Intermittent internet connectivity issue since updating to 6.11

[GriGri]

HI there,

 

I have scoured through the forums, reddit, ... but so far to no avail.  Here is my problem and troubleshooting so far.

 

I have been running unraid for many years with the same config without issue but since upgrading to 6.11 I experience sporadic internet connectivity loss (not a complete loss of connectivity, ie. I can still access the server and the server can access my local network).  I can reproduce it at will by simply enabling or disabling docker.  my dockers are fairly standard (plex, nginx, duckdns, ...) but even when docker is enabled and all dockers are shut down  the problem still happens.  I have eliminated hardware and dns issues, i simply ping 8.8.8.8 and if docker is on it pings for about one minute than stops for a couple minutes than comes back on automatically.  no other device on my network experience packet loss

 

I have now disabled docker and not a single packet loss.

 

Any ideas ?

 

Thanks in advance

[JorgeB]

Are you using custom docker network? If yes macvlan or ipvlan?

[GriGri]

Hi Jorge,

 

Yes, MacVlan.

 

 Most of my containers (code-server, mosquitto, ....) are on bridge network, a couple use host (duckdns, plex) and 2 of them use proxynet custom: nginx and home assistant.

 

My home network is on 192.168.1.0/24, bridge is on 172.17.0.0/16 and proxynet on 172.18.0.0/16

[JorgeB]

Try switching to ipvlan.

[GriGri]

Switched to ipVlan, same issue

 

Tried to disable "Host access to custom network" after switching back to MacVlan as suggested in 

but it did not resolve it.

 

I should also mention that I have 2 ssh connections running at the same type pinging 8.8.8.8.  One to unraid , one to one of the containers (any container) and both stop pinging at the exact same time and restart at the saem time

 

[JorgeB]

Sorry, no more ideas, maybe try resetting all network settings (including all docker network settings) and start over.

[GriGri]

Thanks for trying to help.  Will keep at it 🙂

[GriGri]

A little more info for those brave souls that are following.  I have removed all custom networks, switched back and forth between mac and ip vlan, with or without the "host access custom network" switch.

 

I then stopped all containers, changed them so they do not autostart at docker startup. turned docker off then back on and ... network stayed up.

I turned each container on one by one, starting with the bridge ones, network stayed up

I turned plex (binhex) on and network stopped pinging immediately. turned it on/off back and forth and it is definitely that container, with host network, not bridge or custom

i turned binhex/plex off and installed lscr.io/plex and so far no network issue

 

Will test extensively and see if it holds.  If anybody needs it, I have the diags from before I changed everything

 

[GriGri]

Spoke too soon.  lscr.io plex induces same behavior.

 

I thought maybe it is due to heavy network activity or disk activity so I installed Jellyfin and so far no issues.

 

Something plex does on my install of 6.11.3 but am at a loss

[GriGri]

Solved for now.  At least patched if not solved.

 

I reconfigured lscr.io/plex in bridge mode instead of host and added the following ports per Plex documentation:

TCP: 32400 (access to the Plex Media Server) [required]

TCP: 8324 (controlling Plex for Roku via Plex Companion)

UDP: 32410, 32412, 32413, 32414 (current GDM network discovery)

TCP: 32469 (access to the Plex DLNA Server)

 

So far network is stable