Netserver LC2000R MSCS Woes..

[CA964633]

Hoping to get some help here for my cluster setup. I have the following HW & SW components:

- (2) HP Netserver LC2000R Servers
- (2) D8602B Fiber Channel Adapters
- (1) Netserver Rack Storage/12 FC
- (1) D6976A Fiber Channel Switch
- Windows 2000 Adv. Server with SP3

I recently rebuilt this lab configuration and am now having issues with what appears to be contention for the shared disk subsystem. Individually each server/node can access the shared logical drives. However if both nodes are brought up it will bring the cluster down. The console for the first server becomes unresponsive as well like an underlying HW issue is eating up cpu cycles. I am using the latest drivers from the HP web site for my FC adapters. Interestingly they show up as Adaptec (didn't know they were OEMed) so I am not sure if these newer drivers are the issue. Anyone else seen any similar problems? Other troubleshooting ideas are appreciated as well! Thanks!

[CA706919 · Accepted solution]

Hello Rob,

I didn't looked deep into your problem yet but I would like to point you to some URL's that give more information about clustering and Rackstorage 12FC.

LC2000 clustering information: http://h20000.www2.hp.com/bizsupport/TechSupport/SupportTaskIndex.jsp?locale=en_US&taskId=115&prodSeriesId=62572&prodTypeId=15351&supportTaskId=405

Rackstorage 12 FC information:
http://h20000.www2.hp.com/bizsupport/TechSupport/Home.jsp?locale=en_US&prodTypeId=15351&prodSeriesId=62538

See drivers for latest Fiber channel assistant information.

See setup/install/configure >>>> clustering for Clustering on RS12FC.

The latest FC D8602B driver is version: 2.0.25.44 and can be found on: ftp://ftp.hp.com/pub/ccd/soar2sdd/ns1998/ns1998en/drv_fc_win2000_202544_1.exe

I'll hope that gives you some more background to solve your issue.

Regards, Jeroen

[CA942904]

Hello,
you should check ...
... the firmware of your RS12/FC. This should be 5.49.00; versions 5.46.00 and 4.xx.xx are NOT clusteraware. Please check this document if it is not the current version "http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=lpn10782&locale=en_US"
... the version of your FC-Array Assistent. This should be A.00.02-05.
... the FC-HBA-driver correct configured to NL-Port (InitAsNport=0)
Your FC-HBA-driver "Adaptec 1Gb/2Gb ..." is correct (version 2.0.25.44).
Regards,
Tobias Winkelmann

[meloni]

hi
two more thinks:

check lun mapping from FCarray assistant on the RS/12
be sure to set the rear jumper of the RS/12 in the correct position

[CA964633]

Thanks for all the advice/suggestions. I have been trying to use it to figure out my clustering issue but I just found out my Rack Server/12FC has a couple bad fans. I will be trying to get this resolved asap then get back to you on the cluster problem.

Best regards,
Rob

[CA964633]

Update: It looks like my fan problem was an aberration so back to my original problem.

I have uninstalled MS Cluster Services to hopefully simplify my situation. Here is a bit more information about the symptoms I am seeing. If I start from cold starts (powered off to on) with one server I can see everything fine. Sometimes if I power down the 1st server completely then bring up the 2nd server it appears the Rack Storage/12FC will not allow the 2nd server to see 2 of my 3 logical drives. Their drive letters appear but I am unable to view their file systems. Of course this is resolved if I start from scratch again. I believe I am using the latest drivers/firmware/reg hacks:

Adaptec 1Gb/2Gb HBAs.. (2.0.25.44)
FC-HBA-driver registry file for configuring the NL-Port (InitAsNport=0)
FC-Array Assistent (A.00.02-05)
RS/12FC Firmware (5.49.00)

I believe I have misconfigured something. Are there any specfic things I need to setup so the RS/12FC can be handed off between servers??

[CA706919]

Hi,

Could you maybe attach the fa1cl.log file from your FC array assistant application and save as well the configuration within the FC array tool and have that file attached as well.

I suspect a issue with you lunmapping but these files should give some clearity about it.

Best regards, jeroen

[CA964633]

Hi Jeroen - I am not in the office now but I think I can remember how I have the LUN mappings setup. Take a look:

C0:0 C0:1 C1:0 C1:1
LogDisk1 0 None None None
LogDisk2 1 None None None
LogDisk3 None None None 0

I will grab those files asap.

[CA706919]

Hi,
that doesn't look very good....

It should be like:
0 x x 0
1 x x 1
2 x x 2

But the files will say more!
Regards, jeroen

[CA964633]

Hmmm the plot thickens.. I did that LUN configuration and I ended up with 2 of every logical drive under Windows. That must mean I have another misconfiguration somewhere else?? I will be able to send those files tonight so thanks again for the help!

[CA942904]

Hi Rob,

what cabling do you have? You can find them in this document http://h200002.www2.hp.com/bc/docs/support/SupportManual/lpn13402/lpn13402.pdf.
Because the correct LUN-Mapping depends on your cabling.

Regards,
Tobias Winkelmann

[CA964633]

Final reply here.. I ended up calling HP support on this a week ago and we removed the FC hub from the cabling configuration. Basically each server/cluster node has a single FC connection to a RS/12FC controller. This seems to have greatly improved the situation. This workaround is odd though since the official HP docs show the FC hub working in the configuration I was using. No matter.. Thanks for all the replies/help!

Rob

[CA706919]

Hi,

Normally you should be able to get it to work with the hub it's even a supported way!

I attached as reference the clusterguide with all the connections if you have 1 of 2 hubs (like SPOFless support).

Goodluck maybe it clears it up for the full 100%.
Best regards, jeroen