Problem moving database after fresh install

We’re doing a fresh install of the OS and cryosparc (v4.4.1+240110, standalone) on a system. The install went smoothly, and afterwards both “cryosparc install test” and web gui invocation of the scheduling confirmation job pass. However problems arise after restoring a database backup saved from an earlier system which A) had a master / server configuration (both on the same host), and B) had cryosparc installed in a different directory. Specifically, after doing the restore and restarting cryosparc, “cryosparc install test” from the CLI still passes, but attempting to run the scheduling confirmation job from the web gui (either existing or freshly created) now hangs at running the job. Looking at the metadata, it’s easy to see why:

“queue_message”: “[Errno 2] No such file or directory: ‘/home/cryosparc_user/V4.X/cryosparc_worker/bin/cryosparcw’”, “queued_at”: “2024-03-25T01:58:58.596Z”,

Somehow after restoring the database, it’s seeing the worker bin path which was associated with the previous system, not the current one. Is there some easy way to reset the bin path for all the projects / workspaces / jobs in the database?

FWIW, I’ve discovered a kludge which does work but is too hideous to use long term:
I can make a symbolic link for the new cryosparc_worker/bin path to the old one listed above.

Thanks in advance!
-jh-

Hi @jmh ,
The easiest way to correct this issue would be to reconnect your workers, using cryosparcw connect with the --update flag. See this link for more details.

1 Like

Thanks! Actually, I’d already looked at that documentation and tried a couple variants of cryosparcw connect --update with no luck. This time I found a combination that did work… specify master and worker with FQDNs (not localhost) and sshstr.