MSA 2050 ... another question about Raid-5 vs Raid-6

[WillyEB70]

Dear all,

I would like to have your opinion on the configuration of the MSA2050. I have 12 HDD SAS 2TB  (archive).

Which of these three solutions is the correct one ?

1. no. 11  HDD RAID-5 + no. 1 hot-spare
2. no. 11  HDD RAID-6 + no. 1 hot-spare
3. no. 12  HDD RAID-6 with no hot spares

there are no particular performance requirements. I have to choose the safest thing without wasting too much disk space.

In the case of the first solution, when a HDD fails, does the hot spare come into operation automatically?

MSA 2050 best practices say that the number of hdds for raid5 must be odd while for raid6 even.

If I did no. 10 HDD in raid6 I could put two hot-spare but I would lose a lot of space.

Some idea ?

Thaks a lot

Bet Regards

Willy

[ArunKKR · Accepted solution]

Hi Willy,

In solutions 1 and 2 mentioned the spare will kick in as soon as one of the drives fails in the disk group.
RAID 6 disk group with 11 disks and one global spare would be the safest option(Option 2).
RAID 5 can survive a single disk failure in disk group whereas RAID 6  can survive up to 2 disks failure in disk group.

Please note that none of the configurations mentioned above meet best practices guidelines for performance.

For optimal write sequential performance, parity-based disk groups (RAID 5 and RAID 6) should be created with “the power of 2” method. This
method means that the number of data drives (non-parity) contained in a disk group should be a power of 2.
We have noticed significant sequential write performance improvements when disk groups are configured as per power of 2 recommendations.

Please refer to pages 18 and 19 of best practices guide.

https://www.hpe.com/psnow/doc/a00015961enw

[JonPaul]

@WillyEB70 @ArunKKR 
HPE does not recommend RAID 5 for Archive class drives (See MSA Best Practices)
The only class of drive that would be acceptable to use RAID 5 is the Performance class drives, SSDs.  This is due to the low failure rate and the quick rebuild times.
Beyond that the decision is between capacity and ability to react to a drive fault.  If you have spare drives onhand and can access the system within a short timeframe then using all the drives could be more acceptable. If the system may have to run with a degraded drive for a longer timeframe it would be better to have 'live' spares in the system.

[WillyEB70]

@JonPaul  @ArunKKR  No .12  HDD in raid-6 without hot-spare a good solution ?

[JonPaul]

@WillyEB70 
RAID 6 will allow 2 drive faults without data loss.  In the event of a drive fault the system will start notifications to the owner that the system is degraded.  IF the system is setup with notifications then the owner will be notified and can remediate.  During this time the system is able to sustain an additional drive fault.  Following all the Best Practices of RAID level and notifications AND being able to remediate issues quickly makes a RAID 6 with no spares a recommended best practice.  A better Best Practice would be to also have SPAREs for the RAID 6 but it does come at the cost of capacity.
RAID 5 + SPARE is not a best practice as it leaves the system in a CRITICAL state while the re-build is happening.  Rebuilds of very large drives (10TB+) can be in the DAYS range, depending on a lot of factors, during which time the data is at risk of a second drive fault.

[WillyEB70]

ok thanks