Plex Server Continually Crashes - MAC OS - UPNP - SSDP hanging on Plex. Effects versions 1.10 - 1.13

Plex Media Server Desktops & Laptops
1

Hi,

I continually have crashes that are causing my Plex to hang randomly after UPNP requests that timeout on my local network. The odd thing is that the UPNP request are coming from my XBOX One (10.0.1.114 in the attached logs). The reason this is odd is that I removed the Plex client from the XBOX One so I have no idea why it is pinging back to my Plex server for an addressable location.

This behavior continually results in my Plex server crashing. The last stable build I’ve found is 1.9 that doesn’t seem to mimic these issues. Other than that build, I tend to have crashes once a week or so.

This is getting quite frustrating and I’m hoping that I can get some guidance on the best route to go to solve this.

I’ve attached logs of the hanging incident, as well as my restart of Plex below:

"Jul 01, 2018 01:06:50.017 [0x7000036ee000] DEBUG - HTTP requesting GET http://10.0.1.114:2869/upnphost/udhisapi.dll?content=uuid:3c75de5d-c7a5-4192-b350-9e416d206fb4
Jul 01, 2018 01:06:50.031 [0x7000036ee000] DEBUG - HTTP 200 response from GET http://10.0.1.114:2869/upnphost/udhisapi.dll?content=uuid:3c75de5d-c7a5-4192-b350-9e416d206fb4
Jul 01, 2018 01:07:51.725 [0x7000036ee000] DEBUG - Checking if time for scheduled update
Jul 01, 2018 01:11:51.665 [0x70000289a000] DEBUG - NetworkServiceBrowser: SSDP departed after not being seen for 28.635988 seconds: 10.0.1.114
Jul 01, 2018 01:12:51.726 [0x700002ebe000] DEBUG - Checking if time for scheduled update
Jul 01, 2018 01:16:30.538 [0x7000036ee000] DEBUG - Sync: uploadStatus
Jul 01, 2018 01:16:49.559 [0x700002ebe000] DEBUG - PublicAddressManager: Obtaining public address and mapping port.
Jul 01, 2018 01:16:49.559 [0x7000036ee000] DEBUG - PublicAddressManager: Obtaining public IP.
"

For the record, I’m manually setting my own static IP address/Port for Plex.

Please let me know what other information is needed. I’d love to know if others have had this issue as well.

2

To clarify further, the only way I’m able to get Plex to no longer hang - it never actually shuts down, I need to actually end the server since it becomes unresponsive and appears offline - is to quit the application and relaunch it.

Full logs available upon request.

3

can’t see any attachments to your post. Please attach the logs zip

If it is crashing in the routine that is processing the response from uPnP request to the router for the public IP then a wireshark capture would help see what is coming back from the router

But would need full logs after the restart following the crash and of course crash reporting has to be enabled beforehand (in server settings)

4

Plex Media Server Logs_2018-07-01_09-29-48.zip (6.2 MB)
Plex Media Server - Crash - 7:1.log (3.1 MB)

Thanks for the fast reply @sa2000! I’ve attached the log zip, and the specific log that I noticed the issue in.

I’ve got crash reporting enabled so hopefully you can find that info.

Right now I’ve downgraded to 1.9 but if you need me to update to 1.13 again and see the issues, let me know.

If captures are needed, also let me know. Have a wireshark equivalent that I can use.

5

Away from my pc at the moment but if the log stops at obtain public IP address before each crash and you can reproduce it, I would love a wireshark filtered on router IP address with logical or for pms ip addresses saved as wireshark default pcap file and sent to me by private message together with corresponding server logs

6

I can do the PMS capture.

Timing/reproducing it is hard and leaving Wireshark running all the time is a challenge but I’ll try and reproduce when I get a chance this week.

Hopefully the logs attached will be a good starting point.

Appreciate the rapidity of responses!

7

I will have a look at the logs and locate the uploaded crash report later this evening

8

Thanks for the logs. There was only one log file for version 1.13.2.5154

It was actually not a crash. The reference to public address upnp request at the end of the log was just a coincidence and not relevant to the issue.

The problem was a lockout that happened hours earlier - at 21:52:51 on 30 June. Between that time and the time log was picked up at 01.17 am on 1 July, no requests were handled by the server

I do have a development build I can make available to you to try - it has the fix for a lockout that is similar to what happened at 21:52:51 and may well be the same issue (for lockouts / deadlock one would need process dumps - on a Mac obtained through Activity Monitor) to supplement logs - see https://support.plex.tv/articles/201373203-gather-a-process-dump-or-sample-process/

I will send you a link to download this build - it is a pre 1.13.4 beta

9

Thanks @sa2000. Will give it a shot.

10

I’m having a similar issue - and I’m seeing the same messages in my log. Any chance I can try out the pre 1.13.4 beta build?

11

Please attach the logs zip

12

@sa2000 noticing one oddity around the latest test release you provided - seems like I’m either getting a false positive on remote access disconnecting according to the web UI - or I’m actually disconnecting on my remote access. I don’t see any alerts from my Log analysis tool (PlexPy) nor do I see anything in logs indicating such behavior - but I did log into the web portal and was notified that the server was not externally accessible.

Can DM you logs if you’d like but I’d enabled verbose logging so not sure if that is helpful.

Advise on how you’d like me to proceed.

Thanks!

13

If the logs cover the time when it was showing as unavailable and include the checks we do then fine - I can have a look.

14

Sounds good. I’ll dm them over in the morning.

15

@sa2000 So, it seems to be just the WEB UI that is showing remote as unavailable. If I go into the application over LTE, or check via the web over remote networks, it is all fine. I’ll send the logs so you can see the latest. If you need a screenshot of the portal, I’ll also attached that in the DM.

16

Please do. What is important is the time

Can you explain. Web UI shows wrong indication of status. But rge application over WAN / Cellular ok - which application? If all are using Plex Web - which url’s give a true indication and which do not (at the same time) ?

The status indicator for remote access has not been 100% accurate for some time. There are timing issues and leaving the screen without clicking on retry button repeatedly would normally correct itself after a browser refresh with F5

It is possible that there is now different timing with the new version

17

Both the web browser/Plex/Web and the Plex client on iOS.

None of the pages give a true indication…both the /settings/server/{key}/General and /settings/server/{key}/remoteAccess show up as “red” remote access not available → despite true status of being available.

This has been on all day - so, not like past where clicking on “remoteAccess” refreshed the UI.

18

Thank you - i suspect this was caused by changes to some pubsub plex.tv servers where some servers were replaced. These servers take part in the connectivity tests for the remote access status

I will look at the info you sent me by PM

19

This has now been confirmed as due to the recent migrations of pubsub plex.tv servers which led to some Plex Media Servers not getting the relevant events for remote access connection testing invoked by Plex Web. Steps have been taken to correct that.

20

@sa2000 Just to follow-up, things seem to be ok with the build you provided for quite a while but it looks like I recently had an old issue pop up - the SSL certificate/secure version of Plex becomes unable to be reestablished after 2 days.

I have the “green lock” on plex.tv for about two days before a flip over where the only way to reach my server is via http only. I’m not sure if it’s related to the build or not. A simple restart of the server seems to work. I can download and provide old logs if needed.

I’ve fixed this in the past with DNS adjustments and ensuring that my DNS provider enables DNS Rebind. If that is the same cause, then it would mean that the documentation on the main page may not be up to date as Cloudflare’s DNS (1.1.1.1, 1.0.0.1) is now the provider I use (the last time this happened, I was local before pointing everything to 8.8.8.8) and they’re listed on this page as working fine with the DNS rebind setup - https://support.plex.tv/articles/206225077-how-to-use-secure-server-connections/

Any thoughts/assistance would be appreciated.