The HTCondor central manager does not have a database, The job queue does not reside there, it resides with the condor_schedd process, which we call the Access Point. A HTCondor pool can have many Access Points and each will have their own job queue.
You can find the file that holds the job queue by running
condor_config_val SPOOL JOB_QUEUE_LOG
on each Access Point to find the location of the local SPOOL directory and
the explicit location of the transaction log for the job queue (JOB_QUEUE_LOG). If the configuration for JOB_QUEUE_LOG is not set, then the job queue log will be in the SPOOL directory. it is the file named
job_queue.log
You will want to back up the entire SPOOL directory to save the jobs and their associated temporary files. you should do this while the condor_schedd process is not running.
To save the jobs, but not the files associated, you can back up just the job_queue.log file.
-tj
From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of acraft@xxxxxxxxx <acraft@xxxxxxxxx>
Sent: Tuesday, November 16, 2021 1:17 PM To: htcondor-users@xxxxxxxxxxx <htcondor-users@xxxxxxxxxxx> Subject: [HTCondor-users] Can we backup Central Manager's database and states? Hello,
I'm planning to use Docker containers to host a Central Manager (CM) service and I have some questions:
1. Does CM keep a database to store it's job queue to make it persistent?
2. If yes, what folders and files should I save to backup CM's database and states?
Aaron
|