April 11, 2017 at 6:37 am
Hi Guys,
Just experienced an unplanned failover. Before I go on I should add this box has been neglected in terms of memory throttling so its about 94-96% utilised out of 64GB.
Throttling is planned for this evening during the Maintenance window.
However we just had a Lease Timeout cause a failover.
Could this be caused my a low memory condition for Windows?
This is the juicy parts from the ERRORLOG:
The local replica of availability group 'grpName' is going offline because either the lease expired or lease renewal failed. This is an informational message only. No user action is required
Windows Server Failover Cluster did not receive a process event signal from SQL Server hosting availability group 'grpName' within the lease timeout period..
Timeout waiting for external dump process 11912.
So the above looks unpleasant -- could it be caused a shortage of memory available to Windows. Prior to this performance was fine for the SQL Server, but RDP to the desktop could be a bit laggy..
Cheers
Alex
April 12, 2017 at 5:06 am
How was the CPU at the same time? I've heard of maxed out CPU causing lease timeouts. Is there anything helpful in the cluster log? Also, was there anything useful in the error log right before the bit you've quoted?
Questions, questions :).
April 12, 2017 at 5:18 am
Howdy,
Well, CPU was pretty erratic because of a MAXDOP oversight... (and by oversight, I mean someone changed the Cost Threshold to 1 rather than the MAXDOP with a miss-click).
So CPU was high but at that point was not pegged. So between low-mid 90's. It normally cruises at about 24%
So to answer yes CPU was considerably higher but not flat out.
As I can't get any more useful info (there was nothing more that appeared relevant in the ERRORLOG) I am chalking it up to being tight on resources.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply