Server Version#: Cant get to the server because of the network storm. But less then a month old (27/12/21)
Player Version#:
Have no idea why this is happening, but Plex is basically spamming this IP address. Here is the Wireshark trace (or a little part to illustrate the point)
It basically makes the windows server its running on unreachable and stops devices on the network from working.
How can I stop this?
Thanks
How do I contact Plex support, I’m paying for this software and now the main component is unusable.
I reinstalled the latest version of PMS and it seems to start happening when I add a library.
Can anyone tell me why?
How frequently are the packets sent?
Plex looks for clients using multicast at least every 10 seconds.
Thanks mate…
Its waaay more then every 10 seconds. That first column there is time, its pumping them out so hard and fast that its swamping the network and cannot access the server.
I did get it stable-ish for a while but I deleted and reinstalled my Shield TV client, as soon as I did that It just started going crazy again.
I just don’t understand how even a client can set it off, I’m looking at WireShark right now and its the same server to multicast packets. Hundreds of thousands.
EDIT: Sorry, first column is number of packets. the second is time 
Please provide some server logs.
Just to add the Wireshark packet text.
^ú0#mBE=rÊÀ¨¾ïú~~)ÍÍHELLO * HTTP/1.0
Content-Type: plex/media-player
Device-Class: pc
Name: Chrome
Port: 32400
Product: Plex Web
Protocol: plex
Protocol-Capabilities: timeline,playback,navigation,mirror,playqueues
Protocol-Version: 3
Resource-Identifier: diilmmto59fppk6a4so4xrxu
Version: 4.66.1
I cant get logs because I cant get to the server.
That first number from the capture is the number of packets… I started that about an hour ago.
Its just a flood.
See the section “Manually Finding Log Files”
Thanks mate, but when I say I cant get to the server I mean, I cant use a UNC path or RPD to it.
I just checked the capture again and what I said before was way off.
it went from 0 to 164680 of all the same packets in about 7 seconds
Its a headless windows server, is there anything else we can do first? otherwise Ill have to try and pull it our and connect it to a monitor ect tomorrow
Like the NIC is being that hammed that much, I cant even connect to it from another NIC (Hyper V V-switch) on the same hardware 
Ok I managed to get the logs this morning.
Plex Media Server Logs_2021-12-29_11-38-28.zip (6.6 MB)
Please advise?
Just to add to this. I reset my Shield and reinstalled Plex client and its doing a very similar thing,
The client itself keeps freezing/ not connecting to the Plex server and when you look at the capture, its a storm.
What the hell is going on with Plex ?
So is there a support team or what? I mean I’m paying as much for others services per year that actually have a support team, that actually have SLAs
If your not going to do it, thats fine. But its time for both of us to move on.
Do you have Mac spoofing enabled on the VM Vnic?
Do you have bonding of some sort running between your hyper-v server and your physical switch(es)?
Do you have both wifi and ethernet enabled simultaneously on the Shield?
Any non default IGMP settings on your physical switch?
What’s your network setup look like?
Hey thanks for getting back. I’ve put in some comments here, I could have been a little more clear.
I just wish I knew which log was for this kind of stuff. I looked in the system log, But cant see anything outstanding there.
Both NICs on the same subnet? Do both physical NICs connect into the same Vswitch and then you assigned different ports on that vswitch to the hypervisor and VMs? (I’m not 100% sure on Hyper-V these days, so hope that’s the right lingo)
Where is the old PMS VM, still hanging around or is it dead and buried?
Deep in the bowls of networking land, your PMS logs show nothing wrong so my guess is you might have a layer 2 loop somewhere causing the multicast to have a heart attack.
Single cable between the 2 switches?
Some of the earlier packet capture information posted looks like Plex’s GDM multicast messages. It’s the older way servers and players found each other on local networks and is not strictly necessary any longer.
If you can get to Plex Media Server’s settings, navigate to Settings → Network → Enable local network discover (GDM) and disable it. You may also want to disable a similar setting in any web clients you use: Settings → (Plex Web) General → Advertise As Player.
If you’re unable access settings, you can disable the server’s use of GDM via the Windows registry:
HKEY_CURRENT_USER\Software\Plex, Inc.\Plex Media Server\GdmEnabled = 0
I’m almost positive a server restart is required for this to take effect.
This should minimize some of the broadcast traffic from Plex Media Server and its clients, but there’s no way to completely eliminate it short of configuring (Windows) firewall rules to completely block it. However, If you have some sort of loop issue, that will need to be corrected.
Thanks guys! I really appreciate this so much. Right between Christmas and NY with no Plex is pretty crappy LOL.
To have two devices doing the same thing, it must be a loop right. Its not really my area and I discounted it initially because I just didn’t see where it could be. But maybe one of the switches is no good. I’ve ordered a new core switch (just a 24p GB TP link from Officeworks) I swap that out this morning and see if it helps at all.
Both NICs on the same subnet? Do both physical NICs connect into the same Vswitch and then you assigned different ports on that vswitch to the hypervisor and VMs? (I’m not 100% sure on Hyper-V these days, so hope that’s the right lingo)
Dont worry about the lingo, I cant remember half the time either LOL. Both NICs are on the same subnet, but one is the v switch and the other appears as a standard NIC to the server
Where is the old PMS VM, still hanging around or is it dead and buried?
Totally dead and removed
Deep in the bowls of networking land, your PMS logs show nothing wrong so my guess is you might have a layer 2 loop somewhere causing the multicast to have a heart attack.
Thanks for taking a look, ill check out that doco later. But if it says nothing wrong, then, again a loop must be the culprit
Single cable between the 2 switches?
Yes, ethernet though the walls
Thanks so much for writing that up. It must be a loop right, as I said before, two separate devices doing the same thing…
But ill hang on to this and keep it as a last resort if I cant find this loop.
Ill let you know how I get on.
Thanks both for spending the time to help in a pretty stressful time lol 
Just wanted to thank you guys for pointing me in the right direction!
It was a loop in the end, but interestingly only really manifested itself when I did a fresh install of Plex. I guess the VM switch was stopping the storms before, but the network has had performance issues for sometime. But because it was so low level I couldn’t pin point it till now. Then you start looking in the wrong direction ect ect ect LOL
It the end it was the auto backhaul between the mesh routers/APs that was causing the loop. It wasn’t actually working, setting it explicitly to ethernet backhaul seems to have everything working now.
So thanks again guys, see how the network goes over the next few days, but I’m really hopping to see much improved performance, especially for Home Assistant stuff
Out of curiosity, what were the devices? How are they connected within the network?
“Automatic mesh” sounds like another way to say “maybe loops”.
It was the back haul for my ASUS mesh wifi. I had the back haul set to auto. Thinking that would be the best setting. But apparently you need to specify Ethernet. You think it would tell you of a possible loop or something. But there you go 
