6.12.6 every week or so will lose all incoming network connectivity for all services/outbound ok

[cr08]

This has started probably within the past month and seems to happen on roughly a weekly basis, but doesn't seem to have strict timing. What happens is all inbound network connectivity is seemingly lost. Pings work and using a port scanner it shows all my open ports, docker container ports, etc. but nothing can be accessed. Outbound seems to work fine. I have some scripts that access external resources that continue to work during this apparent dead time. I've also had syslog pointed at a Pi that has continued to send stuff over normally as well (sadly it looks like it went offline during this last occurrence so I don't have anything to provide in that area unless the attached diagnostics include something).

 

Around the time this started occurring I did notice I got an alert about the whole macvlan/ipvlan thing. It never seemed to cause issues before but I went in and did the suggested changes but it doesn't seem to have helped here.

 

This is a Dell Poweredge T20. Mostly stock. Xeon E3-1225v3 and a Perc H200 with a number of drives. Also a GTX 950 for transcoding duties. Unfortunately the 950 makes things a bit of a pain since Dell's design of this machine disables the iGPU with a dGPU installed and no config in the BIOS to change this. This also kills the KVM access via IPMI since it relies in the iGPU. So when this happens I'm basically forced to do a hard restart since there's no soft reset option via IPMI. Really frustrating given the system seems to be functional otherwise.

 

Diagnostics are attached. Hopefully someone has some good insight on this. Such an annoying bug to deal with. Haven't done much with it so far beyond the ipvlan changes and haven't had a clue where to start troubleshooting.

nas-diagnostics-20240203-2055.zip