Server Version#:1.19.5.3112-b23ab3896 and 1.20.1.3213-67a1137a9
OS: MacOS Catalina 10.15.6
I’ve been a plexpass holder for quite some time, and never had any issues that a reboot didnt solve. Now for some reason starting yesterday, my plex server wont stay running for more than 30 seconds or so, then it closes unexpectedly. I’ve tried the two versions available to me on the download screen and the same effect. I also tried disabling any plugins I’ve added, and that did not change the behavior.
I’m not even sure what else to do. I see there is a .dmp file in the crashreports folder, but do not know how to open/read it.
This could be a sign of a corrupted database (among other things). Your logs may show if it if that’s the case (make sure Debug is enabled and Verbose disabled under Settings -> General). You can search them for “corrupted” or “malformed” or upload them here for us to have a look.
This support article also describes how to check for and repair a corrupt database using sqlite3:
And this article describes how to restore a backed up database:
Thanks so much for your quick reply. I tried the database restore (it was only a couple days old) and that didn’t help.
I ran through the repair steps just copy/pasting commands into a terminal not really understanding what I was doing, and I think that did the trick! Its been up for a few minutes and hasn’t crashed yet!
Ok, thanks. I see that the log files have many files. Which one would be most helpful to put in pastebin for posting here? Also, is there any personal/private info in the logs that should be scrubbed first?
If you’re only going to upload one log, it should probably be Plex Media Server.log. It would probably be best to just upload the entire package from Settings -> Troubleshooting -> Download Log Files. There’s no need to use pastebin, you can just drag/drop it to the post window.
Plex does a pretty good job of scrubbing personal details from the logs, there’s nothing in there of which I’m aware which can be used to compromise your account. However, it can include public IP addresses and user names, so bear that in mind.
Well here is an update. If I’m on 1.19, it stays up after following the repair steps. I tried going to 1.20, and the same problem came back. I reverted back to 1.19, and it kept crashing, so I did the repair steps again, and its staying up. I’m happy staying on 1.19 for now but I assume this might be an issue again later.
Any recommended next steps?
Did you attempt the repair after upgrading to 1.20.x.x and experiencing the problem?
The next step is likely to attempt the upgrade again and, if it fails, upload the resulting logs. And perhaps try the database repair while on 1.20.x.x.
Ok, well I guess I spoke too soon. It kicked off a library scan, then crashed partway through. I enabled verbose logging, and let it crash again. Here is the log zip: Plex Media Server Logs_2020-08-20_10-47-15.zip (779.7 KB)
Ok, verbose disabled. I cleaned out the logs and let it crash. The local time was ~10:59:41 when it crashed. Here are the new log files:Plex Media Server Logs_2020-08-20_11-00-16.zip (37.4 KB)
Thanks. There’s no obvious database corruption, or at least none logged. There are some timeouts connecting to network resources:
Aug 20, 2020 10:59:33.018 [0x70000e950000] ERROR - Error issuing curl_easy_perform(handle): 28
Aug 20, 2020 10:59:33.018 [0x70000e9d3000] ERROR - Error issuing curl_easy_perform(handle): 28
Aug 20, 2020 10:59:38.018 [0x70000d143000] WARN - Connection Tester: Error testing media server HendersonPlex in 10005 ms.
Aug 20, 2020 10:59:39.196 [0x70000d143000] ERROR - Unknown metadata type:
Aug 20, 2020 10:59:39.768 [0x70000d2cc000] ERROR - Error issuing curl_easy_perform(handle): 28
Aug 20, 2020 10:59:40.394 [0x70000e84a000] ERROR - Error issuing curl_easy_perform(handle): 28
Aug 20, 2020 10:59:41.378 [0x70000d5de000] DEBUG - EventSource: Failure in IdleTimeout (0 - Undefined error: 0).
Aug 20, 2020 10:59:41.379 [0x70000d5de000] ERROR - EventSource: Retrying in 15 seconds.
These immediately precede the crash, so they could be relevant. However, I see (a fewer number of) similar errors on my functioning system, so take that with a grain of salt.
There’re also some indications in the logs that it’s initializing and/or communicating with plug-ins near when the crash occurs. If you use any third-party plug-ins, it may be worth removing those for testing.
Finally, near where the crash occurs, it appears to be testing for server connectivity. These checks fail for a couple of servers. This may be related to the timeouts I mentioned above.
Other than that, I’m not sure what’s going on. To rule out corruption with the larger set of existing Plex date, you could try:
Stop PMS.
Rename ~/Library/Application Support/Plex Media Server to Plex Media Server.old.
Start PMS.
Perform a basic configuration, perhaps add a single test library.
See how stable it is at this point. You can alway revert to your existing configuration by deleting the newly-created Plex Media Server folder and renaming your old one back.
Thanks. I tried renaming the pluin folder to load without any plugins and no change. I’m trying these steps now. Just adding back in one library at a time and letting it fully scan. If this ends up fixing my issues, I assume I’ll have lost all play history plus any manual edits I made to posters, collections etc? That will be a pain.
Yes, that’s right. I mainly suggested this as a troubleshooting step though. If it works after renaming, it suggests that there’s nothing wrong with the binaries or your configuration settings, just something in the Plex data.
If it works, the next thing I would suggest would be to just move your old database file (com.plexapp.plugins.library.db), and nothing else, to your new ~/Library/Application Support/Plex Media Server/Plug-in Support/Databases folder. Delete the .wal/.shm files first, and do all of this with PMS stopped. Back up the current, new database first so you can put it back if necessary.
Ok, so interesting. it crashed. I decided to start these steps over again and just let it sit without adding any libraries. It crashed.
Here are the logs. Crash around 12:35Plex Media Server Logs_2020-08-20_12-35-39.zip (979.0 KB)
These new logs look fairly similar with regard to the timeouts I mentioned earlier, and these still precede the crash. I’m not sure what’s going on with that. Have you changed anything with your networking configuration lately?
Also, do you use a custom domain for your server? There’s an incoming request to the server immediately preceding the crash with a custom domain:
There may be something about this page request which is causing the server issues. You may want to try restoring defaults on your web client (Settings -> (Plex Web) Debug -> Restore Defaults) in case there’s some bad filter for one of your library views causing an issue. It may also be prudent to do the same for other clients you may use. And if you use remote access and share with others, temporarily disable that.
I recently experienced a issue with latest beta. Refreshing of the EPG used extremely high amounts of RAM and made my iMac Crash 3 times before i worked it out.
In scheduled tasks I unchecked refreshing EPG, no more issues
I did recently switch my router’s dns to use cloudflare (1.1.1.1, 1.0.0.1) but i wouldn’t have guessed that would cause issues. I could turn that off, and use my ISPs dns again.
I am not using a custom domain, but I do share so some family. Which log file did you see that in?
It’s the third-from-the-last line in Plex Media Server.1.log:
HTTP requesting GET https://plex.mistergav.co:443/library/sections/2/all?type=1&excludeFields=summary&excludeElements=Director,Producer,Writer,Role,Genre,Country&X-Plex-Token=xxxxxxxxxxxxxxxxxxxx
I just browsed to mistergav.co to see what it was and it’s just a “coming soon” page. I’m not sure what that’s about or how it’s ending up in your logging. But whois says the IP range that lives in belongs to CloudFlare. I use their DNS too, so I don’t necessarily suspect them as being problematic. And I don’t see any references to that domain in my logging. But, it won’t hurt to test with a different DNS provider.
And EPG is Electronic Program Guide. It’s only relevant if you use a tuner and Plex’s Live TV & DVR feature.
Edit: A Plex Media Server lives at that URL. Not sure how that’s ending up in your server logs.
