The database of jobs in the $(SPOOL) directory by default. SPOOL will also have intermediate files that the jobs need to run stored in various subdirectories.
Also, jobs also often need to copy input files from user's HOME directories.
So if you copy over all the user home directories and also SPOOL and it's subdirectories, you should be able to then start the new schedd and have it read in the job database and pick up where you left off.
You can run this command
condor_config_val SPOOL JOB_QUEUE_LOG
To locate the spool directory and the database of jobs. If there is no config entry for the job database, it will be $(SPOOL)/job_queue.log
Note that HTCondor must not be running while you copy the SPOOL directory and the job database.
From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of David Cohen <cdavid@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, December 16, 2021 6:25 AM
To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>
Subject: [HTCondor-users] Moving jobs from one schedd to another
Is there a way to move jobs from one schedd to another?
We've got new hardware to replace one of the aging schedd.
Waiting for all the jobs to naturally terminate will put the server out of order for days. While the installation itself should take half an hour at most.