vCenter inventory all objects briefly disconnected every so often

[Justins]

I have this really annoying issue where all inventory tree objects within vCenter go into the grey disconnected state seemingly randomly once an hour to a couple times an hour through the day. I briefly recall seeing this one other time but at the time we couldn't troubleshoot it on a different install.

The vCenter servers specs:

Win2k3 Standard R2 64-bit VM

2 vCPU's

4 GB RAM

vCenter 4.1 (previous time I seen this was on 4.0 Update 1 IIRC on Win2k3 R2)

VM stored on EMC SAN (iSCSI)

SQL 2005 DB 64-bit DSN

-

---

The logs also indicate all types of alarms such as storage disconnect, vswitch uplink disconnects etc. What you see on the inventory tree is all objects (for example in the hosts and clusters) view to go into the grey italisized (Disconnected) state. Then they come back up. This brief disconnect states lasts about a second or 2. Enough to see it happen on the inventory tree but then it goes away. This happened during a deploy from template operation and the VM which was deploying from template went into a (Orphaned) grey state. This state lasted even while the disconnect state cleared up. However the copy opreation of the clone job continued successfully with a normal progress bar. Once the clone job finished the VM came back up to a normal state with no issues and not listed as oprhaned any more.

Of course this doesnt happen when I want to capture it with a screen cap but it happens enough I see it once in a while. Also when I was running continuous pings to vCenter no pings dropped during the state of disconnectivity to the hosts/VMs. It seems like a really bizarre issue, any body else encounter this?

J

[jcwuerfl]

I was seeing something similar with update 1 but seems better now that I put update 2 on. What I saw though was 1 host disconnect and all of the VM's assocated to that host disconnected and then quickly reconnected and not ALL objects. However, I also moved my SC network ports back to access ports as I did have those on a trunked port prev. with update 1 so it may have been something with that and spanning tree protocol possibly even though I also had the portfast on. I'm hoping I don't see that again with 4.1.

Different issue from yours that I have is that I also see every once in a while that I have lost access to my volumes being reported and then a bit later it successfully restored the connection? in my vmkernel logs I see:

NMP: nmp_DeviceRequestFastDeviceProbe: NMP device "naa...." state in doubt; requested fast path state update...

Which VMware said that its switching path's and that's normal to see when it does that, however, I verified with my san vendor that there is no issues there, and also with my switch vendor that there was no issues there as well? so no idea what is going on there.

So one question for you I guess is how are your Service Console uplink ports configured? ALso, are you running your vCenter in a VM w/in the cluster, outside the cluster or on a physical server?