MSA1000 Ouch

[AFyodorov]

Just had a problem with one of our Exchange 2003 clusters.

Two DL380 nodes attached to MSA1000. Firmware 4.32

300GB 10K RPM disk in Bay 14, used as hotspare for a few RAID volumes.

We decided to replace this disk with a 300GB 15K RPM disk.

- I go to ACU, disable the hotspare.
- Local IT guys pull out the disk
- so far so good
- Local IT guys insert a new 300GB 15K RPM disk (HPB9 disk firmware)
- cluster fails
- rebooted both cluster nodes. Cluster can't start. No volumes from MSA1000 are visible on either node.
- Shut down both nodes, shut down MSA1000, turn MSA1000 back on, wait for MSA1000 to boot, turn on one of the cluster nodes.
- Everything is fine, except the new 300GB 15K RPM disk is not showing up in the ACU, local IT guys are telling me it is flashing with a red X.

We ended up putting the old 300GB 10K RPM disk back in and calling it the day.

I have never seen an MSA1000 dropping off like this just because a new disk has been inserted. How could this happen?

[Patrick Terlisten]

Hello,

the MSA1000 is a scsi based storage (internal bus for disks and shelfs is scsi, host connect is FC). So if you plug in a bad scsi drive, the scsi bus (on of the two busses in the controller shelf) could fail. The 14 bays in a MSA1000 are split into two busses. The flashing X is an indicator for a bad drive.

Log a call at the HP support center. Otherwise connect with the serial controller cable to one of the controllerports on the front-side and run the command "show tech_support". You can post it here in the forum.

Best regards,
Patrick

[Andy_Stor]

Hi,

I totally agree with Patrick, if you have a bad SCSI drive then the whole bus might freeze. Few things I wanted to share/ask,
- MSA1000 F/W is low, 4.48 has got lots of fixes which were in 4.32, so why the F/W is low. Is there any specific need running MSA on low F/W.
- The new disk could also be a DOA.
- If you have any other spare disk you can try using that disk

Andy

[AFyodorov]

Thanks.

We haven't had a chance to upgrade firmware on this busy server. In our environment, when something is running fine, we try to have as little downtime as possible. Sometimes it takes months to get business approval to even deploy patches and reboot.

We have upgraded firmware on other MSA1000 as needed, when we had a window of oppotunity, just not this one.

We put this disk into an internal slot on another server and it looks good.

[Rob Leadbeater]

Hi,

The release notes for Firmware 4.48 here:

ftp://ftp.compaq.com/pub/products/storageworks/msa1000/MSA1000_448_Release_Notes.pdf

Have got the following entry:

"Fixes an issue where the MSA1000 array controller could get a Critical Lockup F112h when inserting new disk drives or migrating disk drive from non smart array based SCSI controllers"

That might be the reason...

Cheers,

Rob

[AFyodorov]

Awesome, thanks! I should read release notes better :)

[Rob Leadbeater · Accepted solution]

Cool. Remember to assign points ;-)

Cheers,

Rob