Unable to monitor SQL instances after completing a rolling upgrade in an AG
andy8888
Posts: 2 New member
Hi folks,
We're trying to complete some rolling upgrades from SQL2012 to SQL2019, which involves adding 2 new servers (that have a newer OS but SQL2012 installed) into the existing cluster, then we upgrade SQL on the new servers after they're sync'd in the AG and failover to them- at this point the 2 old servers drop out of the AG but remain part of the cluster until we decommission them a few days later.
SQL Monitor can't currently monitor the non-prod AG where we have already down this because it says "Access is denied." even though it has the same permissions in SQL and the OS on all 4 servers. SQL Monitor can't see the new servers in the cluster when i try pausing and resuming monitoring- is there anything else I can try without having to remove and add the cluster back in again- will this remove the historical data we have about the monitoring for them?
We're trying to complete some rolling upgrades from SQL2012 to SQL2019, which involves adding 2 new servers (that have a newer OS but SQL2012 installed) into the existing cluster, then we upgrade SQL on the new servers after they're sync'd in the AG and failover to them- at this point the 2 old servers drop out of the AG but remain part of the cluster until we decommission them a few days later.
SQL Monitor can't currently monitor the non-prod AG where we have already down this because it says "Access is denied." even though it has the same permissions in SQL and the OS on all 4 servers. SQL Monitor can't see the new servers in the cluster when i try pausing and resuming monitoring- is there anything else I can try without having to remove and add the cluster back in again- will this remove the historical data we have about the monitoring for them?
Tagged:
Answers
What is it showing "Access is denied" for? Is this at the cluster level for General or WMI?
I would expect SQL Monitor to query the WMI class MSCluster to the Windows cluster name level (which ends up querying the active node) to determine if it has more nodes since I think the discovery process is performed on occasion. At that point it would see there were more nodes, then discover the instances on those nodes.
If you go to the Configuration > Monitored servers page, do you see the cluster level with only the original WSFC nodes and the instances on those nodes underneath that? OR do you see all four WSFC nodes, but only the SQL instances for the original two?
If you remove and re-add the cluster, as long as you DO NOT tick the box to delete historical data, it will keep it and that process will definitely trigger the discovery of the new nodes (but if there is still any permissions error, it will still run into that then).
If you can send in a screenshot of what is discovered on the Configuration > Monitored servers page along with what is shown on the "Show log" link (if it is there where you are seeing the "Access is denied") that will help determine what may be going on. If needed I can reach out via a support ticket to get that information rather than having it on the forum here.
Kind regards,
Alex
Have you visited our Help Center?
Thanks for getting back to me. Short answer is that I got it sorted but not in the way that I thought I would. The error was occurring with the cluster and related to WMI. We were seeing all 4 nodes but only the SQL instances for the original 2. I had checked we had the same OS permissions as other servers which were being monitored fine but I thought it was worth giving the account local admin rights on the new servers; this did the trick. SQL Monitor then connected fine to the cluster, saw the old nodes weren't there anymore and removed them and started monitoring the new servers.
I can only think there was something a little screwy with our GPOs on the new servers that were impacting the access.
Excellent, thanks for letting me know it's up and running!
Kind regards,
Alex
Have you visited our Help Center?