I originally asked this question on SO, but i realized Serverfault was probably the correct place to ask.
I have an Artifactory Pro server running on an EC2 instance. Up yesterday morning, it was running version 7.41.13 and worked just fine.
It has a bunch of repositories that are hosted in the local Derby database with the blobs stored in S3.
Yesterday around 8:20 UTC, it crashed and rebooted. After the reboot, something was wrong. I can still see all packages under the "Artifactory -> Packages" in the UI, but there are no repositories under "Artifactory -> Artifacts", and I can't download the packages that I find in packages.
I restored some of the earlier hourly backups, and noticed that right as they boot, everything is fine, but after about 30-60 seconds, the Artifactory service restarts, and now the repos are gone.
Looking in the logs, I can see that the server starts up running v. 7.41.13, then restarts the service as version 7.49.3 and runs a lot of DB migration steps. After this, it says that there were no remote or local repos to map, corresponding to the missing repos in the UI.
I've looked through the documentation, logs, and config files, and I can't find anything about automatic upgrades so I'm very confused.
The server has been running fine for about 6 months prior to this, and nothing has changed in its configuration between now and yester day (AFAICT).
Has anyone seen this behaviour before, and do you have any suggestions as to how to fix the issue? Even stopping the upgrade and reboot would allow my users to extricate their packages long enough for me to do a full reinstall, but I can't see how to that either.