Ghost Drive Letters on 6 node Exchagne cluster
From: Jason (jdheltne_at_sbcglobal.net)
Date: 04/17/04
- Next message: Steven W: "Losing Shared on Fail Over"
- Previous message: janderson_at_optonline.net: "My little sister giving a handjob"
- Next in thread: Rodney R. Fournier [MVP]: "Re: Ghost Drive Letters on 6 node Exchagne cluster"
- Reply: Rodney R. Fournier [MVP]: "Re: Ghost Drive Letters on 6 node Exchagne cluster"
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 17 Apr 2004 14:16:13 -0500
We have setup a 6 node Exchagne 2003 cluster for our users, and we have done
months of time in our test labs doing testing, and now that we are doing our
production build, we have ran into a quirk. What we had prevuiously done is
test both QLogic 2300 HBAs with an IBM FasT SAN, and Emulex 9802 HBAs with a
EMC Symmetric SAN. In both rounds of testing, we saw exactly what we should
regarding failover, when a disk resource is failed from node 1 to node 2, in
WIndows explorer you only see the disks that the node currently owned. Now,
we have had to replace our SAN with an EMC CLARiion CX600 SAN, and we now
want to use VSS for backups via the EMC SIME product. For that reason we
ahve had to now switch to use the Emulex 9002 HBAs because according to EMC,
this is the only HBA that would support VSS. Anyway, since we have switched
to this HBA, we are now seeing the ghost drive letter issues where when we
fail a disk resource from node 1 to node 2, the disk still shows up in
Windows Explorer, just with the name "Local Disk" and then a drive letter.
The disk is not accessable from Windows Explorer. I have tried logging off
and back on both via terminal services and at the console, tried rescanning
the bus, etc... the only thing that makes the drive letter go away is a
reboot of the node. I can repeat this at will. We built everything exactly
like I should, had all nodes but node 1 off did all config from node 1,
etcx... then turned node 2 on, added it to the cluster, then node 3, etc....
Anyway, has anyone seen this issue and if so, is it causing you any issues?
I am leaning towards it being an HBA/driver/PNP subsystem issue, but it
concerns me because we did not see this when we did our previous testing.
thanks,
Jason
- Next message: Steven W: "Losing Shared on Fail Over"
- Previous message: janderson_at_optonline.net: "My little sister giving a handjob"
- Next in thread: Rodney R. Fournier [MVP]: "Re: Ghost Drive Letters on 6 node Exchagne cluster"
- Reply: Rodney R. Fournier [MVP]: "Re: Ghost Drive Letters on 6 node Exchagne cluster"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|