Unstable communication L2 and ARP

Hi, I have a very wired problem with one of our F5. This is a single armed partition, so the LB VS and pool members and everything is all on the same L2 network segment.

The thing is, the pool memebers (four) are going down every other minute, and then come back after a while, maybe a few minutes.

Digging into the issue, I found that I am not able to ping those nodes from the F5 tmsh when they are down, while I can ping them from my workstation just fine. Just the F5 looses communication for a reason.

I checked the ARP table, and the entries for those hosts are in there with the right MAC address.

However, when the problem occurs, as soon as I clear the ARP table entry for any of these hosts, I am immideately able to ping them again - for some minutes, then the ping dies again. Clearing the ARP again brings them back to life right away - and so on.

As I said, I can see the correct ARP table entry when the ping is not working, so I dont get why clearing the ARP entry brings them back to life.

All other communication to those hosts is just running fine, e.g. I run a RDP session from my workstation to them which just runs fine while they are not ping-able from the tmsh.

Question is, whats up with the F5 it looses communication. I tried to add static ARP entries for those pool members as I am running out of ideas, but that didnt change anything.

Also, we have the same set up in our dev environment, same F5, same versions, all the same, which just runs fine.

Any help or ideas are appreciated, Tx&Greetings, Jo