Make server bulk RPC timeouts shorter than the client timeouts, so we don't
have cascading failures. Server bulk timeout is 1/4 of the client timeout. Also fix /proc variables to be int, as that is what the functions expect. b=1845
Showing
- lustre/include/linux/obd_support.h 3 additions, 3 deletionslustre/include/linux/obd_support.h
- lustre/mds/handler.c 1 addition, 1 deletionlustre/mds/handler.c
- lustre/obdclass/class_obd.c 3 additions, 4 deletionslustre/obdclass/class_obd.c
- lustre/ost/ost_handler.c 12 additions, 9 deletionslustre/ost/ost_handler.c
- lustre/ptlbd/rpc.c 1 addition, 1 deletionlustre/ptlbd/rpc.c
Please register or sign in to comment