I used the VIOPS document to help setup a vCenter 4.0 with Windows 2008 Failover Clustering. I had to adapt the document a bit (given it was written for MSCS 2003) but it is mostly working at this point. There are two servers for the cluster and a third housing the SQL database.
The one thing I notice is that the vCenter client must reconnect after a failover. It doesn't actually disconnect, but after any attempts to perform functions throws the error:
"The request failed because the server 'clustername' closed the connection. (The underlying connection was closed: A connection that was expected to be kept alive was closed by the server.)"
Is this behavior expected, and as such, should we expect to be required to close and reopen the vCenter client? Or should the client maintain the connection regardless of failover? If so, what did I miss in terms of setup?
Also, it seems like the connection may just take some time to reestablish. Could it be I'm not waiting long enough for the client to "resync" after failover?
Thanks,
Jason
Also, it seems like the connection may just take some time to reestablish. Could it be I'm not waiting long enough for the client to "resync" after failover?
Seems that this is the problem.
Old 2.5 client was more smart during re-connection.
The new one seems that in some cases (for example when VC chash and you manually restart the service) is not always able to auto-reconnect.
Andre
Hi,
Since the service gets restarted as part of the failover, the client will fail to reconnect. and so you will be forced to reconnect by entering the credentials after the failover. I don't think this is due to some issues in your setup.
Thanks
Sandeep