Hi, Jaime:
Neither DAEMON_SHUTDOWN nor DAEMON_SHUTDOWN_FAST is defined:
node029:~# condor_config_val -master -verbose
DAEMON_SHUTDOWN_FAST
Not defined: DAEMON_SHUTDOWN_FAST
node029:~# condor_config_val -master -verbose DAEMON_SHUTDOWN
Not defined: DAEMON_SHUTDOWN
node029:~# condor_config_val -master -config
Configuration source:
/etc/condor/condor_config
Local configuration sources:
/etc/condor/config.d/security.config
/etc/condor/config.d/wn-slots.config
/etc/condor/config.d/wn-wn.config
/etc/condor/condor_config.local
Also checked all files under /etc/condor and nothing related
with those 2 parameters.
Cheers,Gang
On Aug 25, 2015, at 11:01 AM, Gang Qin <
Gang.Qin@xxxxxxxxxxxxx> wrote:
Dear condor Experts:
From time to time the condor service on one
machine turns off automatically, and in the
MasterLog I can see:
08/25/15 15:39:54 The DaemonShutdownFast _expression_
"1000000" evaluated to TRUE: starting fast shutdown
08/25/15 15:39:54 Got SIGQUIT. Performing fast
shutdown.
08/25/15 15:39:54 Sent SIGQUIT to STARTD (pid 3333)
08/25/15 15:40:00 AllReaper unexpectedly called on
pid 3333, status 0.
08/25/15 15:40:00 The STARTD (pid 3333) exited with
status 0
08/25/15 15:40:00 All daemons are gone. Exiting.
08/25/15 15:40:00 **** condor_master (condor_MASTER)
pid 3306 EXITING WITH STATUS 99
However, in the configuration files I didn't see the
setting of DAEMON_SHUTDOWN or DAEMON_SHUTDOWN_FAST.
node029:~# condor_config_val -dump | grep SHUTDOWN
EVENTD_SHUTDOWN_CLEANUP_INTERVAL = 3600
EVENTD_SHUTDOWN_CONSTRAINT =
EVENTD_SHUTDOWN_SLOW_START_INTERVAL = 0
EVENTD_SHUTDOWN_TIME =
EVENTD_SIMULATE_SHUTDOWNS =
NEGOTIATOR_TRIM_SHUTDOWN_THRESHOLD = 0
SHUTDOWN_FAST_TIMEOUT = 300
SHUTDOWN_GRACEFUL_TIMEOUT =
STARTD_FACTORY_SCRIPT_SHUTDOWN_PARTITION =
STARTD_NOCLAIM_SHUTDOWN = 0
Any idea what lead to this fast shutdown ?
Try running ‘condor_config_val -master
-verbose DAEMON_SHUTDOWN_FAST’. That will query the master
daemon for the value of that parameter and where the value
was set, rather than reading the config files directly.
Your daemons may be using a different configuration file
than the command line tools.
You can also run ‘condor_config_val -master
-config’ to see which configuration files the master read.
Thanks and regards,
Jaime Frey
UW-Madison HTCondor Project