Latest Blog

Move one role at a time after proving the shared-host bottleneck; extra hosts should create isolation, not just more network hops.

Keep enough Plex history to recover before a bad update or silent corruption, and verify multiple restore points instead of hoarding untested copies.

Move the full Plex state tree, preserve media paths and ownership, then validate architecture-specific analysis and acceleration features separately.

Do not edit the first Preferences.xml you find; prove the running Plex instance reads the mounted configuration path you are changing.

A useful Plex health check proves the paths required for playback and state writes, then reports dependency failures separately from process liveness.

Find the fast-growing log first, fix its root cause, then set retention so diagnostics remain useful without consuming the system volume.

Create separate datasets when app data, backups, and downloads need different snapshots, cleanup policies, permissions, and performance settings.

Use dataset-level recordsize for new writes, keep mixed shares conservative, and split media-heavy or small-file workloads when behavior differs.