brequeue
Kills and re-queues a job.
Synopsis
brequeue [-a] [-d] [-e] [-H] [-p] [-r] [-J job_name | -J "job_name[index_list]"] [-u user_name | -u all] [job_ID | "job_ID[index_list]" ...]Description
You can use the brequeue command only on jobs that you own. root or the LSF cluster administrator or LSF group administrator can requeue jobs for any user.
Kills a running (RUN), user-suspended (USUSP), or system-suspended (SSUSP) job and returns it to the queue. A job that is killed and requeued retains its submit time but is dispatched according to its requeue time. When the job is requeued, it is assigned the PEND status or PSUSP if the -H option is used. after the job is dispatched, it starts over from the beginning. The requeued job keeps the same job ID.
When the JOB_INCLUDE_POSTPROC=Y parameter is set in the lsb.params file or in an application profile in the lsb.applications file, job requeue occurs only after post-execution processing, not when the job finishes. When used for host-based post-execution processing, configure a longer time period to allow the operation to run.
Use the brequeue command to requeue job arrays or job array elements.
By default, the brequeue command kills and requeues your most recently submitted job when no job ID is specified.
In the LSF multicluster capability lease model, you can use the brequeue command only on jobs in local queues. A job that is killed and requeued is assigned a new job ID on the execution cluster, but it retains the same job ID on the submission cluster. For example, a job that is submitted from cluster A that is killed and requeued and then runs on execution cluster B is assigned a new job ID on cluster B. However, when the bjobs command runs from submission cluster A, the job is displayed with the original job ID. When the bjobs command runs from execution cluster B, the job is displayed with the new job ID.
In the LSF multicluster capability job forwarding model, use the brequeue -p command to requeue specified remote pending jobs. Use the brequeue -a command to requeue all non-pending jobs (including running jobs, suspended jobs, jobs with EXIT or DONE status) in the local cluster. The brequeue -a command does not requeue pending jobs in the local cluster.
The only difference between the -p and -a options is the job dispatch order. Running the brequeue -p command on the submission cluster requeues a remote job to the top of the queue so that the requeued job is dispatched first no matter which position it is in the pending job list. The brequeue -a command sends the remote job to the end of the queue the same way as in the local cluster.
When absolute job priority scheduling (APS) is configured in the queue, specified requeued jobs are treated as newly submitted jobs for APS calculation. The job priority, system, and the ADMIN APS factors are reset on requeue.
When you use multi-phase rusage resource requirement strings, such as with the bsub -R command, the requeued job is treated as a new job and resources are reserved from the beginning of the first phase.
Options
- -a
- Requeues all local non-pending jobs, including running jobs, suspending jobs, jobs with EXIT or DONE status, and pending remote jobs. It does not requeue pending jobs in the local cluster.
- -d
- Requeues jobs that finished running with DONE job status.
- -e
- Requeues jobs that terminated abnormally with EXIT job status.
- -H
- Requeues jobs to PSUSP job status.
- -p
- In the LSF multicluster capability job forwarding model, requeues specified jobs that are pending in a remote cluster for LSF multicluster capability job forwarding modes.
- -r
- Requeues jobs that are running.
- -J job_name | -J "job_name[index_list]"
- Operates on the specified job.
The job name can be up to 4094 characters long. Job names are not unique.
The wildcard character (*) can be used anywhere within a job name, but it cannot appear within an array index. For example, the pattern job* returns jobA and jobarray[1]. The *AAA*[1] pattern returns the first element in job arrays with names that contain AAA. However, the pattern job1[*] does not return anything since the wildcard is within the array index.
- -u user_name | -u all
- Operates on the specified user’s jobs or all jobs. To specify a Windows user account, include
the domain name in uppercase letters and use a single backslash (DOMAIN_NAME∖user_name)
at a Windows command prompt. On a UNIX or Linux command line, use a double backslash
(DOMAIN_NAME∖∖user_name).
Only root and LSF administrators can requeue jobs that are submitted by other users.
- job_ID |"job_ID[index_list]"
- Operates on the specified job or job array elements.
The value of 0 for job_ID is ignored.
- -h
- Prints command usage to stderr and exits.
- -V
- Prints LSF release version to stderr and exits.
Limitations
The brequeue command cannot be used on interactive batch jobs. The brequeue command kills only interactive batch jobs, it does not restart them.