I have an UnRAID Docker instance of Plex, with /transcode, /data, and /config on a cache ssd. Every 5 seconds, my Plex server executes the following with a corresponding 60-80kb read and write to my disk array (not the Cache SSD where /data, /transcode, and /config reside):
July 31, 2022 09:59:18.216 [0x14f2bbbe0b38] Debug — Request: [xxx.x.x.x:xxxxx (Loopback)] GET /identity (9 live) Signed-in
July 31, 2022 09:59:18.216 [0x14f2bec2ab38] Debug — Completed: [xxx.x.x.x:xxxxx] 200 GET /identity (9 live) 0ms 398 bytes (pipelined: 1)
I’m not really worried about it keeping the drive spun up (it’s hitting my parity drive and Data 1 drive which would be spun-up anyway, Data 2 and Data 3 spin down fine), but the sound of the repeated reads/writes every 5 seconds is becoming annoying. The 60-80kb being read and written at every 5 second cycle correlates directly with the loopback behavior shown in the Plex Console. Stopping the Plex Docker instance is the only way to stop the behavior.
Thoughts on how to remedy this? And why it’s reading/writing from/to the array and not the cache SSD?
Regarding the first issue, it is likely related to the health check being performed by Docker. This thread has some information (there are likely others, but this is the first one I found):
Regarding the second issue, it is the result of Plex Media Server sending SSDP discovery requests on the local network. You have a device (or devices) on your network which is (are) responding, but when queried further replies with invalid information. Plex provides no means by which you can fully disable its usage of SSDP discovery as it is integral to several features. The error messages themselves can be safely ignored, however the errors will always be logged.
You can use iptables or UFW to block outgoing messages to UDP port 1900 (or responses from the specific device) if it causes issues, but be aware it will break certain things (like network tuner automatic discovery).
I wonder, though, why am I seeing (and hearing) the reads/writes on the disk array when I have Plex Appdata set to go to cache and the /Appdata and /System shares set to Use Cache = Only?
Figured out why it was writing to the array instead of to the cache… since I’d installed my cache after I set up my Docker containers, docker.img hadn’t moved to the cache. I had to completely disable Docker (not just the containers), run Mover, then re-enable Docker. Now at least I don’t hear the spinning discs reading/writing every 5 seconds (even if it’s probably harder on the SSD in terms of wear-and-tear than it is on the spinning discs).
