I’ve seen a few other users post on this topic in various subjects. It looks like 1.13.5.5291 has some issues for some users.
I’ll start the server, it will start (according to the logs) but not show up as available n the local server lists.) It will throw all kinds of errors and have issues what looks to be resolving the plex.tv addresses. It will then try again in about 3 minutes. The server will then respond for about 20 seconds then the log will say, “MyPlex: We appear to have regained Internet connectivity.” and at this point the server stops responding to the client.
For the most recent logs (with verbose debugging enabled)
16:06 (approx) I restarted the server
16:06:37 error messages stop for a while, the server does not respond
16:08:41 around this time the server starts responding
16:08:55 (above message) server stops responding to client
Plex Logs(1.13.5.5291).zip (4.2 MB)
Engineering is working on this now.
Please disable VERBOSE logging until requested. DEBUG is all that’s needed.
When enabled, VERBOSE only provides us the last 2 minutes of wall time. This is why we only use it when watching actual server-client handshake & data exchange activities.
If you want to join those other threads, please do. It will help them better than starting new here.
Thanks ChuckPA,
It’s disabled now, I only did it for a few minutes… or at least until the server responded again and I could turn it off. 
I’m assuming there’s an XML file I could alter before I start it (or after I stop it) to alter this behaviour too?
Changing the XML by hand is not something I recommend or advocate.
If it’s up and running, we turn on Verbose, Do that finite test, then turn it off. Then we grab the logs.
We do it this way because the log buffer is only 26 MB (5 rollover files deep)
ChuckPA,
Normally in my line of work I’ll say “WHAT DID YOU DO!” and mean it as an angry question.
Now I’m saying 'WHAT DID YOU DO?" (and meaning it in an amazed way) because suddenly Plex is working, I’m not seeing any DNS resolution errors and the web interface is no longer pausing.
So I’m wondering, was it something You guys (Plex) did? Something I did (or didn’t do?) or did the network gremlins just go home?
If you need some logs or something from me… I (almost) loathe touching my working server, but for Plex it’s always a worthy sacrifice.
Thanks!
BC
There is a handshake between PMS, and “MyPlex” (part of Plex.tv) If they get out of sync badly (which might have been the case for you ) it goes wild. An ISP bouncing IP addresses (more than 2 in a 24 hour period is a good reason to get confused … stuff like that.
It will eventually right itself and, in your case, it did.
So in reply to “WHAT DID YOU DO!”.
I have one reply. “I DIDN’T TOUCH IT! I PROMISE!” lol
a-ha…!
So you’re saying the network gremlins went home. 
Although I didn’t change my firewalls UNTIL I started having this problem. Then I definitely DID cycle through at least 3 different firewalls and therefore 3 different Mac addresses… and 3 IPs… (I’ll do MAC spoofing on them from now on)
Now the only thing that’s left is to figure out why my friends still see 7 “offline Bellerophon” servers and the one that now works. I only see one instance of it in my “device” under servers.
a-ha!
So the truth does come forth!
They can delete those extra servers as well from their left margin (Plex/Web). It’s a “Device” to them as much as it is a server to you.
If not, have Perseus and Cadmus take care of it
Well, that was short lived. I restarted the server and the problem is back.
So we truly did nothing. 
No change. It worked, now it’s back to the “no resolve”
Want fresh logs?
(Oh and the other older server? It was Chimera )
Well Vladimir, What do we do about it? Wait and see what grows?
Yes, Fresh logs would be nice
Fresh logs,
The catch is I have no idea WHEN the server came back online today. I thought I’d left it turned off. But suddenly I saw the server was up and running and running perfectly. I then wanted to re-add a channel (trakt) and it went downhill from there.
I asked a few external people to test and it worked for everyone.
2018-08-09 Plex (Bellerophon)Logs.zip (4.7 MB)
Is this where I tell you to not touch it? 
On a serious note,
If you want to increase the log file retention length to allow more time, it’s easily done.
The default is 5. We can make it 10 by adding LogNumFiles="10" within the scope of the XML statement in Preferences.xml when PMS is stopped. At next start, it will obey the new log file retention value.
Do you have PMS set to listed to a specific interface?
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - Network interfaces:
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - * 1 lo (127.0.0.1) (loopback: 1)
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - * 100 br0 (172.16.0.48) (loopback: 0)
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - * 101 docker0 (172.17.0.1) (loopback: 0)
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - * 119 virbr0 (192.168.122.1) (loopback: 0)
Aug 09, 2018 17:58:19.479 [0x14e99bd61800] DEBUG - Creating NetworkServices singleton.
If you don’t have it set, it can easily drift to the other adapters.
Ya, I touched it…
I Set LogNumFiles = “20” I have a ton of space available.
How would I set PMS to a specific interface? Would I do this in the docker declaration? I have tried “ADVERTISE_IP” as ‘172.16.0.48’ which is the actual “external” IP and port of 32400 via what looks like a bridge (br0) interface.
I could also re-built my VM with 16.04 LTS and reproduce this without docker.
This seems to have done something “ADVERTISE_IP” = 172.16.0.48
so far it’s been stable…
Edit: “nope”
I’m thinking you’re right it’s latching on to the wrong IP address. It’s currently as a “host”
If this works, then leave it.
IF you have to take action again, i really recommend the native app… Docker causes so many problems because what it thinks is a public IP … ISN’T. It’s private LAN. It makes a mess.
Had I known you’re using Docker, I’d have slapped your fingers early 
lol
Remember? Docker + Remote Access != pleasure lol
It’s not working now, I’ve been using it for 9 months without issue until last week when it started this with what appears to be the latest upgrade.
As I said though, I can also reproduce this with a different VM. I’m wondering if I should try Windows to see if it has the same connectivity problems.
As of the latest upgrade, the internal network handling changed.
If you notice, you actually “Select” which ethernet adapter to Listen (bind) to.
This is they key.
As experiment, create an ethernet adapter alias for your computer’s ethernet adapter.
Now you can cycle through which adapter to bind to because you’ll have access to it via eth0 and eth0.1 (the alias)
Really? I can’t recall ever having seen this before, I’m even spinning this up in an ubuntu 16.04 LTS and I’ve not seen it say anything during setup with the NIC… is this in server settings?
PMS 1.13.5
Settings - Server - Network - Show Advanced
I’ve been pushing for this for almost 2 years now.
