Show all of the cluster disk's collected statistics regardless of active node
Currently, a cluster disk's collected statistics are only visible for the duration of being "active/owned" by a cluster node. If I ran on Node A for the last month and fail-over to the B node today then I lose all visibility to the prior month's stats since they were collected while on the node A.
Joey O'Neal commented
It's difficult to see what went wrong on a cluster when you only see connection or disk drive info for the currently active node in a SQL Cluster. When a failover happens, we lose ability to see the passive node (or on an active/active cluster whatever node is deemed to be passive) and any data for network or SAN Disks is not viewable. This data would be helpful in determining what was going on just prior to the failover.
Larry Silverman commented
I would like the same feature added/fixed. We had to fail over our cluster today, and I really wanted to compare Disk Reads/Sec and Disk Writes/Sec before and after. But after the failover, the passive server's drive letters for data, log and temp (M:, L: and T:) were not accessible, and the active node's statistics for the drive letters were empty prior to the event.
For clustered resources like disk, the statistics should probably be associated with the cluster itself, and not the individual servers.
Hello .. first .. my sincere apologies. Not sure how we missed this post. You should be able to browse history. SQL Monitor doesn't stop browsing history when Suspended. If you think that there is bug then please raise it with Support and we will investigate it.
I had a server that I knew was having problems and was getting a high volume of alerts (and thus emails) from SQL Monitor. I suspended monitoring on its instance so that the alerts would stop being generated. I discovered when I went to browse history for the instance that no data would show up! Once I resumed monitoring I could browse the history.
This to me seems like a bug, but I'll add it as a feature request in case it's intentional I am curious what the reasoning behind this would be however.
If this is inappropriate for a suggestion feel free to remove it and I'll file it with support.
Priya Sinha commented
It would be great if you could email the screen shot to email@example.com.
John B commented
Specifically what I'm seeing is that when monitoring a cluster, the "Disk used" statistics for cluster disks, say F: are only visible for the length of time the disks were owned by a specific node.
So, if node1 owns all the cluster resources on Monday and there are 2880 samples, 1 every 30 seconds for the 24 hours of Monday, then the cluster fails to another node, at 12:00:01am Tuesday those 2880 samples take on Monday from node 1 are no longer visible since you have to click on a cluster node to enumerate the Disk Used stats. The issue appears to be due to the filtering mechanism of drilling into a specific node.
I have a screen shot of the issue if someone wants it.
Priya Sinha commented
Please help me to understand this issue. The current behaviour for SQL Monitor is that stats for quorum disk is available under "Cluster Overview". The stats for shared disk + quorum disk is also available under current active node. But SQL Monitor always monitors all the nodes of the cluster. So even if the node is not currently active, it is still being monitored. So after failover, you should still be able to see the information.
Please explain bit more in detail.
Thanks a lot,