Backup.sh procedure is just able to run once

[christianbraun]

Hi everyone,

I'm working on a 4.4.1 GeoNode project using Docker. I'm about to set up the backup procedure, and I noticed something strange. If I run backup.sh just once, it finishes successfully. But if I run it again, it gets stuck trying to contact GeoServer and waits forever.

After restarting the Docker Compose, backup.sh works again, but it fails during a second call.

Any ideas on what could be causing this?

[caocuongngo]

I have the same issue! @christianbraun Did you resolve it?

[caocuongngo]

I found the issue: it's the GeoWebCache that creates a lock, thus backup can't acquire it --> resulting hang up. Override backup.py with BK_SKIP_GWC=true to solve the issue.

[kikislater]

@caocuongngo I think you can make a PR to solve the issue ?

[caocuongngo]

@kikislater I don't have write access.

[kikislater]

It is like every project: make a fork, then modify the code and you can purpose a PR!

[christianbraun]

@caocuongngo , thanks for investigating and finding a fix.
I guess I have to inject the variable into the _options list in line 247 of backup.py?

[caocuongngo]

@christianbraun Yes, I override create_geoserver_backup and create _options = [ "BK_CLEANUP_TEMP=true", "BK_SKIP_SETTINGS=false", "BK_SKIP_SECURITY=false", "BK_SKIP_GWC=true", f'BK_BEST_EFFORT={("true" if ignore_errors else "false")}', f"exclude.file.path={config.gs_exclude_file_path}", ]