Skip to content
Snippets Groups Projects
  • James Simmons's avatar
    6189ae07
    LU-10752 build: fix rpm packaging issues for gss · 6189ae07
    James Simmons authored
    
    Lustre can create rpms in two ways. One is with make rpm and the
    other is using the actual source rpm that is provided. Their are
    several issues with how GSS is handled with rpm packaging.
    
    First problem is that you can ./configure --disable-gss which has
    never been handled. Secondly if you do configure with disable-gss
    it is still possible to have the option enable-gss-keyring set to
    yes. The reason it was never seen before is due to everything
    being treated with the keyring option. Now if the user sets
    enable-gss to no then enable-gss-keyring will also be set to no
    even if the user tries to set it to yes. This was done by properly
    setting $enable_gss and $enable_gss_keyring in lustre-core.m4.
    In the spec file create the bcond gss to handle the gss only case
    and we turn on gss if gss_keyring is true. Move lgssc.conf under
    the with_gss_keyring bcond which is only needed for server builds
    along side lsvcgss.
    
    It is impossible to know if it can build due to the spec file not
    properly handling build dependencies for GSS and not knowing if
    the kernel is too new for GSS. So the user has to provide the
    options --with gss and / or --with gss-keyring to rpmbuild. If
    the user only provides gss-keyring option to rpmbuild make sure
    it enables gss as well. That is handled in the spec file.
    
    For the case of make rpms fix it up so if gss-keyring is enabled
    then by default the core gss handling is enabled. Also handle the
    long ignored enable-gss case.
    
    Test-Parameters: trivial
    
    Change-Id: Ieed9df98a27bd6e77504486762d6e60ddca5a916
    Signed-off-by: default avatarJames Simmons <uja.ornl@yahoo.com>
    Reviewed-on: https://review.whamcloud.com/31757
    
    
    Tested-by: Jenkins
    Tested-by: default avatarMaloo <hpdd-maloo@intel.com>
    Reviewed-by: default avatarSebastien Buisson <sbuisson@ddn.com>
    Reviewed-by: default avatarElena Gryaznova <c17455@cray.com>
    Reviewed-by: default avatarOleg Drokin <oleg.drokin@intel.com>
    6189ae07
    History
    LU-10752 build: fix rpm packaging issues for gss
    James Simmons authored
    
    Lustre can create rpms in two ways. One is with make rpm and the
    other is using the actual source rpm that is provided. Their are
    several issues with how GSS is handled with rpm packaging.
    
    First problem is that you can ./configure --disable-gss which has
    never been handled. Secondly if you do configure with disable-gss
    it is still possible to have the option enable-gss-keyring set to
    yes. The reason it was never seen before is due to everything
    being treated with the keyring option. Now if the user sets
    enable-gss to no then enable-gss-keyring will also be set to no
    even if the user tries to set it to yes. This was done by properly
    setting $enable_gss and $enable_gss_keyring in lustre-core.m4.
    In the spec file create the bcond gss to handle the gss only case
    and we turn on gss if gss_keyring is true. Move lgssc.conf under
    the with_gss_keyring bcond which is only needed for server builds
    along side lsvcgss.
    
    It is impossible to know if it can build due to the spec file not
    properly handling build dependencies for GSS and not knowing if
    the kernel is too new for GSS. So the user has to provide the
    options --with gss and / or --with gss-keyring to rpmbuild. If
    the user only provides gss-keyring option to rpmbuild make sure
    it enables gss as well. That is handled in the spec file.
    
    For the case of make rpms fix it up so if gss-keyring is enabled
    then by default the core gss handling is enabled. Also handle the
    long ignored enable-gss case.
    
    Test-Parameters: trivial
    
    Change-Id: Ieed9df98a27bd6e77504486762d6e60ddca5a916
    Signed-off-by: default avatarJames Simmons <uja.ornl@yahoo.com>
    Reviewed-on: https://review.whamcloud.com/31757
    
    
    Tested-by: Jenkins
    Tested-by: default avatarMaloo <hpdd-maloo@intel.com>
    Reviewed-by: default avatarSebastien Buisson <sbuisson@ddn.com>
    Reviewed-by: default avatarElena Gryaznova <c17455@cray.com>
    Reviewed-by: default avatarOleg Drokin <oleg.drokin@intel.com>