Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

Potential Data Loss at MSA20 !!!

Dmitri Smirnov
Occasional Contributor

Potential Data Loss at MSA20 !!!


MSA20 attached to DL380 with RedHat Linux with 6402 with latest firmware and utilities.

We had a drive failure reported by MSA20 and confirmed by ACU and Diagnostic utilities.

After replacement of failed drive the RAID-5 rebuild process started for a biggest logical drive (2TB) but stopped at 5% and drive went to "Ready for Rebuild" status.

After HP support attached a notebook to a serial port of MSA20 the array generated a message during the rebuild process about unrecoverable read errors on one of the drives (different from replaced).

No errors about this drive reported in ACU, SmartStart, Linux or array diagnostic utilities (before, during or after the problem). Second failed disk has S.M.A.R.T enabled and reports no errors. Media scan were activated every 15 secs but reported no errors.

The only solution at the moment is to migrate all the data (and certain data proved to be corrupted), replace the second drive and rebuild the entire array.

The ONLY way to find the problem is to connect to serial port of MSA20 and monitor array messages.

I still have no solid response from HP (after 7 days) why it is happened and how to prevent it in future.

My advice: if you have MSA20 - make sure to have good and often backups. Make sure you realize how long is to restore data from the backup in case of the failure (2.5TB in my case, still in recovery mode).
If you have MSA20 - I would recommend to address this problem with HP - myself I will not use MSA20(s) until I'll have a clarification how to avoid this problem in future.

2 REPLIES
generic_1
Respected Contributor

Re: Potential Data Loss at MSA20 !!!

Sorry to hear about your problem. Data loss sucks.

First off HP has escalation managers. I would contact HP and ask for one, and talk to your local HP rep too.
Make sure your firmware and drivers are up to date on your MSA.
If your device supports more than one hotspare. You could consider that as an option. Also if you have enough room a mirror set of your data would be nice too.

If you have EMC Symetrix, DMX, or HP XP on site you might check into getting storage off there instead. I have found these to have the best Data integrity out of the san products I work with.
Uwe Zessin
Honored Contributor

Re: Potential Data Loss at MSA20 !!!

Jeff,
the MSA20 is a low-cost SATA solution - you cannot seriously compare it with a high-end disk array!
.