Hello,
I upgraded vCenter from 2.5 to 4.0 last week. I also upgraded VCB to 1.5 U1. All went well. My hosts are still 3.5 U3. My vCenter server is a VM, running in the cluster that it manages.
Last night, the 'Host connection state' alarm triggered for on of my ESX hosts while the VCB snapshot was being removed for the vCenter server VM. The email alert for the alarm was triggered at the same time the snapshot was being deleted. I have a two host cluster, and the alarm was for the host that was not running the vCenter VM.There is no event for the host disconnecting, the host remains connected and all is fine. The snapshot was successfully deleted and the VCB operation completed. It is not the first VCB backup to be taken since the upgrade, and the others completed with no alarms being triggered.
So, apparently a false positive, but is there anything that can be done to prevent this? Is this a timeout issue, and potentially caused by vCenter being in the managed cluster?
Any insight appreciated.
Target: esx1
Previous Status: Green
New Status: Red
Alarm Definition:
Current values for metric/state:
State = Not responding
Description:
Alarm 'Host connection state' on esx1 changed from Green to Red
Hi,
before searching for an answer, you should upgrade to esx 3.5u5 or 4.0. Imagine how many bugs fixed since your esx 3.5u3! Maybe an restart of the VMware managment Service on the ESX can fix this temporary.
Yep, already on the schedule, but a couple weeks out yet. Figured it couldn't hurt to chuck it out there and see if anyone had any ideas.
Thanks!
Just FYI, I'm actually running the most current version of vCenter 4 and I'm having this issue every time a snapshot is taken of my vCenter VM.
We have seen this after our upgrade as well. Actually, not technically an upgrade, we did fresh installations.
vCenter - 4.0 build 208111
Hosts - 4.0 build 244038 (all current patches apr/10)
When the vCenter VCB is finishing up, most hosts loose their connections for a very short period of time. Within the alarm, there is no way to set it to only alert after 30 seconds. Is there way to reduce the sensitivity of the host connection state? With all the false alarms, it makes it difficult to know when there is a real issue.
This is strange, we backup our vcenter server using veeam since years now but just since yesterday it starts throwing host connection state alarms when doing a snapshot of the vcenter server machine (even when doing no vss processing inside the guest at all, no vmware, no veeam vss trigger). Any ideas?
best regards,
Joerg