Skip to content

Plex storage migration approaches

We've started to run into I/O contention on the big, busy storage volumes, which use CephFS to make files available to multiple pods (for example, to Plex for storage, but to FileBrowser for browsing config).

In the next maintenance window, we'll be migrating users' Plex libraries (some of which are >50GB!) to Ceph "block" storage, which bypasses the I/O contention, at the cost of being unable to "view" your Plex metadata / DB / logs via FileBrowser (we'll come up with an alternate solution as required).

This migration has been observed to take a long time, especially for larger libraries (roughly 2 min / GB of Library), so we're trying to "stagger" it as best as possible, by shifting "earlybird" users onto the alpha channel, so that their migration can happen in a time of their choosing, rather than during maintenance.

If you've got a big(ish) Plex library, or you care quite a lot about when it goes down, create an #elf-help ticket to arrange a time (before the next maintenance window) to have your Plex instance migrated.

Auto-delete trash in Plex now disabled

Thanks to a suggestion from @pomnz, we're now explicitly setting Plex's setting not to delete trash automatically after a scan. (This can cause confusion and re-downloads as deleted files disappear, and then Radarr/Sonarr attempt to re-download the content!)

Today's scoreboard

Metric Numberz Delta
Total subscribers 354 +29
๐Ÿ‘ฝ Zurg mounts: 131 +2
๐Ÿ’พ ElfStorage in TBs 65 -
Tenant pods 4368 +271
๐Ÿฆธ Elf-vengers 4 -
๐Ÿง‘โ€๐ŸŽ“ Trainees 2 -
Bugz squished - -
New toyz 2 -

Summary

Thanks for geeking out with us, and please share these posts with related geeks!