- 01 Aug, 2012 18 commits
-
-
Danny Auble authored
-
Danny Auble authored
-
Danny Auble authored
-
Danny Auble authored
-
Morris Jette authored
Change node_req field in struct job_resources from 8 to 32 bits so we can run more than 256 jobs per node.
-
Morris Jette authored
-
Danny Auble authored
-
Danny Auble authored
correctly as of IBM driver V1R1M1 efix 008.
-
Morris Jette authored
-
Danny Auble authored
-
Danny Auble authored
Conflicts: src/common/read_config.c
-
Danny Auble authored
configured correctly.
-
Morris Jette authored
switch/nrt: un/pack job request for PMD use switch/nrt: pass launch_state info to PMD
-
Morris Jette authored
-
Morris Jette authored
Specifically if memory allocations are not tracked, this code could prevent jobs from running due to lack of memory.
-
Danny Auble authored
-
Danny Auble authored
Move code to handle multiple clusters. Previous code could have issues where job_db_inx space could overlap.
-
Morris Jette authored
-
- 31 Jul, 2012 18 commits
-
-
Danny Auble authored
current or in the past.
-
Morris Jette authored
-
Danny Auble authored
-
Mark Nelson authored
from Mark Nelson
-
Janne Blomqvist authored
Using the syscalls directly rather than calling bin/(u)mount via system() avoids a few fork + exec calls, and provides better error handling if something goes wrong. Users of this functionality are also updated to use slurm_strerror in order to provide a more informative error message. The mount and umount syscalls are Linux-specific, but so are cgroups so no portability is lost.
-
Danny Auble authored
-
Danny Auble authored
Using the syscalls directly rather than calling bin/(u)mount via system() avoids a few fork + exec calls, and provides better error handling if something goes wrong. Users of this functionality are also updated to use slurm_strerror in order to provide a more informative error message. The mount and umount syscalls are Linux-specific, but so are cgroups so no portability is lost.
-
Danny Auble authored
Using the syscalls directly rather than calling bin/(u)mount via system() avoids a few fork + exec calls, and provides better error handling if something goes wrong. Users of this functionality are also updated to use slurm_strerror in order to provide a more informative error message. The mount and umount syscalls are Linux-specific, but so are cgroups so no portability is lost.
-
Danny Auble authored
the current plugin has been loaded when using runjob_mux_refresh_config
-
Danny Auble authored
DBD rpc version to SLURM protocol version.
-
Rod Schultz authored
could be called from a dbd pack/unpack as well as from regular rpc's. This attempts to fix the issue buy passing a variable to the function to adjust from the dbd rpc version to the slurm rpc version.
-
Don Lipari authored
These comments were orphaned with this commit: 874f797f Move the start time calculation of pending jobs into a separate pthread on Nov 3, 2010.
-
Don Lipari authored
These comments were orphaned with this commit: 874f797f Move the start time calculation of pending jobs into a separate pthread on Nov 3, 2010.
-
Stephen Trofinoff authored
Code improvements for recording GRES use: The xmalloc() function initialized everything to zero. There is no need for a call to memset() afterwards. The xfree() function tests for a NULl arguments and does nothing in that case. There is no need for an "if" test before calling xfree(). There is an xrealloc() function available that eliminates the need for saving and restoring values. I made some formatting changes to comply with the Linux kernel coding standard, mostly related to lines over 80 characters long. JobID and some other fields should be formatted using "%u" rather than "%d" to avoid negative job ID's being logged. I added tests of buffer sizes in a couple of places to avoid possibly going off the end of a buffer. Code for the function gres_get_value_by_type() was missing from the patches. Gres accounting modifications Fix memory leak, avoid xstrdup(). Bump the rpc_version number in for the slurmdbd. Break up some long lines. gres accounting: save/restore job gres requested/allocated info, get job counts properly gres accounting: updates based upon email from Steve Trofinoff, 31 July` gres accounting: remove job gres_res and just replace gres
-
Danny Auble authored
-
Danny Auble authored
-
Danny Auble authored
-
Danny Auble authored
already open File * to write logs to. This is helpful when something sets up logging outside of SLURM but wants SLURM debug written to the same place as the rest of there logs.
-
- 30 Jul, 2012 4 commits
-
-
Morris Jette authored
-
Morris Jette authored
This particularly effects --multi-slurmd mode and the Port option
-
Morris Jette authored
-
Morris Jette authored
-