Troubleshooting Network Issues

[spalmisano]

I had a USB failure recently and due to lack of viable backup (since fixed) needed to start with a fresh configuration. I had the drives labeled and didn’t lose any of the array data, and the parity check finished successfully.

What Im experiencing now is a temporary complete loss of outgoing network connectivity when initiating any sort of mass request.

 

Example: while browsing Radarr’s Discover section I’ll add several selections for grabbing. Radarr processes them as expected but after one or two I start getting DNS errors in the logs for ‘resource temporarily unavailable’ trying to hit api.Radarr.video (or anything else).

 

Same goes for Sonarr, NZBGet, NZBHydra, and even the console on UnRAID itself. Doing an nslookup/ping/et cetera on anything will fail. What this feels like is DNS saying ‘too much too fast, wait a few before trying again…’.

 

Sonarr/Radarr will keep trying their requests, and NZBHydra will keep failing and eventually block all of the indexers. After a few minutes the whole thing clears itself up and traffic resumes, but all of the requests made during the blackout are lost.

 

Ive tried both Cloudflare and Google’s DNS, and am still currently on Google’s 8.8.8.8 and 8.8.4.4 and DNS 1 and 2 for UnRAID.

 

What else can I provide that will help point in the right direction for what’s going on? Diagnostics from when this happened just now are attached.

 

Any suggestions are welcomed.

media-diagnostics-20220720-2021.zip

[spalmisano]

Update/Additional detail:

 

This doesn’t seem to impact incoming traffic to the same machine, just the outgoing traffic initiated from a high-volume request. Plex users’ activity is unaffected, but any time I make a request from any of the Arr’s the outbound network traffic ceases for several minutes.

 

This is a Dell r720 that’s been in (my) service for years without issue. Would a bad network cable cause this kind of inconsistent behavior? Bad driver? Anything suspicious in the diagnostics?

 

Not sure where else to diagnose; anyone with a thought?

[CasaP]

I had a similar issue, caused by me trying in vain to get OpenVPN running on my server, purging all VPN stuff got me back.

 

I ended up buying a GL-inet Mango and running Open VPN client on it.

 

Do you have VPN installed, i.e. is the server trying to route through TUN0, but it's not setup correctly or corrupt?

[spalmisano]

Thats a good thought, but no VPN on this system. I do use WireGuard for remote access but its configured on a separate UnRAID server and the server with the network issue doesn’t connect to the WireGuard one. I double checked that WireGuard isn't enabled on this server, and there aren’t any other VPN services/apps either.

 

I am using Pi-hole as a Docker container on this server, and also have a secondary Pi-hole on the WireGuard server. Both UnRAID servers have hardcoded public DNS as their DNS servers and to my knowledge (and the logs) dont hit the Pi-hole containers.