UNSOLVED
Powervault 220S with Perc 4/DC cluster problems
Hello,
I have had a great deal of grief trying to configure a NetWare 6.5 cluster with two PowerEdge 1750's abd a Powervault 220S. Both 1750's have an added Perc 4/DC for connecting to the Powervault. The cards are configured for clustering and I did set the SCSI ID on one of them to 6. The Powervault is running in cluster mode.
My present issue is that I have had to replace an apparently failed Perc 4/DC and now I cannot bring both servers up at the same time. No matter what I do-- clear the config on both cards, bring one server up, read config from the disks, reboot, and bring that server all the way up, clustered volumes and all-- if I then try to bring the 2nd server up, it will hang while scanning SCSI devices and will cause the 1st server to abend. Also, the first drive LED in the Powervault starts flashing, and the RAID config will show that drive as failed. Rebooting the servers and/or the powervault fixes this.
Every time I go through this, the 1st server's Perc 4/DC loses its NVRAM config and I have to do it all over again. So I've had the servers alternating back and forth since Friday, crashing each other as I try to bring the 2nd one up.
I know that I have at least a fair grasp of how to set this up because I had it running for weeks with no real problems. Since having replaced one of the PERC cards that apparently died last week, it's been a mess.
Also, the Perc 4/DC cards seem prone to lose their configs if the server is rebooted irregularly, i.e. server abends or hangs, or system suddenly loses power. This has happened to me three times on two different servers now, and when it happens on the cluster, I have to bring the remaining server down as well because I can never bring the downed one back up to reconfigure its NVRAM so long as the 2nd one is still actively accessing the Powervault. It's a mess.
Anyone have any ideas? I would appreciate your help immensely!!
Message Edited by Nuncio on 03-01-2005 04:00 PM
- I believe that my problem was ultimately that I failed to clear the config on the replacement adapter before I brought up the server.What I had to do was this:First, MAKE SURE:-that the powervault is set to cluster bus mode-that you are using the same channel on each adapter (if it is a dual channel) to connect to the powervault.-that SCSI initiator IDs for one card is 6 and the other is 7-- they must not be the same. These must be correctly set before you attempt to configure the logical drive.-that cluster mode is enabled on both adapters.Power everything down. Bring up server A; go into the adapter BIOS and clear the adapter configuration. Save and exit. Power down server A and do the same thing with server B.Next, bring up the powervault. Then, bring up server A again and enter adapter BIOS. Next, bring up server B; enter the system BIOS and just let it sit there. On server A; reconfigure the logical drive as desired (tag the drives, end the array, select the span, and accept. DON'T INITIALIZE if you have valued data on the array!) Save and exit.Now, boot server A into the system bios and let it sit. Reboot server B, and go into the adapter BIOS. Select view/add configuration; it should give you the option to read from the disk, but if not, go ahead and follow the same procedure for reconfiguring the adapter as you did on server A. Save and exit. Bring both servers down. Bring up either node, let it boot all the way into the OS (or if you don't have the OS loaded, just let it boot past the point where it successfully scans and spins up the disks on the array.) Then, bring up the other node. You should find that both come up okay.
hildich
1 Message
212
0
Posted April 27th, 2005 08:00