completing,
+ false, &job_ptr->node_bitmap))) {
error("Invalid nodes (%s) for job_id %u",
job_ptr->nodes_completing,
job_ptr->job_id);
job_fail = true;
- } else if ((job_ptr->nodes)
- && (node_name2bitmap(job_ptr->nodes, false,
+ } else if ((job_ptr->node_bitmap == NULL) && job_ptr->nodes &&
+ (node_name2bitmap(job_ptr->nodes, false,
&job_ptr->node_bitmap))) {
error("Invalid nodes (%s) for job_id %u",
job_ptr->nodes, job_ptr->job_id);
@@ -4786,6 +4786,8 @@ extern int job_requeue (uid_t uid, uint32_t job_id, slurm_fd conn_fd)
else
job_ptr->end_time = now;
deallocate_nodes(job_ptr, false, suspended);
+ if (job_ptr->details)
+ xfree(job_ptr->details->req_node_layout);
job_completion_logger(job_ptr);
//FIXME: Test accounting
@@ -4834,12 +4836,11 @@ extern void update_job_nodes_completing(void)
job_iterator = list_iterator_create(job_list);
while ((job_ptr = (struct job_record *) list_next(job_iterator))) {
- if ((job_ptr->job_state & JOB_COMPLETING) == 0)
+ if (((job_ptr->job_state & JOB_COMPLETING) == 0) ||
+ (job_ptr->node_bitmap == NULL))
continue;
- if (job_ptr->nodes_completing) /* no change */
- continue;
- node_name2bitmap(job_ptr->nodes_completing,
- false, &job_ptr->node_bitmap);
+ xfree(job_ptr->nodes_completing);
+ job_ptr->nodes_completing = bitmap2node_name(job_ptr->node_bitmap);
}
list_iterator_destroy(job_iterator);
}
diff --git a/src/slurmctld/node_scheduler.c b/src/slurmctld/node_scheduler.c
index 3aeb4c69e6eb6074ff9c7ef829bdff6bf2bf4c85..20cf46bccef4a17b41a262b220614918ed0c4d80 100644
--- a/src/slurmctld/node_scheduler.c
+++ b/src/slurmctld/node_scheduler.c
@@ -2,7 +2,7 @@
* node_scheduler.c - select and allocated nodes to jobs
* Note: there is a global node table (node_record_table_ptr)
*
- * $Id: node_scheduler.c 12452 2007-10-05 19:07:07Z da $
+ * $Id: node_scheduler.c 12631 2007-11-06 22:48:18Z da $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -137,7 +137,7 @@ extern void allocate_nodes(struct job_record *job_ptr)
* globals: node_record_count - number of nodes configured
* node_record_table_ptr - pointer to global node table
*/
-extern int count_cpus(unsigned *bitmap)
+extern int count_cpus(bitstr_t *bitmap)
{
int i, sum;
diff --git a/src/slurmctld/node_scheduler.h b/src/slurmctld/node_scheduler.h
index 7febc9ad74a628daca6aab9bcf92a0f264911bc9..d44fa4eff0ccf26c3f1ea96459ef5df98cb92130 100644
--- a/src/slurmctld/node_scheduler.h
+++ b/src/slurmctld/node_scheduler.h
@@ -61,7 +61,7 @@ extern void build_node_details(struct job_record *job_ptr);
* globals: node_record_count - number of nodes configured
* node_record_table_ptr - pointer to global node table
*/
-extern int count_cpus(unsigned *bitmap);
+extern int count_cpus(bitstr_t *bitmap);
/*
* deallocate_nodes - for a given job, deallocate its nodes and make
diff --git a/src/slurmctld/read_config.c b/src/slurmctld/read_config.c
index 09617451a21ca7964eb802b1bbae1c4dcd4cf7fb..aca9e9bba65c27f6e9d00a335723c7b0d9dea0ad 100644
--- a/src/slurmctld/read_config.c
+++ b/src/slurmctld/read_config.c
@@ -697,6 +697,12 @@ int read_slurm_conf(int recover)
/* initialization */
START_TIMER;
+ if (recover == 0) {
+ /* in order to re-use job state information,
+ * update nodes_completing string (based on node_bitmap) */
+ update_job_nodes_completing();
+ }
+
/* save node states for reconfig RPC */
old_node_record_count = node_record_count;
old_node_table_ptr = node_record_table_ptr;
@@ -708,11 +714,6 @@ int read_slurm_conf(int recover)
node_record_table_ptr = NULL;
node_record_count = 0;
- if (recover == 0) {
- /* in order to re-use job state information,
- * update nodes_completing string (based on node_bitmap) */
- update_job_nodes_completing();
- }
if ((error_code = _init_all_slurm_conf())) {
node_record_table_ptr = old_node_table_ptr;
return error_code;
@@ -1005,6 +1006,7 @@ static int _sync_nodes_to_comp_job(void)
deallocate_nodes(job_ptr, false, false);
}
}
+ list_iterator_destroy(job_iterator);
if (update_cnt)
info("_sync_nodes_to_comp_job completing %d jobs",
update_cnt);
diff --git a/src/slurmctld/step_mgr.c b/src/slurmctld/step_mgr.c
index 483fbe39a577670db16a3277a829ab5f1868ab75..43aa77897f19ea43a3e85d9bb5c09bafee175639 100644
--- a/src/slurmctld/step_mgr.c
+++ b/src/slurmctld/step_mgr.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* step_mgr.c - manage the job step information of slurm
- * $Id: step_mgr.c 11969 2007-08-08 23:13:12Z da $
+ * $Id: step_mgr.c 12681 2007-11-26 18:56:25Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -701,7 +701,7 @@ step_create(job_step_create_request_msg_t *step_specs,
if (job_ptr == NULL)
return ESLURM_INVALID_JOB_ID ;
- if (job_ptr->job_state == JOB_SUSPENDED)
+ if ((job_ptr->job_state == JOB_SUSPENDED) || IS_JOB_PENDING(job_ptr))
return ESLURM_DISABLED;
if (batch_step) {
@@ -716,9 +716,6 @@ step_create(job_step_create_request_msg_t *step_specs,
(step_specs->user_id != 0))
return ESLURM_ACCESS_DENIED ;
- if (IS_JOB_PENDING(job_ptr))
- return ESLURM_INVALID_JOB_ID ;
-
if (IS_JOB_FINISHED(job_ptr) ||
(job_ptr->end_time <= time(NULL)))
return ESLURM_ALREADY_DONE;
diff --git a/src/slurmd/slurmstepd/mgr.c b/src/slurmd/slurmstepd/mgr.c
index 7d1f3aa16ede8b9fa5d3b16480713de7adc48fe3..0091fadc28f0b3c48fb7fa12437a9a1170f4aadf 100644
--- a/src/slurmd/slurmstepd/mgr.c
+++ b/src/slurmd/slurmstepd/mgr.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* src/slurmd/slurmstepd/mgr.c - job manager functions for slurmstepd
- * $Id: mgr.c 12580 2007-10-29 20:17:09Z jette $
+ * $Id: mgr.c 12647 2007-11-12 17:09:47Z da $
*****************************************************************************
* Copyright (C) 2002-2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -537,7 +537,7 @@ _one_step_complete_msg(slurmd_job_t *job, int first, int last)
for (i = 0; i < REVERSE_TREE_PARENT_RETRY; i++) {
if (i)
sleep(1);
- retcode = slurm_send_recv_rc_msg_only_one(&req, &rc, 10000);
+ retcode = slurm_send_recv_rc_msg_only_one(&req, &rc, 0);
if (retcode == 0 && rc == 0)
goto finished;
}
diff --git a/src/smap/job_functions.c b/src/smap/job_functions.c
index 8536e27b773e7bff9216cc33827217d72ffcbfdf..21eedfc16d9ddb797df147b1db7b014b3a33b572 100644
--- a/src/smap/job_functions.c
+++ b/src/smap/job_functions.c
@@ -251,7 +251,7 @@ static int _print_text_job(job_info_t * job_ptr)
int i = 0;
int width = 0;
char time_buf[20];
- char tmp_cnt[7];
+ char tmp_cnt[8];
uint32_t node_cnt = 0;
char *ionodes = NULL;
time_t now_time = time(NULL);
@@ -271,9 +271,9 @@ static int _print_text_job(job_info_t * job_ptr)
if ((node_cnt == 0) || (node_cnt == NO_VAL))
node_cnt = _get_node_cnt(job_ptr);
#ifdef HAVE_BG
- convert_num_unit((float)node_cnt, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_cnt, tmp_cnt, sizeof(tmp_cnt), UNIT_NONE);
#else
- sprintf(tmp_cnt, "%d", node_cnt);
+ snprintf(tmp_cnt, sizeof(tmp_cnt), "%d", node_cnt);
#endif
if(!params.commandline) {
mvwprintw(text_win, main_ycord,
diff --git a/src/smap/partition_functions.c b/src/smap/partition_functions.c
index 2a04de9abecf8b6bd8706ee5afea83f1b4fb61be..2dd270b00740fe0bfd3aece334267731ad25059b 100644
--- a/src/smap/partition_functions.c
+++ b/src/smap/partition_functions.c
@@ -517,12 +517,13 @@ static int _print_text_part(partition_info_t *part_ptr,
int i = 0;
int width = 0;
char *nodes = NULL, time_buf[20];
- char tmp_cnt[7];
+ char tmp_cnt[8];
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->total_nodes, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_nodes, tmp_cnt,
+ sizeof(tmp_cnt), UNIT_NONE);
#else
- sprintf(tmp_cnt, "%u", part_ptr->total_nodes);
+ snprintf(tmp_cnt, sizeof(tmp_cnt), "%u", part_ptr->total_nodes);
#endif
if(!params.commandline) {
diff --git a/src/squeue/print.c b/src/squeue/print.c
index e43b17f4418d9de54dcf8a699e974dec9aa634e5..ff4c3ffe642c3106e89832bccd8e437f43ca5145 100644
--- a/src/squeue/print.c
+++ b/src/squeue/print.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* print.c - squeue print job functions
- * $Id: print.c 12594 2007-10-31 22:27:56Z jette $
+ * $Id: print.c 12627 2007-11-06 19:48:55Z jette $
*****************************************************************************
* Copyright (C) 2002 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -543,7 +543,7 @@ int _print_job_reason_list(job_info_t * job, int width, bool right,
char* suffix)
{
char *ionodes = NULL;
- char tmp_char[6];
+ char tmp_char[16];
if (job == NULL) { /* Print the Header instead */
#ifdef HAVE_BG
@@ -567,7 +567,8 @@ int _print_job_reason_list(job_info_t * job, int width, bool right,
_print_nodes(job->nodes, width, right, false);
if(ionodes) {
- sprintf(tmp_char, "[%s]", ionodes);
+ snprintf(tmp_char, sizeof(tmp_char), "[%s]",
+ ionodes);
_print_str(tmp_char, width, right, false);
}
}
@@ -600,7 +601,7 @@ int _print_job_node_inx(job_info_t * job, int width, bool right, char* suffix)
int _print_job_num_procs(job_info_t * job, int width, bool right, char* suffix)
{
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("CPUS", width, right, true);
else {
@@ -610,9 +611,11 @@ int _print_job_num_procs(job_info_t * job, int width, bool right, char* suffix)
cnt += job->cpus_per_node[i] *
job->cpu_count_reps[i];
}
- convert_num_unit((float)cnt, tmp_char, UNIT_NONE);
+ convert_num_unit((float)cnt, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
} else {
- convert_num_unit((float)job->num_procs, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->num_procs, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
}
_print_str(tmp_char, width, right, true);
}
@@ -625,7 +628,7 @@ int _print_job_num_nodes(job_info_t * job, int width, bool right_justify,
char* suffix)
{
uint32_t node_cnt = 0;
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("NODES", width, right_justify, true);
@@ -639,9 +642,10 @@ int _print_job_num_nodes(job_info_t * job, int width, bool right_justify,
node_cnt = _get_node_cnt(job);
#ifdef HAVE_BG
- convert_num_unit((float)node_cnt, tmp_char, UNIT_NONE);
+ convert_num_unit((float)node_cnt, tmp_char, sizeof(tmp_char),
+ UNIT_NONE);
#else
- sprintf(tmp_char, "%d", node_cnt);
+ snprintf(tmp_char, sizeof(tmp_char), "%d", node_cnt);
#endif
_print_str(tmp_char, width, right_justify, true);
}
@@ -675,9 +679,12 @@ int _print_job_num_sct(job_info_t * job, int width, bool right_justify,
char threads[10];
char sct[(10+1)*3];
if (job) {
- convert_num_unit((float)job->min_sockets, sockets, UNIT_NONE);
- convert_num_unit((float)job->min_cores, cores, UNIT_NONE);
- convert_num_unit((float)job->min_threads, threads, UNIT_NONE);
+ convert_num_unit((float)job->min_sockets, sockets,
+ sizeof(sockets), UNIT_NONE);
+ convert_num_unit((float)job->min_cores, cores,
+ sizeof(cores), UNIT_NONE);
+ convert_num_unit((float)job->min_threads, threads,
+ sizeof(threads), UNIT_NONE);
sct[0] = '\0';
strcat(sct, sockets);
strcat(sct, ":");
@@ -701,7 +708,8 @@ int _print_job_num_sockets(job_info_t * job, int width, bool right_justify,
if (job == NULL) /* Print the Header instead */
_print_str("SOCKETS", width, right_justify, true);
else {
- convert_num_unit((float)job->min_sockets, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->min_sockets, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -716,7 +724,8 @@ int _print_job_num_cores(job_info_t * job, int width, bool right_justify,
if (job == NULL) /* Print the Header instead */
_print_str("CORES", width, right_justify, true);
else {
- convert_num_unit((float)job->min_cores, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->min_cores, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -731,7 +740,8 @@ int _print_job_num_threads(job_info_t * job, int width, bool right_justify,
if (job == NULL) /* Print the Header instead */
_print_str("THREADS", width, right_justify, true);
else {
- convert_num_unit((float)job->min_threads, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->min_threads, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -789,12 +799,13 @@ int _print_job_contiguous(job_info_t * job, int width, bool right_justify,
int _print_job_min_procs(job_info_t * job, int width, bool right_justify,
char* suffix)
{
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("MIN_PROCS", width, right_justify, true);
else {
- convert_num_unit((float)job->job_min_procs, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->job_min_procs, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -805,12 +816,13 @@ int _print_job_min_procs(job_info_t * job, int width, bool right_justify,
int _print_job_min_sockets(job_info_t * job, int width, bool right_justify,
char* suffix)
{
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("MIN_SOCKETS", width, right_justify, true);
else {
- convert_num_unit((float)job->job_min_sockets, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->job_min_sockets, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -821,12 +833,13 @@ int _print_job_min_sockets(job_info_t * job, int width, bool right_justify,
int _print_job_min_cores(job_info_t * job, int width, bool right_justify,
char* suffix)
{
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("MIN_CORES", width, right_justify, true);
else {
- convert_num_unit((float)job->job_min_cores, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->job_min_cores, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -837,12 +850,13 @@ int _print_job_min_cores(job_info_t * job, int width, bool right_justify,
int _print_job_min_threads(job_info_t * job, int width, bool right_justify,
char* suffix)
{
- char tmp_char[6];
+ char tmp_char[8];
if (job == NULL) /* Print the Header instead */
_print_str("MIN_THREADS", width, right_justify, true);
else {
- convert_num_unit((float)job->job_min_threads, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job->job_min_threads, tmp_char,
+ sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
if (suffix)
@@ -862,11 +876,13 @@ int _print_job_min_memory(job_info_t * job, int width, bool right_justify,
else {
tmp_char[0] = '\0';
if (job->job_max_memory < job->job_min_memory) {
- convert_num_unit((float)job->job_max_memory, max_mem, UNIT_NONE);
+ convert_num_unit((float)job->job_max_memory, max_mem,
+ sizeof(max_mem), UNIT_NONE);
strcat(tmp_char, max_mem);
strcat(tmp_char, "-");
- }
- convert_num_unit((float)job->job_min_memory, min_mem, UNIT_NONE);
+ }
+ convert_num_unit((float)job->job_min_memory, min_mem,
+ sizeof(min_mem), UNIT_NONE);
strcat(tmp_char, min_mem);
_print_str(tmp_char, width, right_justify, true);
}
@@ -880,13 +896,13 @@ int
_print_job_min_tmp_disk(job_info_t * job, int width, bool right_justify,
char* suffix)
{
- char tmp_char[6];
+ char tmp_char[10];
if (job == NULL) /* Print the Header instead */
_print_str("MIN_TMP_DISK", width, right_justify, true);
else {
convert_num_unit((float)job->job_min_tmp_disk,
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
_print_str(tmp_char, width, right_justify, true);
}
diff --git a/src/srun/allocate.c b/src/srun/allocate.c
index 92ea573ee6a1ae677678bc1939909717042c775a..a06ea95966d6af57affd4ec8dfbbc3c5fa0b218f 100644
--- a/src/srun/allocate.c
+++ b/src/srun/allocate.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* src/srun/allocate.c - srun functions for managing node allocations
- * $Id: allocate.c 12574 2007-10-26 17:00:52Z jette $
+ * $Id: allocate.c 12700 2007-11-27 23:39:24Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -363,8 +363,7 @@ static bool
_retry()
{
static int retries = 0;
- static char *msg = "Slurm controller not responding, "
- "sleeping and retrying.";
+ static char *msg = "Slurm job queue full, sleeping and retrying.";
if (errno == ESLURM_ERROR_ON_DESC_TO_RECORD_COPY) {
if (retries == 0)
@@ -558,13 +557,14 @@ job_desc_msg_create_from_opts (char *script)
xassert (opt.batch);
j->environment = NULL;
- if (opt.get_user_env >= 0) {
+ if (opt.get_user_env_time >= 0) {
struct passwd *pw = NULL;
pw = getpwuid(opt.uid);
if (pw != NULL) {
j->environment =
env_array_user_default(pw->pw_name,
- opt.get_user_env);
+ opt.get_user_env_time,
+ opt.get_user_env_mode);
/* FIXME - should we abort if j->environment
is NULL? */
}
diff --git a/src/srun/opt.c b/src/srun/opt.c
index 17a78067d998f589dfc12563de3b4931cb2d040d..e00bf4f232e8ae13d2daed56fc06f40459eb3e54 100644
--- a/src/srun/opt.c
+++ b/src/srun/opt.c
@@ -1,6 +1,6 @@
/*****************************************************************************\
* opt.c - options processing for srun
- * $Id: opt.c 12583 2007-10-30 17:01:31Z jette $
+ * $Id: opt.c 12711 2007-11-29 00:04:01Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -78,7 +78,6 @@
#include "src/common/xmalloc.h"
#include "src/common/xstring.h"
#include "src/common/slurm_rlimits_info.h"
-#include "src/common/parse_time.h"
#include "src/common/plugstack.h"
#include "src/common/optz.h"
#include "src/api/pmi_server.h"
@@ -193,8 +192,8 @@ static void _opt_default(void);
/* set options based upon env vars */
static void _opt_env(void);
-
static void _opt_args(int argc, char **argv);
+static void _proc_get_user_env(char *optarg);
/* list known options and their settings */
static void _opt_list(void);
@@ -1045,7 +1044,8 @@ static void _opt_default()
opt.msg_timeout = 15;
}
- opt.get_user_env = -1;
+ opt.get_user_env_time = -1;
+ opt.get_user_env_mode = -1;
}
/*---[ env var processing ]-----------------------------------------------*/
@@ -1884,6 +1884,10 @@ void set_options(const int argc, char **argv, int first)
break;
case LONG_OPT_BEGIN:
opt.begin = parse_time(optarg);
+ if (opt.begin == 0) {
+ fatal("Invalid time specification %s",
+ optarg);
+ }
break;
case LONG_OPT_MAIL_TYPE:
opt.mail_type |= _parse_mail_type(optarg);
@@ -2006,9 +2010,9 @@ void set_options(const int argc, char **argv, int first)
break;
case LONG_OPT_GET_USER_ENV:
if (optarg)
- opt.get_user_env = strtol(optarg, NULL, 10);
+ _proc_get_user_env(optarg);
else
- opt.get_user_env = 0;
+ opt.get_user_env_time = 0;
break;
default:
if (spank_process_option (opt_char, optarg) < 0) {
@@ -2027,6 +2031,25 @@ void set_options(const int argc, char **argv, int first)
spank_option_table_destroy (optz);
}
+static void _proc_get_user_env(char *optarg)
+{
+ char *end_ptr;
+
+ if ((optarg[0] >= '0') && (optarg[0] <= '9'))
+ opt.get_user_env_time = strtol(optarg, &end_ptr, 10);
+ else {
+ opt.get_user_env_time = 0;
+ end_ptr = optarg;
+ }
+
+ if ((end_ptr == NULL) || (end_ptr[0] == '\0'))
+ return;
+ if ((end_ptr[0] == 's') || (end_ptr[0] == 'S'))
+ opt.get_user_env_mode = 1;
+ else if ((end_ptr[0] == 'l') || (end_ptr[0] == 'L'))
+ opt.get_user_env_mode = 2;
+}
+
/* Load the multi_prog config file into argv, pass the entire file contents
* in order to avoid having to read the file on every node. We could parse
* the infomration here too for loading the MPIR records for TotalView */
diff --git a/src/srun/opt.h b/src/srun/opt.h
index 1bd844e00b1c1178f87cacfeda0ab7ec3a9a83ae..e7e67e2b7c397f529ec5c13800f574154a724e8c 100644
--- a/src/srun/opt.h
+++ b/src/srun/opt.h
@@ -1,6 +1,6 @@
/*****************************************************************************\
* opt.h - definitions for srun option processing
- * $Id: opt.h 12574 2007-10-26 17:00:52Z jette $
+ * $Id: opt.h 12697 2007-11-27 22:02:29Z jette $
*****************************************************************************
* Copyright (C) 2002-2006 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
@@ -212,7 +212,8 @@ typedef struct srun_options {
uint16_t mail_type; /* --mail-type */
char *mail_user; /* --mail-user */
char *ctrl_comm_ifhn; /* --ctrl-comm-ifhn */
- int get_user_env; /* --get-user-env[=secs] */
+ int get_user_env_time; /* --get-user-env[=secs] */
+ int get_user_env_mode; /* --get-user-env=[S|L] */
} opt_t;
extern opt_t opt;
diff --git a/src/srun/srun.c b/src/srun/srun.c
index 4fb3f0b001c143e2a824ee06b0717be9f7fb1770..82b01cd39ad14b1d60739fed76384b7f491f70d9 100644
--- a/src/srun/srun.c
+++ b/src/srun/srun.c
@@ -597,12 +597,12 @@ _print_job_information(resource_allocation_response_msg_t *resp)
static int
_run_batch_job(const char *argv0)
{
- int file_type, retries;
+ int file_type, retries = 0;
int rc = SLURM_SUCCESS;
job_desc_msg_t *req;
submit_response_msg_t *resp;
char *script;
- void (*log_msg) (const char *fmt, ...) = (void (*)) &error;
+ static char *msg = "Slurm job queue full, sleeping and retrying.";
if ((remote_argc == 0) || (remote_argv[0] == NULL))
return SLURM_ERROR;
@@ -629,15 +629,15 @@ _run_batch_job(const char *argv0)
if (!opt.jobid_set)
req->job_id = NO_VAL;
- retries = 0;
- while ( (retries < MAX_RETRIES)
- && (rc = slurm_submit_batch_job(req, &resp)) < 0) {
-
- if (errno != ESLURM_ERROR_ON_DESC_TO_RECORD_COPY)
+ while ((rc = slurm_submit_batch_job(req, &resp)) < 0) {
+ if ((errno != ESLURM_ERROR_ON_DESC_TO_RECORD_COPY) ||
+ (retries >= MAX_RETRIES))
return (error("Unable to submit batch job: %m"));
- (*log_msg) ("Controller not responding, retrying...");
- log_msg = &debug;
+ if (retries == 0)
+ error(msg);
+ else
+ debug(msg);
sleep (++retries);
}
diff --git a/src/sview/block_info.c b/src/sview/block_info.c
index 47654ea7f6defd63557a2143ad8bb4359ee0001c..ee1afa88e4063eaec4d3fdbedc030eb268244fc9 100644
--- a/src/sview/block_info.c
+++ b/src/sview/block_info.c
@@ -207,7 +207,7 @@ static void _layout_block_record(GtkTreeView *treeview,
sview_block_info_t *block_ptr,
int update)
{
- char tmp_cnt[7];
+ char tmp_cnt[8];
GtkTreeIter iter;
GtkTreeStore *treestore =
GTK_TREE_STORE(gtk_tree_view_get_model(treeview));
@@ -238,7 +238,8 @@ static void _layout_block_record(GtkTreeView *treeview,
SORTID_USE),
_convert_node_use(block_ptr->bg_node_use));
- convert_num_unit((float)block_ptr->node_cnt, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)block_ptr->node_cnt, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_block,
SORTID_NODES),
@@ -269,7 +270,7 @@ static void _layout_block_record(GtkTreeView *treeview,
static void _update_block_record(sview_block_info_t *block_ptr,
GtkTreeStore *treestore, GtkTreeIter *iter)
{
- char tmp_cnt[7];
+ char tmp_cnt[8];
gtk_tree_store_set(treestore, iter, SORTID_BLOCK,
block_ptr->bg_block_name, -1);
@@ -284,7 +285,8 @@ static void _update_block_record(sview_block_info_t *block_ptr,
gtk_tree_store_set(treestore, iter, SORTID_USE,
_convert_node_use(block_ptr->bg_node_use), -1);
- convert_num_unit((float)block_ptr->node_cnt, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)block_ptr->node_cnt, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
gtk_tree_store_set(treestore, iter, SORTID_NODES, tmp_cnt, -1);
gtk_tree_store_set(treestore, iter, SORTID_NODELIST,
@@ -350,7 +352,7 @@ static void _update_info_block(List block_list,
if(block_ptr->node_cnt == 0)
block_ptr->node_cnt = block_ptr->size;
if(!block_ptr->slurm_part_name)
- block_ptr->slurm_part_name = "no part";
+ block_ptr->slurm_part_name = xstrdup("no part");
/* get the iter, or find out the list is empty goto add */
if (!gtk_tree_model_get_iter(model, &iter, path)) {
diff --git a/src/sview/job_info.c b/src/sview/job_info.c
index e07749e1d453ef76c9aa0b58d3aa5d26535ec762..2bbc43be2e2515d27974940737e6aea9107ea35d 100644
--- a/src/sview/job_info.c
+++ b/src/sview/job_info.c
@@ -1188,9 +1188,10 @@ static void _layout_job_record(GtkTreeView *treeview,
#ifdef HAVE_BG
convert_num_unit((float)sview_job_info_ptr->node_cnt,
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#else
- sprintf(tmp_char, "%u", sview_job_info_ptr->node_cnt);
+ snprintf(tmp_char, sizeof(tmp_char), "%u",
+ sview_job_info_ptr->node_cnt);
#endif
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_job,
@@ -1198,9 +1199,10 @@ static void _layout_job_record(GtkTreeView *treeview,
tmp_char);
#ifdef HAVE_BG
- convert_num_unit((float)job_ptr->num_procs, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job_ptr->num_procs, tmp_char, sizeof(tmp_char),
+ UNIT_NONE);
#else
- sprintf(tmp_char, "%u", job_ptr->num_procs);
+ snprintf(tmp_char, sizeof(tmp_char), "%u", job_ptr->num_procs);
#endif
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_job,
@@ -1592,7 +1594,7 @@ static void _update_job_record(sview_job_info_t *sview_job_info_ptr,
#ifdef HAVE_BG
convert_num_unit((float)sview_job_info_ptr->node_cnt,
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#else
sprintf(tmp_char, "%u", sview_job_info_ptr->node_cnt);
#endif
@@ -1600,9 +1602,10 @@ static void _update_job_record(sview_job_info_t *sview_job_info_ptr,
SORTID_NODES, tmp_char, -1);
#ifdef HAVE_BG
- convert_num_unit((float)job_ptr->num_procs, tmp_char, UNIT_NONE);
+ convert_num_unit((float)job_ptr->num_procs, tmp_char, sizeof(tmp_char),
+ UNIT_NONE);
#else
- sprintf(tmp_char, "%u", job_ptr->num_procs);
+ snprintf(tmp_char, sizeof(tmp_char), "%u", job_ptr->num_procs);
#endif
gtk_tree_store_set(treestore, iter,
SORTID_NUM_PROCS, tmp_char, -1);
@@ -1643,57 +1646,43 @@ static void _update_job_record(sview_job_info_t *sview_job_info_ptr,
SORTID_MAX_NODES, tmp_char, -1);
}
if(job_ptr->cpus_per_task > 0) {
- //convert_num_unit((float)job_ptr->cpus_per_task,
- // tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->cpus_per_task);
gtk_tree_store_set(treestore, iter,
SORTID_CPUS_PER_TASK, tmp_char, -1);
}
- //convert_num_unit((float)job_ptr->job_min_procs, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->job_min_procs);
gtk_tree_store_set(treestore, iter,
SORTID_REQ_PROCS, tmp_char, -1);
- //convert_num_unit((float)job_ptr->min_sockets, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->min_sockets);
gtk_tree_store_set(treestore, iter,
SORTID_MIN_SOCKETS, tmp_char, -1);
- //convert_num_unit((float)job_ptr->max_sockets, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->max_sockets);
gtk_tree_store_set(treestore, iter,
SORTID_MAX_SOCKETS, tmp_char, -1);
- //convert_num_unit((float)job_ptr->min_cores, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->min_cores);
gtk_tree_store_set(treestore, iter,
SORTID_MIN_CORES, tmp_char, -1);
- //convert_num_unit((float)job_ptr->max_cores, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->max_cores);
gtk_tree_store_set(treestore, iter,
SORTID_MAX_CORES, tmp_char, -1);
- //convert_num_unit((float)job_ptr->min_threads, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->min_threads);
gtk_tree_store_set(treestore, iter,
SORTID_MIN_THREADS, tmp_char, -1);
- //convert_num_unit((float)job_ptr->max_threads, tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->max_threads);
gtk_tree_store_set(treestore, iter,
SORTID_MAX_THREADS, tmp_char, -1);
- //convert_num_unit((float)job_ptr->job_min_memory,
- // tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->job_min_memory);
gtk_tree_store_set(treestore, iter,
SORTID_MIN_MEM, tmp_char, -1);
- //convert_num_unit((float)job_ptr->job_max_memory,
- // tmp_char, UNIT_NONE);
+
sprintf(tmp_char, "%u", job_ptr->job_max_memory);
gtk_tree_store_set(treestore, iter,
SORTID_MAX_MEM, tmp_char, -1);
- //convert_num_unit((float)job_ptr->job_min_tmp_disk,
- // tmp_char, UNIT_NONE);
sprintf(tmp_char, "%u", job_ptr->job_min_tmp_disk);
gtk_tree_store_set(treestore, iter,
SORTID_TMP_DISK, tmp_char, -1);
@@ -1767,10 +1756,10 @@ static void _layout_step_record(GtkTreeView *treeview,
nodes = step_ptr->nodes;
#ifdef HAVE_BG
convert_num_unit((float)step_ptr->num_tasks,
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#else
convert_num_unit((float)_nodes_in_list(nodes),
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#endif
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_job,
@@ -1818,7 +1807,8 @@ static void _layout_step_record(GtkTreeView *treeview,
SORTID_NAME),
step_ptr->name);
- convert_num_unit((float)step_ptr->num_tasks, tmp_char, UNIT_NONE);
+ convert_num_unit((float)step_ptr->num_tasks, tmp_char, sizeof(tmp_char),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_job,
SORTID_TASKS),
@@ -1857,10 +1847,10 @@ static void _update_step_record(job_step_info_t *step_ptr,
nodes = step_ptr->nodes;
#ifdef HAVE_BG
convert_num_unit((float)step_ptr->num_tasks,
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#else
convert_num_unit((float)_nodes_in_list(nodes),
- tmp_char, UNIT_NONE);
+ tmp_char, sizeof(tmp_char), UNIT_NONE);
#endif
gtk_tree_store_set(treestore, iter,
SORTID_NODES, tmp_char, -1);
@@ -1894,7 +1884,8 @@ static void _update_step_record(job_step_info_t *step_ptr,
gtk_tree_store_set(treestore, iter,
SORTID_NAME, step_ptr->name, -1);
- convert_num_unit((float)step_ptr->num_tasks, tmp_char, UNIT_NONE);
+ convert_num_unit((float)step_ptr->num_tasks, tmp_char, sizeof(tmp_char),
+ UNIT_NONE);
gtk_tree_store_set(treestore, iter,
SORTID_TASKS, tmp_char, -1);
diff --git a/src/sview/node_info.c b/src/sview/node_info.c
index 6348522e5736d1c288e826bb06c78b2027c0176b..2330f0f8f71ece64c86c29da46cb8e2517bced10 100644
--- a/src/sview/node_info.c
+++ b/src/sview/node_info.c
@@ -137,43 +137,50 @@ static void _layout_node_record(GtkTreeView *treeview,
lower);
xfree(lower);
- convert_num_unit((float)node_ptr->cpus, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_ptr->cpus, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_CPUS),
tmp_cnt);
- convert_num_unit((float)node_ptr->used_cpus, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_ptr->used_cpus, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_USED_CPUS),
tmp_cnt);
- convert_num_unit((float)node_ptr->cores, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_ptr->cores, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_CORES),
tmp_cnt);
- convert_num_unit((float)node_ptr->sockets, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_ptr->sockets, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_SOCKETS),
tmp_cnt);
- convert_num_unit((float)node_ptr->threads, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)node_ptr->threads, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_THREADS),
tmp_cnt);
- convert_num_unit((float)node_ptr->real_memory, tmp_cnt, UNIT_MEGA);
+ convert_num_unit((float)node_ptr->real_memory, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_MEGA);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_MEMORY),
tmp_cnt);
- convert_num_unit((float)node_ptr->tmp_disk, tmp_cnt, UNIT_MEGA);
+ convert_num_unit((float)node_ptr->tmp_disk, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_MEGA);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_node,
SORTID_DISK),
@@ -217,9 +224,11 @@ static void _update_node_record(node_info_t *node_ptr,
node_ptr->sockets, -1);
gtk_tree_store_set(treestore, iter, SORTID_THREADS,
node_ptr->threads, -1);
- convert_num_unit((float)node_ptr->real_memory, tmp_cnt, UNIT_MEGA);
+ convert_num_unit((float)node_ptr->real_memory, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_MEGA);
gtk_tree_store_set(treestore, iter, SORTID_MEMORY, tmp_cnt, -1);
- convert_num_unit((float)node_ptr->tmp_disk, tmp_cnt, UNIT_MEGA);
+ convert_num_unit((float)node_ptr->tmp_disk, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_MEGA);
gtk_tree_store_set(treestore, iter, SORTID_DISK, tmp_cnt, -1);
gtk_tree_store_set(treestore, iter, SORTID_WEIGHT,
node_ptr->weight, -1);
diff --git a/src/sview/part_info.c b/src/sview/part_info.c
index 369da9dda97146eb16b01ab185846c95fa976e9d..48c274ea1ea0330d8c7a3d84df225f31f8202a14 100644
--- a/src/sview/part_info.c
+++ b/src/sview/part_info.c
@@ -258,11 +258,13 @@ static int
_build_min_max_16_string(char *buffer, int buf_size,
uint16_t min, uint16_t max, bool range)
{
- char tmp_min[7];
- char tmp_max[7];
- convert_num_unit((float)min, tmp_min, UNIT_NONE);
- if(max != (uint16_t) INFINITE)
- convert_num_unit((float)max, tmp_max, UNIT_NONE);
+ char tmp_min[8];
+ char tmp_max[8];
+ convert_num_unit((float)min, tmp_min, sizeof(tmp_min), UNIT_NONE);
+ if(max != (uint16_t) INFINITE) {
+ convert_num_unit((float)max, tmp_max, sizeof(tmp_max),
+ UNIT_NONE);
+ }
if (max == min)
return snprintf(buffer, buf_size, "%s", tmp_max);
@@ -281,10 +283,10 @@ static int
_build_min_max_32_string(char *buffer, int buf_size,
uint32_t min, uint32_t max, bool range)
{
- char tmp_min[7];
- char tmp_max[7];
- convert_num_unit((float)min, tmp_min, UNIT_NONE);
- convert_num_unit((float)max, tmp_max, UNIT_NONE);
+ char tmp_min[8];
+ char tmp_max[8];
+ convert_num_unit((float)min, tmp_min, sizeof(tmp_min), UNIT_NONE);
+ convert_num_unit((float)max, tmp_max, sizeof(tmp_max), UNIT_NONE);
if (max == min)
return snprintf(buffer, buf_size, "%s", tmp_max);
@@ -721,9 +723,9 @@ static void _layout_part_record(GtkTreeView *treeview,
GtkTreeIter iter;
ListIterator itr = NULL;
char time_buf[20];
- char tmp_cnt[7];
- char tmp_cnt1[7];
- char tmp_cnt2[7];
+ char tmp_cnt[8];
+ char tmp_cnt1[8];
+ char tmp_cnt2[8];
partition_info_t *part_ptr = sview_part_info->part_ptr;
sview_part_sub_t *sview_part_sub = NULL;
sview_part_sub_t *temp_part_sub = NULL;
@@ -797,7 +799,7 @@ static void _layout_part_record(GtkTreeView *treeview,
snprintf(time_buf, sizeof(time_buf), "infinite");
else {
convert_num_unit((float)part_ptr->min_nodes,
- time_buf, UNIT_NONE);
+ time_buf, sizeof(time_buf), UNIT_NONE);
}
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_part,
@@ -807,7 +809,7 @@ static void _layout_part_record(GtkTreeView *treeview,
snprintf(time_buf, sizeof(time_buf), "infinite");
else {
convert_num_unit((float)part_ptr->max_nodes,
- time_buf, UNIT_NONE);
+ time_buf, sizeof(time_buf), UNIT_NONE);
}
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_part,
@@ -844,7 +846,8 @@ static void _layout_part_record(GtkTreeView *treeview,
temp_char);
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->total_nodes, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_nodes, tmp_cnt,
+ sizeof(tmp_cnt), UNIT_NONE);
#else
sprintf(tmp_cnt, "%u", part_ptr->total_nodes);
#endif
@@ -852,7 +855,8 @@ static void _layout_part_record(GtkTreeView *treeview,
find_col_name(display_data_part,
SORTID_NODES),
tmp_cnt);
- convert_num_unit((float)part_ptr->total_cpus, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_cpus, tmp_cnt, sizeof(tmp_cnt),
+ UNIT_NONE);
add_display_treestore_line(update, treestore, &iter,
find_col_name(display_data_part,
SORTID_CPUS),
@@ -891,9 +895,12 @@ static void _layout_part_record(GtkTreeView *treeview,
other_part_sub.reason = temp_part_sub->reason;
}
}
- convert_num_unit((float)alloc_part_sub.node_cnt, tmp_cnt, UNIT_NONE);
- convert_num_unit((float)idle_part_sub.node_cnt, tmp_cnt1, UNIT_NONE);
- convert_num_unit((float)other_part_sub.node_cnt, tmp_cnt2, UNIT_NONE);
+ convert_num_unit((float)alloc_part_sub.node_cnt,
+ tmp_cnt, sizeof(tmp_cnt), UNIT_NONE);
+ convert_num_unit((float)idle_part_sub.node_cnt,
+ tmp_cnt1, sizeof(tmp_cnt1), UNIT_NONE);
+ convert_num_unit((float)other_part_sub.node_cnt,
+ tmp_cnt2, sizeof(tmp_cnt2), UNIT_NONE);
snprintf(tmp, sizeof(tmp), "%s/%s/%s",
tmp_cnt, tmp_cnt1, tmp_cnt2);
add_display_treestore_line(update, treestore, &iter,
@@ -915,7 +922,7 @@ static void _update_part_record(sview_part_info_t *sview_part_info,
GtkTreeIter *iter)
{
char time_buf[20];
- char tmp_cnt[7];
+ char tmp_cnt[8];
char *temp_char = NULL;
partition_info_t *part_ptr = sview_part_info->part_ptr;
GtkTreeIter sub_iter;
@@ -959,7 +966,7 @@ static void _update_part_record(sview_part_info_t *sview_part_info,
snprintf(time_buf, sizeof(time_buf), "infinite");
else {
convert_num_unit((float)part_ptr->min_nodes,
- time_buf, UNIT_NONE);
+ time_buf, sizeof(time_buf), UNIT_NONE);
}
gtk_tree_store_set(treestore, iter, SORTID_MIN_NODES,
time_buf, -1);
@@ -967,7 +974,7 @@ static void _update_part_record(sview_part_info_t *sview_part_info,
snprintf(time_buf, sizeof(time_buf), "infinite");
else {
convert_num_unit((float)part_ptr->max_nodes,
- time_buf, UNIT_NONE);
+ time_buf, sizeof(time_buf), UNIT_NONE);
}
gtk_tree_store_set(treestore, iter, SORTID_MAX_NODES,
time_buf, -1);
@@ -993,7 +1000,8 @@ static void _update_part_record(sview_part_info_t *sview_part_info,
gtk_tree_store_set(treestore, iter, SORTID_GROUPS, temp_char, -1);
#ifdef HAVE_BG
- convert_num_unit((float)part_ptr->total_nodes, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)part_ptr->total_nodes, tmp_cnt,
+ sizeof(tmp_cnt), UNIT_NONE);
#else
sprintf(tmp_cnt, "%u", part_ptr->total_nodes);
#endif
@@ -1031,7 +1039,7 @@ static void _update_part_sub_record(sview_part_sub_t *sview_part_sub,
GtkTreeStore *treestore, GtkTreeIter *iter)
{
char time_buf[20];
- char tmp_cnt[7];
+ char tmp_cnt[8];
partition_info_t *part_ptr = sview_part_sub->part_ptr;
char *upper = NULL, *lower = NULL;
char tmp[MAXHOSTRANGELEN];
@@ -1067,7 +1075,8 @@ static void _update_part_sub_record(sview_part_sub_t *sview_part_sub,
sview_part_sub->max_weight, false);
gtk_tree_store_set(treestore, iter, SORTID_WEIGHT, time_buf, -1);
- convert_num_unit((float)sview_part_sub->node_cnt, tmp_cnt, UNIT_NONE);
+ convert_num_unit((float)sview_part_sub->node_cnt, tmp_cnt,
+ sizeof(tmp_cnt), UNIT_NONE);
gtk_tree_store_set(treestore, iter, SORTID_NODES, tmp_cnt, -1);
hostlist_ranged_string(sview_part_sub->hl, sizeof(tmp), tmp);
diff --git a/testsuite/expect/test1.32 b/testsuite/expect/test1.32
index da2d9bd348e365fd0197bb4de544c289ff61e96b..f833cb4ec90bb06e6b684560ba80795ae2e41fb7 100755
--- a/testsuite/expect/test1.32
+++ b/testsuite/expect/test1.32
@@ -88,12 +88,12 @@ expect {
incr matches
exp_continue
}
- -re "SIGUSR1" {
- set usr1cnt [expr $usr1cnt + 1]
- exp_continue
- }
- -re "SIGUSR2" {
- set usr2cnt [expr $usr2cnt + 1]
+ -re "SIGUSR($number)" {
+ if {$expect_out(1,string) == 1} {
+ set usr1cnt [expr $usr1cnt + 1]
+ } else {
+ set usr2cnt [expr $usr2cnt + 1]
+ }
exp_continue
}
-re "error.*not running" {
diff --git a/testsuite/expect/test10.5 b/testsuite/expect/test10.5
index 349aade2d14f509afd8f6e9a7f4fd154ae23c1e7..d48da0b53189df911b0a2e722386a93d451d413a 100755
--- a/testsuite/expect/test10.5
+++ b/testsuite/expect/test10.5
@@ -35,7 +35,7 @@ source ./globals
set test_id "10.5"
set exit_code 0
set matches 0
-set non_bg 0
+set non_bg 0
set too_small 0
print_header $test_id
@@ -97,7 +97,16 @@ expect {
send "q"
exp_continue
}
-
+ -re "error" {
+ send_user "\nFAILURE: smap error\n"
+ set exit_code 1
+ exp_continue
+ }
+ -re "fatal" {
+ send_user "\nFAILURE: smap error\n"
+ set exit_code 1
+ exp_continue
+ }
timeout {
send_user "\nFAILURE: smap not responding\n"
set exit_code 1
diff --git a/testsuite/expect/test10.6 b/testsuite/expect/test10.6
index 48388e84a2179ae21c92e11fb38b6a86bfd3126d..038e9331be98d992465eef326637926f15e54b76 100755
--- a/testsuite/expect/test10.6
+++ b/testsuite/expect/test10.6
@@ -86,7 +86,16 @@ expect {
incr matches
exp_continue
}
-
+ -re "error" {
+ send_user "\nFAILURE: smap error\n"
+ set exit_code 1
+ exp_continue
+ }
+ -re "fatal" {
+ send_user "\nFAILURE: smap error\n"
+ set exit_code 1
+ exp_continue
+ }
timeout {
send_user "\nFAILURE: smap not responding\n"
set exit_code 1
diff --git a/testsuite/expect/test17.32 b/testsuite/expect/test17.32
index b26af694904dab4fbdf46cd997d6cde1b9ddd5ae..0a408b0197b6c9fa48cf84e837c68c3d92d76653 100755
--- a/testsuite/expect/test17.32
+++ b/testsuite/expect/test17.32
@@ -51,7 +51,7 @@ make_bash_script $file_in "
#
# Submit a slurm job that will execute 'id' on 1 node and over task_cnt tasks
#
-spawn $sbatch --tasks=$task_cnt --overcommit -N1 --output=$file_out -t1 $file_in
+spawn $sbatch --ntasks=$task_cnt --overcommit -N1 --output=$file_out -t1 $file_in
expect {
-re "Submitted batch job ($number)" {
set job_id $expect_out(1,string)
diff --git a/testsuite/expect/test17.6 b/testsuite/expect/test17.6
index 2d845385d160f5a65b47092639fbdaee2ef6fd80..ef1df747e2c0396d6f277358b8f2983bbcbcd9fb 100755
--- a/testsuite/expect/test17.6
+++ b/testsuite/expect/test17.6
@@ -42,13 +42,13 @@ set tasks 0
print_header $test_id
#
-# Submit a slurm job that will execute 'id' on 4 tasks (or try anyway)
+# Submit a slurm job that will execute 'id' on $task_cnt tasks (or try anyway)
#
file delete $file_in $file_out
make_bash_script $file_in "$slaunch $bin_id"
set job_id 0
set no_run 0
-spawn $sbatch --tasks=$task_cnt --output=$file_out -t1 $file_in
+spawn $sbatch --ntasks=$task_cnt --output=$file_out -t1 $file_in
expect {
-re "Submitted batch job ($number)" {
set job_id $expect_out(1,string)
@@ -107,7 +107,7 @@ file delete $file_in $file_out
make_bash_script $file_in "$srun -n $task_cnt --overcommit $bin_id"
set job_id 0
set tasks 0
-spawn $sbatch --tasks=1 --output=$file_out -t1 $file_in
+spawn $sbatch --ntasks=1 --output=$file_out -t1 $file_in
expect {
-re "Submitted batch job ($number)" {
set job_id $expect_out(1,string)
diff --git a/testsuite/expect/test7.8 b/testsuite/expect/test7.8
index 61e8c9d9da59ad90934041640eb9b128fc81ac5d..8f5c0a134c203c43b46b3566700a5983c277bdda 100755
--- a/testsuite/expect/test7.8
+++ b/testsuite/expect/test7.8
@@ -94,6 +94,14 @@ if {$sched_port == 0} {
exit 1
}
+if { [test_bluegene] } {
+ set is_bluegene 1
+ set node_cnt "1-1024"
+} else {
+ set is_bluegene 0
+ set node_cnt "1-2"
+}
+
#
# Submit two jobs to work with
#
@@ -106,15 +114,15 @@ make_bash_script $file_in "
set job_id1 0
set job_id2 0
-set srun_pid [spawn $srun -N1-6 --output=$file_out --comment=test --batch -t1 $file_in]
+set sbatch_pid [spawn $sbatch -N $node_cnt --output=$file_out --comment=test -t1 $file_in]
expect {
- -re "jobid ($number) submitted" {
+ -re "Submitted batch job ($number)" {
set job_id1 $expect_out(1,string)
exp_continue
}
timeout {
- send_user "\nFAILURE: srun not responding\n"
- catch {exec $bin_kill -KILL $srun_pid}
+ send_user "\nFAILURE: sbatch not responding\n"
+ catch {exec $bin_kill -KILL $sbatch_pid}
set exit_code 1
}
eof {
@@ -126,15 +134,15 @@ if {$job_id1 == 0} {
exit 1
}
-set srun_pid [spawn $srun -N1-6 --output=$file_out --comment=test --batch -t1 $file_in]
+set sbatch_pid [spawn $sbatch -N $node_cnt --output=$file_out --comment=test -t1 $file_in]
expect {
- -re "jobid ($number) submitted" {
+ -re "Submitted batch job ($number)" {
set job_id2 $expect_out(1,string)
exp_continue
}
timeout {
- send_user "\nFAILURE: srun not responding\n"
- catch {exec $bin_kill -KILL $srun_pid}
+ send_user "\nFAILURE: sbatch not responding\n"
+ catch {exec $bin_kill -KILL $sbatch_pid}
set exit_code 1
}
eof {
@@ -154,7 +162,7 @@ if {$job_id2 == 0} {
exec $bin_rm -f $test_prog
exec $bin_make -f /dev/null $test_prog
set success 0
-set moab_pid [spawn $test_prog $control_addr $job_id1 $job_id2 $sched_port]
+set moab_pid [spawn $test_prog $control_addr $job_id1 $job_id2 $sched_port $is_bluegene]
expect {
-re "SUCCESS" {
set success 1
diff --git a/testsuite/expect/test7.8.prog.c b/testsuite/expect/test7.8.prog.c
index 046ab7f65b6e9bd7de12aedbedf4ae9af1af468a..d88ca12450c9d535e0d57d44872910796533feee 100644
--- a/testsuite/expect/test7.8.prog.c
+++ b/testsuite/expect/test7.8.prog.c
@@ -33,7 +33,7 @@
/* global variables */
char *control_addr;
-int sched_port;
+int is_bluegene, sched_port;
long job_id1, job_id2;
static int _conn_wiki_port(char *host, int port)
@@ -202,6 +202,33 @@ static void _cancel_job(long my_job_id)
_xmit(out_msg);
}
+static void _modify_job(long my_job_id)
+{
+ time_t now = time(NULL);
+ char out_msg[256];
+
+ snprintf(out_msg, sizeof(out_msg),
+ "TS=%u AUTH=root DT=CMD=MODIFYJOB ARG=%ld "
+ /* "PARTITION=pdebug " */
+ /* "NODES=2 " */
+ /* "DEPEND=afterany:3 " */
+ /* "INVALID=123 " */
+ "TIMELIMIT=10 BANK=test_bank",
+ (uint32_t) now, my_job_id);
+ _xmit(out_msg);
+}
+
+static void _resume_job(long my_job_id)
+{
+ time_t now = time(NULL);
+ char out_msg[128];
+
+ snprintf(out_msg, sizeof(out_msg),
+ "TS=%u AUTH=root DT=CMD=RESUMEJOB ARG=%ld",
+ (uint32_t) now, my_job_id);
+ _xmit(out_msg);
+}
+
static void _start_job(long my_job_id)
{
time_t now = time(NULL);
@@ -214,10 +241,21 @@ static void _start_job(long my_job_id)
_xmit(out_msg);
}
+static void _suspend_job(long my_job_id)
+{
+ time_t now = time(NULL);
+ char out_msg[128];
+
+ snprintf(out_msg, sizeof(out_msg),
+ "TS=%u AUTH=root DT=CMD=SUSPENDJOB ARG=%ld",
+ (uint32_t) now, my_job_id);
+ _xmit(out_msg);
+}
+
int main(int argc, char * argv[])
{
- if (argc != 5) {
- printf("Usage: %s, control_addr job_id1 job_id2 sched_port\n",
+ if (argc != 6) {
+ printf("Usage: %s, control_addr job_id1 job_id2 sched_port is_bluegene\n",
argv[0]);
exit(1);
}
@@ -226,12 +264,19 @@ int main(int argc, char * argv[])
job_id1 = atoi(argv[2]);
job_id2 = atoi(argv[3]);
sched_port = atoi(argv[4]);
- printf("control_addr=%s job_id=%ld,%ld sched_port=%d\n",
- control_addr, job_id1, job_id2, sched_port);
+ is_bluegene = atoi(argv[5]);
+ printf("control_addr=%s job_id=%ld,%ld sched_port=%d is_bluegene=%d\n",
+ control_addr, job_id1, job_id2, sched_port, is_bluegene);
_get_jobs();
_get_nodes();
+ _modify_job(job_id1);
+ _get_jobs();
_start_job(job_id1);
+ if (!is_bluegene) {
+ _suspend_job(job_id1);
+ _resume_job(job_id1);
+ }
_cancel_job(job_id2);
sleep(5);
_get_jobs();