strigger - Used set, get or clear Slurm trigger information.
strigger --set [OPTIONS...]
strigger --get [OPTIONS...]
strigger --clear [OPTIONS...]
strigger is used to set, get or clear Slurm trigger
information. Triggers include events such as a node failing, a job reaching
its time limit or a job terminating. These events can cause actions such as
the execution of an arbitrary script. Typical uses include notifying system
administrators of node failures and gracefully terminating a job when it's
time limit is approaching. A hostlist expression for the nodelist or job ID
is passed as an argument to the program.
Trigger events are not processed instantly, but a check is
performed for trigger events on a periodic basis (currently every 15
seconds). Any trigger events which occur within that interval will be
compared against the trigger programs set at the end of the time interval.
The trigger program will be executed once for any event occurring in that
interval. The record of those events (e.g. nodes which went DOWN in the
previous 15 seconds) will then be cleared. The trigger program must set a
new trigger before the end of the next interval to ensure that no trigger
events are missed OR the trigger must be created with an argument of
"--flags=PERM". If desired, multiple trigger programs can be set
for the same event.
IMPORTANT NOTE: This command can only set triggers if run
by the user SlurmUser unless SlurmUser is configured as user
root. This is required for the slurmctld daemon to set the
appropriate user and group IDs for the executed program. Also note that the
trigger program is executed on the same node that the slurmctld
daemon uses rather than some allocated compute node. To check the value of
SlurmUser, run the command:
scontrol show config | grep SlurmUser
- -a,
--primary_slurmctld_failure
- Trigger an event when the primary slurmctld fails.
- -A,
--primary_slurmctld_resumed_operation
- Trigger an event when the primary slurmctld resuming operation after
failure.
- -b,
--primary_slurmctld_resumed_control
- Trigger an event when primary slurmctld resumes control.
- -B,
--backup_slurmctld_failure
- Trigger an event when the backup slurmctld fails.
- -c,
--backup_slurmctld_resumed_operation
- Trigger an event when the backup slurmctld resumes operation after
failure.
- -C,
--backup_slurmctld_assumed_control
- Trigger event when backup slurmctld assumes control.
- --burst_buffer
- Trigger event when burst buffer error occurs.
- --clear
- Clear or delete a previously defined event trigger. The --id,
--jobid or --user option must be specified to identify the
trigger(s) to be cleared. Only user root or the trigger's creator can
delete a trigger.
- -d, --down
- Trigger an event if the specified node goes into a DOWN state.
- -D, --drained
- Trigger an event if the specified node goes into a DRAINED state.
- -e,
--primary_slurmctld_acct_buffer_full
- Trigger an event when primary slurmctld accounting buffer is full.
- -F, --fail
- Trigger an event if the specified node goes into a FAILING state.
- -f, --fini
- Trigger an event when the specified job completes execution.
- --flags=type
- Associate flags with the reservation. Multiple flags should be comma
separated. Valid flags include:
- PERM
- Make the trigger permanent. Do not purge it after the event occurs.
- --front_end
- Trigger events based upon changes in state of front end nodes rather than
compute nodes. Applies to Cray ALPS architectures only, where the slurmd
daemon executes on front end nodes rather than the compute nodes. Use this
option with either the --up or --down option.
- -g,
--primary_slurmdbd_failure
- Trigger an event when the primary slurmdbd fails.
- -G,
--primary_slurmdbd_resumed_operation
- Trigger an event when the primary slurmdbd resumes operation after
failure.
- --get
- Show registered event triggers. Options can be used for filtering
purposes.
- -h,
--primary_database_failure
- Trigger an event when the primary database fails.
- -H,
--primary_database_resumed_operation
- Trigger an event when the primary database resumes operation after
failure.
- -i,
--id=id
- Trigger ID number.
- -I, --idle
- Trigger an event if the specified node remains in an IDLE state for at
least the time period specified by the --offset option. This can be
useful to hibernate a node that remains idle, thus reducing power
consumption.
- -j,
--jobid=id
- Job ID of interest. NOTE: The --jobid option can not be used
in conjunction with the --node option. When the --jobid
option is used in conjunction with the --up or --down
option, all nodes allocated to that job will considered the nodes used as
a trigger event.
- -M,
--clusters=<string>
- Clusters to issue commands to. Note that the SlurmDBD must be up for this
option to work properly.
- -n,
--node[=host]
- Host name(s) of interest. By default, all nodes associated with the job
(if --jobid is specified) or on the system are considered for event
triggers. NOTE: The --node option can not be used in
conjunction with the --jobid option. When the --jobid option
is used in conjunction with the --up, --down or
--drained option, all nodes allocated to that job will considered
the nodes used as a trigger event. Since this option's argument is
optional, for proper parsing the single letter option must be followed
immediately with the value and not include a space between them. For
example "-ntux" and not "-n tux".
- -N,
--noheader
- Do not print the header when displaying a list of triggers.
- -o,
--offset=seconds
- The specified action should follow the event by this time interval.
Specify a negative value if action should preceded the event. The default
value is zero if no --offset option is specified. The resolution of
this time is about 20 seconds, so to execute a script not less than five
minutes prior to a job reaching its time limit, specify
--offset=320 (5 minutes plus 20 seconds).
- -p,
--program=path
- Execute the program at the specified fully qualified pathname when the
event occurs. You may quote the path and include extra program arguments
if desired. The program will be executed as the user who sets the trigger.
If the program fails to terminate within 5 minutes, it will be killed
along with any spawned processes.
- -Q, --quiet
- Do not report non-fatal errors. This can be useful to clear triggers which
may have already been purged.
- -r,
--reconfig
- Trigger an event when the system configuration changes. This is triggered
when the slurmctld daemon reads its configuration file or when a node
state changes.
- --set
- Register an event trigger based upon the supplied options. NOTE: An event
is only triggered once. A new event trigger must be set established for
future events of the same type to be processed. Triggers can only be set
if the command is run by the user SlurmUser unless SlurmUser
is configured as user root.
- -t, --time
- Trigger an event when the specified job's time limit is reached. This must
be used in conjunction with the --jobid option.
- -u, --up
- Trigger an event if the specified node is returned to service from a DOWN
state.
- --user=user_name_or_id
- Clear or get triggers created by the specified user. For example, a
trigger created by user root for a job created by user adam
could be cleared with an option --user=root. Specify either a user
name or user ID.
- -v, --verbose
- Print detailed event logging. This includes time-stamps on data
structures, record counts, etc.
- -V ,
--version
- Print version information and exit.
- TRIG_ID
- Trigger ID number.
- RES_TYPE
- Resource type: job or node
- RES_ID
- Resource ID: job ID or host names or "*" for any host
- TYPE
- Trigger type: time or fini (for jobs only), down or
up (for jobs or nodes), or drained, idle or
reconfig (for nodes only)
- OFFSET
- Time offset in seconds. Negative numbers indicated the action should occur
before the event (if possible)
- USER
- Name of the user requesting the action
- PROGRAM
- Pathname of the program to execute when the event occurs
Some strigger options may be set via environment variables.
These environment variables, along with their corresponding options, are
listed below. (Note: commandline options will always override these
settings)
- SLURM_CONF
- The location of the Slurm configuration file.
Execute the program
"/usr/sbin/primary_slurmctld_failure" whenever the primary
slurmctld fails.
> cat /usr/sbin/primary_slurmctld_failure
#!/bin/bash
# Submit trigger for next primary slurmctld failure event
strigger --set --primary_slurmctld_failure \
--program=/usr/sbin/primary_slurmctld_failure
# Notify the administrator of the failure using by e-mail
/usr/bin/mail slurm_admin@site.com -s Primary_SLURMCTLD_FAILURE
> strigger --set --primary_slurmctld_failure \
--program=/usr/sbin/primary_slurmctld_failure
Execute the program "/usr/sbin/slurm_admin_notify"
whenever any node in the cluster goes down. The subject line will include
the node names which have entered the down state (passed as an argument to
the script by Slurm).
> cat /usr/sbin/slurm_admin_notify
#!/bin/bash
# Submit trigger for next event
strigger --set --node --down \
--program=/usr/sbin/slurm_admin_notify
# Notify administrator using by e-mail
/usr/bin/mail slurm_admin@site.com -s NodesDown:$*
> strigger --set --node --down \
--program=/usr/sbin/slurm_admin_notify
Execute the program "/usr/sbin/slurm_suspend_node"
whenever any node in the cluster remains in the idle state for at least 600
seconds.
> strigger --set --node --idle --offset=600 \
--program=/usr/sbin/slurm_suspend_node
Execute the program "/home/joe/clean_up" when job 1234
is within 10 minutes of reaching its time limit.
> strigger --set --jobid=1234 --time --offset=-600 \
--program=/home/joe/clean_up
Execute the program "/home/joe/node_died" when any node
allocated to job 1234 enters the DOWN state.
> strigger --set --jobid=1234 --down \
--program=/home/joe/node_died
Show all triggers associated with job 1235.
> strigger --get --jobid=1235
TRIG_ID RES_TYPE RES_ID TYPE OFFSET USER PROGRAM
123 job 1235 time -600 joe /home/bob/clean_up
125 job 1235 down 0 joe /home/bob/node_died
Delete event trigger 125.
> strigger --clear --id=125
Execute /home/joe/job_fini upon completion of job 1237.
> strigger --set --jobid=1237 --fini
--program=/home/joe/job_fini
Copyright (C) 2007 The Regents of the University of California.
Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
Copyright (C) 2008-2010 Lawrence Livermore National Security.
Copyright (C) 2010-2013 SchedMD LLC.
This file is part of Slurm, a resource management program. For
details, see <https://slurm.schedmd.com/>.
Slurm is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2 of the License, or (at your option)
any later version.
Slurm is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
more details.