We recently upgraded to 7.4 and we have about a dozen pools.
We started seeing that several VM's are in error state. The pools are set to refresh on log out and some of these pools will no longer refresh.
After the user logs out we get this event:
Refit operation refresh failed
And the VM has an error state of:
Invalid powerstate: The attempted operation cannot be performed in the current state (Powered off).
If you power up the VM via vCenter it powers up just fine and seems to work complete the refresh without error.
Digging through the composer logs does not give any additional error beyond the invalid powerstate.
The only KB I could find was not relevant as i checked the DB and did not have multiple entries: VMware Knowledge Base
Even if I do a full recompose of the pool (using the existing snapshot), the recompose works fine but refreshes still after log off.
We upgraded from 7.1.0 to 7.4.0 last week and I've seen this on one VM that was manually refresh. I looked at our pools that are set to refresh on log off and we have seen this a few times. Did you open a ticket with support? eccl1213
We are about to upgrade to 7.4 from 7.1 in production but I have not seen this on the development environment I have...interesting. What versions are your vcenter and esxi?
Our vCenter is 6.51d and the Host are 6.0p6.
Its very hit or miss but we never saw this before the upgrade to 7.4
We have vCenter 6.0 Update 3d and ESXi 6.0 Patch 4. We are working on upgrading to ESXi 6.0 Patch 6.
Here is a SQL query that will find all events with the error. I found that we were also seeing this prior to our upgrade to 7.4.0 when we were still on 7.1.0. I've opened an SR with VMware.
SELECT [EventType]
,[ModuleAndEventText]
,[Time]
,[Source]
,[Severity]
,[Node]
,[DesktopId]
FROM [ViewEventsCORP].[dbo].[VE_event]
where [ModuleAndEventText] like '%refit%refresh%'
order by Time DESC
I opened an SR and heard "refit seems to be known issue impacting Horizon 7.4". I'm still troubleshooting with VMware to confirm.
Keep us posted - wondering if i should hold off on the upgrade now.
Our first log bundle was generated a few days after the issue occurred and no longer had the relevant entries. I just had another one today and I've generated a log bundle for VMware which will hopefully have the log entries they are looking for.
VMware confirmed that we are experiencing the known bug in 7.4 and it has been fixed in a future version. They are offering a hotfix for affected customers.
man, that is a bummer. What is the hotfix solution?
I'm waiting on confirmation of that.
Any confirmation of the hotfix? Is it a patch or config change etc?
I wanted to save the trouble of opening a ticket...going through 10 iterations of log bundles and two tiers of support before support points me to the hotfix.
Support said this will be fixed in 7.4.1 is Q2
I was offered a sandbox build but declined it. We are waiting on the next release where we are told it's fixed.