MSA 2000 - Remove disk from enclosure

[TTT_70]

Hi, I have a disk that is throwing errors on a daily basis (both "recovered" and "unrecovered" medium errors), but it doesn't fail on its own.

I'd like to change it with a spare one, but support told me that I can't just remove it from the enclosure. According to HP, I have to wait until it fails, so that it is the controller itself removing it from the VDISK it belongs to, and then I can pull it out.

I was asking to myself what's the difference between pulling it out and waiting until it fails...

Why can't I just take it out?

The VDISK is in RAID5 and there is already a hot spare in place, ready to start reconstruction as soon as I would pull out the bad disk...

[Johan Guldmyr]

Hi,

Maybe HP support don't want to replace it yet.

It might be a matter of quantity, as in how many errors are you seeing on this disk on a daily basis.

That's how it was for the EVA-disks anyway, if the disk only sends 1-2 unrecovered read errors per day it may not be enough to warrant a replacement.

Which ASC/ASCQ code are you seeing?
http://t10.org/lists/asc-num.htm

I only see two types of 'MEDIUM ERROR' in there (but maybe it doesn't have all either).
And if you see either of those two it sounds to me like it definitely should be replaced because then it's an internal scan on the disk that produces an error..

[SUBHAJIT KHANBARMAN_1]

Go ahead and replace the disk.

There will be no issue as the Vdisk is in RAID5 and hot spare is avaialble.

[TTT_70]

Thanks for your replies.

In a matter of fact, I was pretty sure I could pull the disk out, but the guy from support scared me with "it's not supported" and blah blah blah.

I mean: if it breaks, it HAS to be similar to pulling it out...

Anyway, the errors that I see are as follows (just some of them):

---
Disk detected error (Channel:0 ID:3 SN:3QQ0PRC400009933KW9Y Encl:0 Slot:3)
Key,Code,Qual=(01h,17h,03h) cdb:Rd 047165e3 0008 Info:047165e3h CmdSpc:0h FRU:0h SnsKeySpc:80000eh
Recovered Error recovered data with negative head offset

Disk detected error (Channel:0 ID:3 SN:3QQ0PRC400009933KW9Y Encl:0 Slot:3)
Key,Code,Qual=(03h,11h,00h) cdb:Rd 0471a4db 0005 Info:0471a4dbh CmdSpc:0h FRU:81h SnsKeySpc:800096h
Medium Error unrecovered read error
---

I will proceed with the pull, and check that it starts reconstructing on the Dedicated Spare. Then I will put in a new disk and that will be the new spare, when reconstruction is finished

Thanks again!

[TTT_70]

For the posterity: everything went just fine.

I've been thinking about the reason why HP told me not to remove the drive brutally.

I came to the conclusion that maybe they feared that sooner or later I'd use that disk again in the same enclosure.

Probably, putting that same disk back in the enclosure would cause the disk's metadata to be read by the controller, messing up the configuration of the now-newly-reconstructed V-disk.

But it's all an hypothesys, as HP did not really tell me anything explaining the statement "don't take it out".

Conclusions:

- Everything worked!
- I'll never use that disk again
- I paid for the new disk, even if the bad one was still in warranty... ahem