So over the last 5 or 6 days we have been dealing with a mess of a SQL cluster here at the office. After speaking at length with MS on the issue we needed to perform the following actions.
Move the entire SQL Cluster group along with the MSDTC resource to the cluster group.
Fix the dependencies on the MSDTC resource.
Add a new clustered instance of SQL that can fail-over between all 3 nodes.
According to MS having the MSDTC resource depending on the SQL drive is NOT a supported configuration. MS SQL Team and the MS Windows Cluster Team both had to be involved in case things went bad during the move. MS wouldn't allow me to perform these cluster changes myself due to some possibility of the whole thing going south.
So we moved the MSDTC and SQL group. Then deleted the MSDTC and created a new one. Then we deleted the folder on the SQL data drive that had been there. We made the MSDTC dependent on the Quorum drive. So at this point MSDTC was all set. From there we moved all the SQL stuff to the right place in the SQL cluster first instance.
Then we installed the second instance and then upgraded it to fail over between all 3 nodes of the cluster.
Now that all the cluster stuff was fixed we had to move TEMPDB and the SQL DB and Tlogs and Backups. From there we were good. Since we were on an EMC array PowerPath FULL was then installed on all 3 nodes to ensure that connectivity via fibre-channel was as reliable as it could be.
Since it was 5am at this point and I had been working all night I calmed down and wrapped everything up. Then as quickly as it started it was all over and I could get some sleep :). LOL
No comments:
Post a Comment