The time of the black screenshot was: August 8, 7:32PM.
I am sorry for the long time to respond, I was very busy this week. I just manually upgraded to 5291 after producing the requested logs on 5271, and for some reason now 5291 is working just fine. I can access 127.0.0.1:32400/web, I can access plex TV, and remote streaming all work just fine now for some unknown reason.
I will attach the logs for 5271. If you want those same logs for the now working 5291, let me know and I’ll grab those too. I will report back if it quits working again or becomes unstable.
I had the same. I left it once for like 30 mins and then realized it was working (it seems to do a full library scan after the upgrade and during this scan the server is appears offline in the client). However after about a couple of hours the web interface was unavailable again and this kind of continued thereafter. Sometimes it works for a couple of hours - day, than it doesn’t. Rebooting the server did not fix this.
I would be curious to know if it stays stable after you upgraded…
very useful to know. I think there is some lockout issue whilst requests to pubsub servers are being made. The evidence suggests there may have been a lockout until 7:35pm - by then the connection to the pubsub server was made
With this information i am able now to pass the original logs from the 8th August to the development team
There may be lockouts during initial connectivity stage
The netstat are all ok - showing listening on all ip addresses for 32400 and on localhost only for 32401 (the plex relay route)
I am not aware of any changes between 1.13.4.5271 and 1.13.5.5291 that would give rise to the timeouts/lockouts seen for the 8th August. The issue that I am aware of and is with the development team came in with 1.13.3
I use Ubuntu 14.04.5 and have experienced quite a few issues with 5291 where the server just becomes ‘unavailable’. Stopping the services results in many threads still running. They have to be manually killed before starting up the service again. This behavior is not necessarily new. It seems to pop up every now and then on different versions. There is always one thread that once is killed will result in all the other threads shutting down. It happens more often when doing library updates and as such I have disabled scheduled library scans. In any case, I have also dropped back to 5271 to see if this resolves the issue. Will post results one they come in. Usually takes about 3 days to fail.
further to my earlier response, if you do notice a lockout please could you try to get the output from this request saved into a text file
http://127.0.0.1:32400/connections?X-Plex-Token=xxxxxxxxxxxxxx
and send to me by private message - and of course collect the logs as well with debug logging enabled beforehand
From what i have seen the lockouts are transient and not permanent - of course the /connections request needs to be done when http://127.0.0.1:32400/web hangs and you know the server is actually running
So far so good. No additional lockouts, or server unavailability after about 24 hours. It has been stable since the 5291 update took. Strangely, when I updated outside of the built-in update button (which often showed as greyed out on 5271), and upgraded via downloading the latest PMS from the website directly, there were NO lockouts or server downtime. This is the only difference between my upgrade path initially. However, this doesn’t mean that the built in updater isn’t working correctly. This is likely a coincidence. It’s just strange that the issues I had when I used the built in updater did not manifest what-so-ever on the manual upgrade. I tried the built in updater multiple times when I first discovered this issue.
None-the-less, I will continue to follow this thread and report any instability for me or my users on 5291. If the issue arrises again, I will collect the new logs as instructed above.
I did the same thing last night and had the same results. However, I restarted my computer this morning and when the server started back up I was back to the same problems with it not working. I have again gone back to the earlier version and will stay there until whatever the issue is with this update is corrected.
I uploaded logs with the latest update in the other thread last night and this morning. Thank you for looking into the issue. I have been following the threads on this issue and have tried fixing on my own with no luck. Very frustrating.
