SQL Server Performance Forum – Threads Archive
Move Group No Longer WorksWhen I open Cluster Admin. and attempt to move the Cluster Group, it goes offline like usual but comes back online on the same NODE, which is NODE 1. This has been a stable MS Cluster between two Dell 6650s since Nov. last yr. Not sure what the problem is. THis is a client of mine. They informed me that maybe a week or two ago they had some problems and with Dell’s support they had to re-set the configuration on one of the Nodes PERC4 cards, I believe they told me it was NODE 2. I havent had a chance to get on-site yet, but looking through the cluster config. I noticed that the Cluster Group has no preffered owners, whereas the SQL Group has both Nodes listed, Node 1 being first. I cannot test to see if this missing of preffered owner means anything because its a production server and I do not want any extra headaches. I did some reading and I dont believe this missing of preffered owner would give me this result. So I was wondering if anyone could point me in the right direction. I will be on-site tomorrow, taking a look at NODE 2’s PERC4 card’s BIOS settings. Any other helpful hints would be greatly appreciated.
I have not seen this issue before. But my best guess is that node 2 is no longer a functioning node of the cluster, and that is why the cluster group cannot be moved. Why, I don’t know, as I don’t have enough data. I would investigate if this is the case, and if you determine that node 2 is not functioning, then of course you need to figure out why. It might be something as simple as a misconfiguration on the PERC4 card, that once fixed, will resolve the issue, or in the worst case, it may require you to remove node 2 from the cluster, and then readd it. —————————–
Brad M. McGehee, MVP
Thanks Brad. I will post my findings after I visit them tomorrow
There are couple of things one may try: FYI:—> Cluster Group not having Preferred ownesr (PO) is not a problem. I have two SQL clusters and CLuster Group does not have PO specified. So at least we can ignore this. 1. Check both cluster nodes are up, if they are up then proceed to step #2 2. Create a test cluster resource on working node and try to failover to other node 3. If #2 works then cluster fuctionality is working fine at least from failover perspective. IF #3 is working then check the properties of CLuster Group, especilly the "failback" option Try and let us know
Great idea Vinny, I had not though of it. —————————–
Brad M. McGehee, MVP