- Apr 04, 2003
-
-
Moe Jette authored
-
- Apr 01, 2003
-
-
Mark Grondona authored
but semaphore doesn't. o shm.c : added new _shm_sane() call that attempts to ensure shared memory is not improperly locked on attach. o slurmd.c : move more initialization code into slurmd_init() o slurmd.c : _reconfigure() now calls _read_config()
-
- Mar 28, 2003
-
-
Mark Grondona authored
o chdir to /tmp to avoid filling spooldir with corefiles (just in case)
-
- Mar 27, 2003
-
-
Mark Grondona authored
to be sure" o src/slurmd/slurmd.c (_reconfigure) : reinitialize job credential key
-
- Mar 25, 2003
-
-
Moe Jette authored
-
- Mar 23, 2003
-
-
Mark Grondona authored
does not exist. o exit early from slurmd if slurmd spooldir cannot be created so the bug above would've never been tripped.
-
- Mar 14, 2003
-
-
Mark Grondona authored
slurm_cred_t job credential.
-
- Mar 13, 2003
-
-
Mark Grondona authored
o src/common/slurm_cred.[ch] : add function slurm_cred_insert_jobid() for updating the internal job state on the verifier side without verifying a credential. (Used when running batch jobs)
-
- Mar 03, 2003
-
-
jwindley authored
-
- Feb 26, 2003
-
-
Mark Grondona authored
that job/job step session leader runs as owner of job and not root. Split "session manger" code from mgr.c (which runs as root) into smgr.c (most session manager code runs as user) o reorganize interconnect functions (see interconnect.h) to fit new dual-process model for slurmd. o add pipe to job structure for communication between slurmd job manager and session manager. o add "mpid" field to shared memory for job steps and shm_update_step_mpid() to update this information. o add ulimits.[ch] to support setting user limits for jobs based off SLURM_RLIMIT* env vars.
-
- Feb 20, 2003
-
-
Moe Jette authored
-
- Feb 19, 2003
-
-
Moe Jette authored
Add error test on close of pid file in slurmd.
-
- Feb 14, 2003
-
-
Mark Grondona authored
o src/srun/ : srun code cleanup. Do not set stdin nonblocking as this has a side-effect of setting stdout and stderr nonblocking too. o src/slurmd : code cleanup. Terminate if read of public key fails. o src/common/credential/signature_utils* : code cleanup
-
- Feb 12, 2003
-
-
Mark Grondona authored
revoked after second use. o common/daemonize.c : do not report inability to open pidfile as an error (since it invariably is not) o slurmd/ : more initialization cleanup.
-
- Feb 11, 2003
-
-
Mark Grondona authored
-
Mark Grondona authored
circumstances. o README : add information on how to generate slurm keys
-
- Jan 31, 2003
-
-
Moe Jette authored
scontrol reports slurmd for "show daemons" only if there is some ControlMachine configured for the slurmd to talk with. Slurmd and Slurmctld report errors and exit gracefully if no ControlMachine.
-
Mark Grondona authored
o src/srun : various signal handling fixes, retry task launch req on failure
-
- Jan 28, 2003
-
-
Mark Grondona authored
creation functions into allocate.[ch] o src/common/xsignal.[ch] : xsignal returns old signal handler o src/slurmd : remove list of running threads - use number of active threads instead. (terminating slurmd sleeps until all active threads have vanished)
-
Moe Jette authored
update in progress for reconfig).
-
- Jan 24, 2003
-
-
Mark Grondona authored
can be waited on in shutdown.
-
Moe Jette authored
Slurmctld treats prolog failure (reported as error code in node_registration RPC) as sign to set node DOWN. Slurmd reports prolog failure on srun step task launch back to slurmctld to set node down.
-
- Jan 23, 2003
-
-
Moe Jette authored
ALLOCATED to IDLE. Save creds to file on each cred revoke and on slurmd shutdown.
-
- Jan 22, 2003
-
-
Mark Grondona authored
o Try harder to clean up shm semaphore in slurmd o call daemon(1,0) instead of daemon(0,0) in slurmctld to preserve chdir()
-
- Jan 17, 2003
-
-
Moe Jette authored
-
- Jan 16, 2003
-
-
Mark Grondona authored
prolog is run as root and failure is reported as a non-zero return code from the launch tasks or launch batch job messages epilog is run as result of revoke credential message
-
Moe Jette authored
Modify slurmd to dump/restore credentials at termination and restart.
-
- Jan 14, 2003
-
-
Mark Grondona authored
o slurmd tries to kill any old slurmd's on startup o create_pidfile() locks pid file and keeps it open, so that we can test pidfile for lock to definitively know whether old slurmd is still running.
-
- Jan 13, 2003
-
-
Moe Jette authored
config file values.
-
- Dec 19, 2002
-
-
Moe Jette authored
-
- Dec 14, 2002
-
-
Mark Grondona authored
o slurmd: if there is room in cmdline, change argv to show what job a slurmd is managing.
-
- Dec 10, 2002
-
-
Moe Jette authored
-
- Dec 05, 2002
-
-
Mark Grondona authored
still remain. o cbuf.[ch], list.[ch] : Checked in new revisions from lsd-tools o srun/io.c,msg.c : added io_thr_create(), msg_thr_create() to simplify code in main(). o Other minor fixes/changes
-
- Nov 22, 2002
-
-
Moe Jette authored
Changed "#include <locals.h>" to "#include "locals.h"" in the other files.
-
- Nov 21, 2002
-
-
Mark Grondona authored
o slurmd shared memory fixes -- no longer detach/reattach around interconnect_init o put slurm.lock in spooldir by default o slurmd daemonize is now the default o log.c fixes -- use a shorter timestamp in logfiles
-
- Nov 15, 2002
-
-
Mark Grondona authored
along with other general bugfixes and cleanup. o cbuf.[ch] list.[ch] : begin using Chris' list and cbuf routines from the LSD-Tools project. o srun/srun.c : *Always* wait for I/O to complete. This still needs work. o srun/launch.c : code cleanup
-
- Nov 13, 2002
- Nov 08, 2002
-
-
Moe Jette authored
-
- Oct 31, 2002
-
-
Mark Grondona authored
o make dist now makes a proper distribution
-