MSA 1000 failed HDD replaced, rebuild not starting

[Joe Conner]

They are all "73GB" disks, the new one is a match to the existing disks in the array.

How do I assign the replacement disk as a global spare", I don't see that option anywhere (in ACU anyway). That's pretty much my problem, I can't seem to change anything to get the rebuild to begin.

I'll be looking at the Cli options today, and at least one full backup of the array should be done by then, at least throwing a safety net out there in case it gets worse not better.

[Johnny von Heimerall]

Hi Joe,

Can you still access the LUN?

Best regards,
Johnny

[Joe Conner]

Well, no Cli utility on the server. I didn't find any installed that I could use (it's not a Compaq server). I do have ACU, and I installed SANsurfer, but that just seems to deal with the host adapters and not the arrays attached.

I'm not sure what you mean if I can still access the LUN - the drive is still mounted and working on the server (Windows 2003), it's obviously running off of the RAID with the failed drive. ACU sees the array just fine, and complains about the failed drive. "#274 The current array controller has a bad or missing physical drive attached to Port 2: SCSI ID 12. To correct the problem, check the data and power connections to the physical drive." and "#272 The current array controller has a bad or missing drive. Logical drive 1 (RAID 5 in array A) is operating with reduced performance and a further physical drive failure may result in data loss" ... "... Configuration changes to this logical drive or any other logical drive in array A are not allowed until the problem is corrected. To correct the problem, check the data and power connections to the physical drives or replace the failed drive. For more information, run the Array Diagnostics Utility".

I don't have the "Array Diagnostics Utility".

I've replaced the drive - tried two different NEW drives, one at a time. When the old drive is in, it lights red as failed. New drives go in and panel on MSA1000 indicated something akin to "drive hot added bay 12". The green hard disk light on the drive does not stay lit. in ACU, there then is a drive listed under MSA1000 Controller as "72.8 GB Parallel SCSI Unassigned Drive at Box 1:Bay12". I just can't seem to assign it to anything, to get the rebuild to begin. Under that, "Parallel SCSI Array A", there is an item "??? Parallel SCSI Drive at Port 2:SCSI ID 12", its icon is a red "X" over the hard drive icon. The other "normal" drives show as "Bay 1" thru "Bay 13", skipping "Bay 12".

So the MSA1000 knows drive 12 failed, and sees the replacement drive arrive (when I removed the failed drive, the panel indicated that also). It just won't start the rebuild, which if memory serves me, should start automatically. Rebuild priority is set to medium btw.

How would you go about assigning the drive as a spare? I suspect I cannot, since no changes are allowed since the array is in "interim recovery mode", as indicated on the MSA panel as part of the startup sequence. Am I missing a key step in the process to cause the rebuild to begin? Right-clicking anything in ACU only brings up "More Information".

BTW, right-clicking the failed drive under the array shows its status as failed. Right-clicking the new drive at the top of the MSA1000 tree shows its status as "OK".

Show physical view and show logical view doesn't seem to add any options, nor have I found anything in the "Configuration Wizards" that doesn't talk about deleting data.

Once backups are done, I'm going to try updating the firmware to 5.20 (it's 4.24 right now), and if that doesn't work, try deleting and recreating the array - might lose the data, might not.

I'm still stumped why the rebuild didn't start automatically, or I can't seem to start it myself. What good is RAID if you can't recover from it without losing all the data, assuming I'm going to if I delete and recreate the array?

[Joe Conner]

Firmware flash to 5.30 didn't help.

I just deleted the array, and recreated it from scratch. No more playing around with it.

Thanks everyone for your advice!

[Joe Conner]

Closing thread. There was no resolution on how to get the array to rebuild.

It should have somehow, and not required a do-over to get it back to proper operation. But at least RAID functioned well enough I didn't lose the data.