3.1.0.125: CLOCK SKEW Fixed over 35 minutes - still "Active"

PDinCA

w32tm /config /update was issued on the problem node and on the Base Monitor server but the HIGH Alert is still "Active".

Why is this, please?

PDinCA

Status: Ended Time raised: 5 Jun 2012 1:36 PM Time ended: 5 Jun 2012 9:41 PM

It appears to have taken about 2 hrs 20 minutes after the time was sync'd on the base and node2 for SQL Monitor to realize there was no longer an issue... Is this a bit excessive a lag?

Strangely, we didn't reset the active node until today but on the same day (Jun 5), for 15 minutes there was an apparent skew between the base and node1 of 17 seconds that lasted for only 15 minutes. Curious :?

PDinCA

Per a very helpful response by Brian Donahue:

Sometimes clock skew can be prolonged by an overworked virtual machine. I did find this on Stack, maybe it explains your situation with SQL Monitor. http://serverfault.com/questions/28803/ ... ck-in-a-vm

Thanks,

Brian Donahue
Red Gate Product Support

My situation is that the Base Monitor is on a VM and monitors a physical active-passive cluster.

There is reference within the post cited above to a VMWare "tinker" setting that I don't profess to understand but looks like it should always be in place for my kind of architecture.

Hope this helps someone else...