Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have metaluns stripped across 6 raid groups of 4+1 (2 DAE's) and I would like to add another DAE. I will create another 3 raid groups on this new DAE and would like the metaluns to be stripped across all 9 raid groups without having to migrate any on the metaluns or create concatenated metas. Is this possible?

share|improve this question

Certainly. Create identically sized luns on the new sets, and simply add those to the metalun. The CX will rebuild the metalun creating new stripes accross all the member luns.

share|improve this answer
This sounds like a migration task.? This will have performance impact as the "live" luns will be read and written to the same spindles again to be re-stripped.? Also, if I add new LUNs to the meta, am I not increasing the size? – Alan Nov 20 '09 at 10:52
just spoke to EMC and they "suggest" that I expand the raid groups and that will automatically stripe to the new disks. The only problem is that these disks are SATA and I don't want to do more that 5 disks in RAID5 because of possible rebuild time issues... – Alan Nov 20 '09 at 12:55
Yes, you are increasing the size of the metalun. Is this not what you're attempting to do? It won't be possible to redistribute the data without a loss of performance during the rebuild. I think, but haven't tested this, that the loss of performance will be minor as parity information does not need to be recalculated. – Roy Nov 20 '09 at 16:40
The rebuild time is dependant on the size of the s-ata spindles, rather than the number of spindles in the set. The increased risk of going from 5 to the reccomended 7-8 spindles per raid 5 set is marginal. If this worries you, you may want to consider migrating to raid 6. – Roy Nov 20 '09 at 16:42

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.