[09:14] Customer: Hi, I upgraded my self-hosted photo library last night, and now some albums show empty even though the photos still appear in the main timeline.

[09:15] Support: Hi! I can help with that. Which version did you upgrade from and to, and how is the application deployed?

[09:16] Customer: From 1.98.2 to 1.101.0. It’s running with Docker Compose on Ubuntu 22.04, with PostgreSQL and the photo storage mounted from my NAS.

[09:17] Support: Thanks. If the photos remain visible in the timeline, the files and primary asset records are likely intact. Empty albums after an upgrade can indicate a pending database migration or an outdated background worker. Are all containers running the same application version?

[09:18] Customer: I’m not sure. I pulled the new image and restarted everything.

[09:19] Support: Could you run `docker compose ps` and check the image tags for the server and worker containers? Please remove any hostnames, usernames, or other sensitive details before pasting the output.

[09:22] Customer: The server is on 1.101.0, but the worker says 1.98.2. I had pinned that image separately and forgot about it.

[09:23] Support: That would explain the inconsistent album indexing. Update the worker to 1.101.0, then run `docker compose pull` followed by `docker compose up -d`. Once it restarts, check the worker logs for migration or indexing errors.

[09:27] Customer: Done. Both containers now show 1.101.0. The logs say, “Album reconciliation job queued,” and then several jobs completed successfully.

[09:28] Support: Great. Please wait a minute, refresh the web interface, and open one of the affected albums.

[09:30] Customer: Most of the albums are back, but one shared album still looks empty.

[09:31] Support: Is that album owned by your account, or was it shared with you by another user?

[09:31] Customer: It belongs to my wife’s account and was shared with me.

[09:32] Support: Please have her sign in and check whether the album contains photos from her account. If it does, she can remove your access and share it with you again. The upgrade may have left the share membership stale even though the album itself is intact.

[09:35] Customer: She can see all 312 photos. She removed me and shared it again, and now they appear for me too.

[09:36] Support: Excellent. The issue was caused by the server and worker running different versions, plus one stale album-sharing record. Your original photo files were not affected.

[09:37] Customer: Is there anything else I should do to make sure the library is healthy?

[09:38] Support: I recommend checking that every application container uses the same version tag, confirming that no migration jobs are still pending, and running a database backup now that the upgrade is stable. It’s also worth verifying that your NAS backup includes the original photo directory rather than only thumbnails.

[09:40] Customer: Everything is green, there are no pending jobs, and the database backup completed. Thanks for the help.

[09:40] Support: You’re welcome! I’m glad the albums are restored. If you notice missing thumbnails, failed uploads, or additional sharing issues, send us the relevant container logs and we’ll take another look.