December 9, 2005 at 5:30 am
Hi guys,
Would appreciate any help on the following error received when installing SQL 2K Enterprise on a configured W2K cluster box.
An error occurred while creating virtual server resources for clustering.Setup could not connect to the database to update cluster-specific information in the system tables.
The various SQL services are set up on the node but they are not starting.
My apologies for the rather vague information here, but if anyone has had any experience of this and knows of a workaround, a swift response would be greatly appreciated.
December 9, 2005 at 11:09 am
Was the SQL Server 2000 EE installed as a clustered instance? If so, did the setup not roll back when it hit this error or are you applying a service pack or the like and this is happening? What errors are you seeing in the Windows 2000 System event log?
K. Brian Kelley
@kbriankelley
December 12, 2005 at 3:04 am
Brian,
Thanks for the reply. To clarify our position, we're installing SQL Server (not a service pack) as a clustered instance. It just so happens that we've managed to resolve the issue. We did have a number of problems with the installation and the System Event log and the logs produced by the installer assisted in helping us resolve the issue.
We experienced an IO problem at one point, which came as the result of some data corruption on the data drives. A full format of the disks resolved this issue.
The error described in my previous post seems to have come from us not putting the domain SQL Service account we use in the local administrators group for the cluster. A minor oversight , but one that had serious reprecussions!
Anyway, all's well in the world now!
December 16, 2005 at 10:20 am
uurrrmmm ... Yeah! Not putting the Cluster and SQL Service accounts in the Local Admins WILL have some minor issues!?
I'd be a little concerned that you have data corruption on the drives already tho?? Did you leave a node on when setting up the disk config?? U have to be very careful to switch off the node you aren't currently working on when doing the disk stuff ... otherwise, oh yeah, u get a load of corruption!?
Rules of thumb when building a cluster ... dont rush, and follow EVERY step maticulously!?
December 18, 2005 at 5:03 am
"... dont rush, and follow EVERY step maticulously"
I certainly agree with you there Vinny! And I'll remember that the next time I'm involved in this process. Unfortunately I wasn't involved in the initial build for the cluster, and despite detailing my prequisites to the team responsible for building the server, it seems there were a couple of oversights (namely not putting the service account in the local admin group). But as far as the data corruption goes on the drive, I can't answer that one since I wasn't involved. But it is a bit suss that this is the second time this has happened on one of our clusters!?
What's more, it's also the second time that I've had to complete something of a convoluted SQL Enterprise installation. For instance, we've never been able to get the installation to complete successfully on the second node when we're doing the primary install. I've always had to resort to doing an installation on one node, and then a second installation on the other node, getting it to affiliate to the virtual server. Mind you, things are never straightforward where I work!
Anyway, I've learnt my lesson ... I will be checking maticulously from now on that the system meets my prequisites, and I'll try and get involved from an earlier stage so we can try and stamp out this data corruption thing.
December 19, 2005 at 2:53 am
Only ever had the "failed on node 2" scenario once ... and i've done quite a few clusters over the years. The site in question did (does!) have some network issues, which is what i decided to attribute that to! Was a pain in the backside ... but when it comes down to it, computers are there to do what I tell them, so it wasn't a long battle!?
Quick check u might wanna try to see if u have similar network issues ... try dragging a number of files from a network share to your machine (make then a decent size as well!) ... what we get on this network is that (randomly!), the transfer fails with a message like "Network name is no longer available"! Tis down to a badly configured Switch as far as i'm aware!
anyway ... enough of my problems ....!!
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply