Service Console network slowness (3.0.2, 52542)/Host Agents periodically crash

[pweinlein]

We have in the last few weeks been experiencing issues with some of our ESX host agents periodically crashing. They restart themselves a minute later, but I am not really seeing much in the VPXA.logs. Also, we've noticed our PINGs between hosts on the same subnet take anywhere from 2-8ms, which seems slow. Has anyone seen this behavior before? Think the slowness in the Service Console network is related to the periodic host agent crashes? We are planning on upgrading to 3.5 in two weeks. Should we push up this timeline? Thanks for any advice.

[RParker]

Upgrading to 3.5 isn't going to fix your network problems, if that's what is causing the problem. It's probably hardware related with your switches, if its all the hosts. I would start with identifying where the network congestion is.

[pweinlein]

We've had one incident where all of the host received "Host connection state - (State = Not responding) " at the same time which we feel is definitely networking related (although our networking folks said they see nothing from the switches at that time). However, we mostly see this occur a few times a week, & it could be with any of our 14 hosts.

Although it seems networking related, could service console memory allocation be a factor? It is configured for 272MB & we have 130 VMs running on a 5 server cluster.