Plex Constantly Loses Remote Access

[bigblueshock]

Thanks for quick response. I have not posted logs yet. I have grabbed them from both my Servers.

Nothing has changed on either server, or either household (other than server version updates). One house has a G1100 Router (which was a pain in the ass to set up, but I finally got it and checked port from outside, and it works) This is the smaller log file.

The bigger log file server is hooked up to a Netgear R7800 with port forwarding correctly set up.

Not sure if it’s a coincidence, but it seems like the timeline issues started lining up when people started having problems on this thread.

Plex Media Server Logs_2018-12-07_08-16-58.zip (621.1 KB)
Plex Media Server Logs_2018-12-07_08-17-31.zip (2.5 MB)

[sa2000]

@bigblueshock

Thanks for quick response. I have not posted logs yet. I have grabbed them from both my Servers.

Nothing has changed on either server, or either household (other than server version updates). One house has a G1100 Router (which was a pain in the ass to set up, but I finally got it and checked port from outside, and it works) This is the smaller log file.

The bigger log file server is hooked up to a Netgear R7800 with port forwarding correctly set up.

Not sure if it’s a coincidence, but it seems like the timeline issues started lining up when people started having problems on this thread.

By the way you have debug logging disabled on Bitfenix.

I am looking at the logs from the other server and also will see what info i can get from the Bitfenix server log even without debug

First i need some clarification / confirmation

The two servers are different locations
The MarkRoomPC server which has a local IP of 192.168.1.125 is on WAN Public IP starting 98.113.xxx.xxx and has port forward of 32400 to 32400 for 192.168.1.125

The Bitfenix server has local IP of 192.168.1.150. I am seeing in the MarkRoomPC logs connections and requests from two different WAN IP Addresses but i think they both relate to Bitfenix. There is a WAN IP starting 108.54.xxx.xxx and there is also one starting with 72.80.xxx.xxx - are these for the same machine ? vpn setup? or multi WAN connections? or different machines?

[bigblueshock]

My apologies. I will enable debug logging on bitfenix.

You are correct. Both servers are in completley different locations/households/internet connections. There is no VPN set up. At this point in time, the only reason i use MarkRoomPC is for OTA live TV. Because i do not get ABC station where my main server is. (Basically, markroomPC is at a family members house closer to the city, where bitfenix is at my house).

Those local IP addresses are just DHCP static addresses for each house/router. I set the reservations so I could do manual port forwarding.

To answer your last question… i may have answered it above. Completley different machines, different WANs, but same Plex account.

[sa2000]

I now understand the references to WAN IP Addresses 108.54.xxx.xxx and 72.80.xxx.xxx - The Bitfenix server had its WAN IP address change sometime between November 29th and December 1st.

With regards to the MarkRoomPC server, it was first launched on version 1.14.0.5470-9d51fdfaa at 12:40 on Nov 27, 2018. By 12:42 it was available for remote access.

The logs show some issue with communication between the PlexScriptHost.exe process for com.plexapp.system and the Plex Media Server.exe process on 127.0.0.1 - Lots of errors like this which is not normal - may be some adware proxy software installed itself on this machine?

2018-11-27 17:59:25,750 (16c) :  DEBUG (networking:166) - Requesting 'http://127.0.0.1:32400'
2018-11-27 17:59:25,763 (16c) :  DEBUG (core:524) - Unable to read server attribute 'serverVersion'

Requests come to the server on tcp port 32400 and if there is anything that intercepts such requests, then outcomes could be indeterminate

If you can search this log file
C:\Users\Mark\AppData\Local\Plex Media Server\Logs\PMS Plugin Logs\com.plexapp.system.log for text Unable to read server attribute and find a recent log entry for it - like the day you check, then please do the following

• Start an elevated command line window : Start / Run / cmd and right click and select Run As Administrator - then in the command line window, type the following command
  netstat -a -b -n > "%LocalAppData%\Netstat_2018-Dec-08a.txt"
• Open settings for Internet Options and select Connections tab and click on LAN Settings and see if the Proxy tick box is ticked. if it is, click Advanced and take a screenshot
• Need to get the xml for a request. First find the server token - see here fo how https://support.plex.tv/articles/204059436-finding-an-authentication-token-x-plex-token/
  Then open a browser on the server PC and type in command
  http://127.0.0.1:32400?X-Plex-Token=xxxxxxxxxxxxxxx substituting the token string for the xxxxxxxxxxxxx
  Copy the response to a text file and send me in a private message

You can attach the "%LocalAppData%\Netstat_2018-Dec-08a.txt" file here together with any screenshots for the intermet options

If the com.plexapp.system.log does not have this text Unable to read server attribute - then wait until you have issues and check then

The logs for this server show that at 23:14 on 5th December, a Plex Relay connection was started - So a client had difficulties getting to the server through the remote access route and decided to try Plex Relay. The logs show previous remote connections requests made at 15:14

We know the issue with requests to port 32400 were present at 18:01

2018-12-05 18:01:40,576 (13c0) :  DEBUG (networking:166) - Requesting 'http://127.0.0.1:32400'
2018-12-05 18:01:40,598 (13c0) :  DEBUG (core:524) - Unable to read server attribute 'serverVersion'

although a plex client app started relay at 23:14:42, i do see successful remote connections coming in to the server eg at 23:14;45 from your other server Bitfenix -
When you opened the remote access settings page and decided to disable and re-enable remote access at 23:16 - it was actually working as a connectivity test just before disabling remote access got through ok at 23:16:02
So probably it was not down - did you try several remote clients? different mobile devices through cellular ?

What followed was hitting a known timing bug that happens sometimes when enabling remote access. It resulted in failure to establish remote access after disabling and re-enabling. And i believe you did a Retry and had same issue. Sometimes rather than retry, a browser refresh with F5 resolves the issue

You then restarted the server at 23;17 and i can see a network issue with dns failures between 23:19 and 23:20 - you probably reset the router then

So we have a few things to follow up as described above. Not clear as to what the issue was

With regards to the other server Bitfenix - server was launched Nov 27th 12:40. We know the WAN IP address changed between Nov 29th and Dec 1st. Without the debug logs, we do not know how Plex Media Server handled that changeover and if there was a glitch and how long for. We know from the other server logs, that it changed sometime after 22:43 on November 29th.

Logs show internet / DNS down during these periods
Nov 28th 16:31 - 16:32
Dec 1st 09:13 (may be when WAN IP changed)

Will wait for debug logs from this server

[pushnoi]

I noticed the same behaviour. I recently moved from a cloud hosted server @ Hetzner (with a direct public IP) to a @home hosted server behind a router (I’m running all Ubiquiti Unifi devices; using a static IP with port mapping and no UPnP). I’m running the official docker. It is running as ‘host’, network wise. I’ve been experiencing these remote access disconnects as well. I do have Tautulli running though. Running latest plexinc/pms-docker:plexpass.

I’ve set the network device from ‘All’ to the specific device it should be listening on and I’ll turn on debug logging. I also enabled Tautulli push messages for remote access (great tip!). If disconnects keep happining I will post logs here.

[sa2000]

@pushnoi

also enabled Tautulli push messages for remote access (great tip!)

Make sure the interval between Tautulli connectivity test requests is not too short - I don;t know if it is a configurable parameter now but the lines to edit to chnage the default was provided here Random loss of remote - hundreds of pubsub async errors - #24 by SwiftPanda16

Suggest every 30 minutes

[Botnick]

Yes please fix this ASAP!

[Forced-Motorworx]

this has been driving me nuts for months now, im paying for this service and it constantly drops out and i have to reset it, if there is a problem with everyone’s “configuration” why hasn’t anyone been sent maybe an email explaining the issue or the fix? instead we have to carry on trying to figure out what the hell is going on, i just want to use my plex that im paying a yearly subscription for.
I am not some IT guru who knows how to reprogram a modem and i shouldnt have to, this is supposed to be an easy to use interface not something you need an IT engineering to back end program all sort of port forwarding configurations and what have you.
What button do i press to fix this thing.
signed
one long time loyal but seriously pissed off customer

[pushnoi]

Tnx, since I’m running tautulli in a docker (and could not find this in the .ini), I haven’t bothered with changing this. Just disabled it for a couple of days to see if things where different. I still have problems however. Attached some logs, hopefully you guys can see what’s going on.
Plex Media Server Logs_2018-12-14_08-53-06.zip (2.9 MB)

[Falkenfighter]

Hey I appear to be having the same issue as everyone else in this thread. For the last few months my Remote Access will connect, then immediately disconnect. Historically I’ve been using port forwarding on 32400 to my static IP however, I recently tried UPnP with no luck. I’ve attached the logs from my Ubuntu Plex server.

Plex Media Server Logs_2018-12-18_22-03-02.zip (2.7 MB)

Thanks!

[mastachaos]

i too am experiencing this bug.

I have another thread going, but I’m posting here so I can follow up on this one too.

In the meantime, setting up a reverse proxy for Emby was a PITA, but now that it’s done, it’s way more reliable for me, and I’m considering making the jump…

[sa2000]

@pushnoi

Thnx, since I’m running tautulli in a docker (and could not find this in the .ini), I haven’t bothered with changing this. Just disabled it for a couple of days to see if things where different. I still have problems however. Attached some logs, hopefully you guys can see what’s going on.

Thanks for the logs.

I do see in the logs instances when Plex Relay was started by client apps but the use of relay is not sustained and lasts only for 1 or more seconds - so these are just transient - possibly due to the server being slow in responding and the client apps try different route or there is an issue with the client app and it is starting a relay connection when it does not need to.

I will summarize what i can see in the logs.

1. Logs covering Dec 13, 2018 15:46 to 23:38 when server was shutdown.
   Plex relay connections were used a number of times but only for short periods, lasting for maximum of 10 seconds. No indication of any issues. Remote WAN connections were active throughout the period

2. Logs covering relaunch at 23:53:45 to shutdown at 23:54:06
   Server was available for remote access on manually specified port 32400 as from 23:53:58

3. Logs covering relaunch at 23:54:44 to shutdown at 08:48:53 14 Dec
   Server was available for remote access on manually specified port 32400 as from 23:54:56
   A few relay connections were established but transient - max of 2 seconds
   WAN connections were active throughout the period. Cannot see any issues

4. Logs covering relaunch at 08:48:57 to 08:53:05 when logs captured
   Server was available for remote access on manually specified port 32400 as from 08:49:08
   There were active wan connections. Cannot see any issue

I need to know at what time there was an issue and if it was a real issue i.e. a client app failed to get through to the server remotely and probably best to get client app log as well at the time - as well as server logs

See this section for logs for client apps
https://support.plex.tv/articles/201869908-log-files/

(just noticed that you have now switched to uPnP automatic mapping - whilst normally this is less reliable than a port forward and manually specified port - in some cases it offers better results, eg as in my case when my ISP has been resetting my router quite frequently recently losing all port forwards and DHCP range definition)

So to summarize - i do need to know when remote access was seen to be down and confirmed to be down and not just the status indicator being wrong (and for the logs to be captured then)

[sa2000]

@Falkenfighter

Hey I appear to be having the same issue as everyone else in this thread. For the last few months my Remote Access will connect, then immediately disconnect. Historically I’ve been using port forwarding on 32400 to my static IP however, I recently tried UPnP with no luck. I’ve attached the logs from my Ubuntu Plex server.

Thanks. Which IP address is the port forward for 32400 to?
Seeing some reference to a 10.xx.xx.xx subnet in the log
Could you also get me the output from ifconfig -a

Is there a vpn ? proxy?

Dec 18, 2018 21:41:51.096 [0x7fc71a857840] DEBUG - Detected primary interface: 10.22.10.10
Dec 18, 2018 21:41:51.096 [0x7fc71a857840] DEBUG - Network interfaces:
Dec 18, 2018 21:41:51.096 [0x7fc71a857840] DEBUG -  * 1 lo (127.0.0.1) (loopback: 1)
Dec 18, 2018 21:41:51.096 [0x7fc71a857840] DEBUG -  * 3 enp0s31f6 (192.168.1.70) (loopback: 0)
Dec 18, 2018 21:41:51.096 [0x7fc71a857840] DEBUG -  * 6 docker0 (172.17.0.1) (loopback: 0)

I need to establish why PMS made reference to 10.22.10.10 and in an earlier launch to 10.94.10.6

What is the route to your server from your public IP 64.187.xxx.xxx and wan port 32400 ?

There was connectivity test that got through but another at the same time did not.

Dec 18, 2018 21:55:51.035 [0x7fc70bfff700] DEBUG - EventSource: Got event [data] '<Message address="64.187.xxx.xxx" port="32400" asyncIdentifier="79f62c79-f0a6-4b0c-82ea-81deac0297a0" connectivity="1" command="notifyConnectivity"/>'

Dec 18, 2018 21:55:56.033 [0x7fc70bfff700] DEBUG - EventSource: Got event [data] '<Message address="64.187.xxx.xxx" port="32400" asyncIdentifier="46d319c1-2db8-4eb4-b283-981ee2ae12ff" connectivity="0" command="notifyConnectivity"/>'

[pushnoi]

Okay, so I’ve noticed the same behaviour again. Attached the most recent logs. I think I checked if PMS was reachable at around 22:45. It’s now 23:26. When I got to Settings => Remote Access I get the infamous red exclamation mark. I was watching some content while on the same network.

Noticed the issue @ ~23:15. At around 23:22 I re-enabled remote access by pressing ‘retry’ in Settings => Remote Access. It’s still set to 32400, port forwarded.

Before I re-enabled I noticed some strange behaviour while testing from outside. I used my Android phone on 4G => when opening the app I noticed nothing was accessible. I tried starting something but this obviously did not work. But now the strange part => I opened my mobile browser, went to plex.tv/web and low and behold - my PMS is reachable from there. Could even start content. While still having the exclamation mark mind you. Weird. Tried the app again, and nothing. Even after re-enabling the remote access I had to force close the app and start it again to get it working.

Notice I have Tautulli running again because leaving it off did not seem to do anything for me. I’m not getting any notifications about not being reachable from the outside from my agent btw, even though notifications are working (tested).

Plex Media (4.0 MB)

In thise case when I’m at my local network going to plex.tv/web on my laptop does seem to work. But sometimes I can’t reach it via that way at all. I ususally end up restarting the docker. But I can also go directly to my open port via http://:32500/web/ and re-enable remote acess that way btw.

Edit: And there it goes again…

New logs to include latest issue…
Plex Media (4.2 MB)

[sa2000]

@pushnoi

But now the strange part => I opened my mobile browser, went to plex.tv/web and low and behold - my PMS is reachable from there. Could even start content. While still having the exclamation mark mind you. Weird. Tried the app again, and nothing.

I will look at the server logs - but what you are saying here requires getting logs from the android app as well (with of course corresponding server logs and indication as to what happened and when)

See Section on emailing zipped logs to your mailbox
https://support.plex.tv/articles/201573117-android-android-tv-fire-tv-logs/

[3090317]

I was having the same issue, remote access going from Green to Red (red most of the time). I was using Geo Blocking and once I disabled it remote access stayed Green. I have re-enabled blocking but only the 10 most troublesome countries and remote access functions is still working like it should.

What particular countries must not be blocked for remote access to work properly? I normally block all except the US.

[autolex]

How are you blocking Geo access? i just started having this issue (several months) after I replaced my router with a Ubiquiti USG Device and started geofencing all but a few countries of origin.

[3090317]

I use IPFire as firewall/router and currently have these countries along with Russia, Germany and the UK GeoIP blocked. When I a block all but the US I lose remote access. I would prefer to block all but the US and any that are specifically required by Plex if I could determine that those would be.

I can’t seem to locate that information and it would appreciated if Plex could provide it.

[jester_tenet]

I suspect (and I’m happy to be contradicted by Plex) that the system uses Content Delivery Networks (CDN), which likely do not respect geo-locations as you expect. This would break either the location blocking or remote access.

The spamhaus list is based on spam-bots, so might not actually be the risk-threat profile that you are looking for. Depending on the risk-threat profile you are aiming for (happy to hear what that is if you specific plex attack data), you could go as far as blocking the US, as those ranges could be the greatest threat (again, breaking CDN).

[3090317]

I believe you are correct about Plex using CDNs. Regional Geo Blocking (RGB) should still be possible. I was wondering what Points of Presence (PoP) need to be open for Plex remote access to work reliably. I know that many regions can be blocked, that is how I currently my network configured. I am adding regions one and a time and see when things break, but was hoping that specific region(s) could be identified.

One suggestion would be to allow mobile apps to connect like browsers can currently connect, pointed directly at our domains. The need to connect through plex.tv is not as much of a issue in that case, only for authorization and subscription status. Mobile apps do not appear to allow this type of configuration and use our plex.tv accounts to phone home. If it were possible in those apps we could minimize that traffic and it would be a win-win.

I host other services on the same networks that contain my Plex Servers and I drop 1000’s of connection attempts daily from all over the world. It is not really any specific Plex threat, it is other services that I try to limit access to.

I prefer to try to limit access to my networks. I don’t have any need for connections to be established from outside the US, so I would prefer to drop them. It would be ideal to block all, including the US, and only allow specific IPs, but that is not really practical.