How long to add 2x300GB SAS disks to RAID5 MSA2000 array?


I need to add a couple of disks to a RAID5 vdisk on an MSA2312i. I've read that it can take a long time. The new disks are actually in a RAID1 vdisk at the moment which I'll drop and add to the RAID5 instead (currently has 4 x 300GB disks). What I'm trying to work out is:

1. How long??? Some estimates sound insane - 4 days or more??
2. Is the array accessible during this time?
3. If so, how badly is performance degraded


Who is Participating?
HossyConnect With a Mentor Commented:
Isn't it always a 7-day week? :-)

Yes, you can change priority on the fly.  In fact, you could manually set to "low" during higher I/O periods and then to "high" during lower I/O periods.

If you are doing an expansion of the array, your redundancy is maintained per your RAID level.  You are doing a RAID 5, so you can still lose 1 disk during the expansion.  Note this includes the new disks.  Once the expansion is started, don't think of the new disks as separate.  They immediately become part of the array and its redundancy, but you don't gain the space until the expansion is complete.
jakethecatukConnect With a Mentor Commented:
a lot will depend on the activity of your array

when you extend a RAID 5 array over more disks, the new drives have to be incorporated into the array.  this involves re-writing and calculating all the parity across the array to make sure the additional space can be used.

If you have a 4x300 GB array then 4 days could be accurate.

The array will be available unless you chose to take it offline

Performance will be degraded but not that badly
1. How long really depends on a number of things which include: the I/O activity on the array, the priority setting of rebuild/expansion on the controller, the speed of the disks, and the speed of the controller.
2. I presume you have the MSA2000 connected to an HP-branded controller.  HP-branded controllers will rebuild/expand an array online (meaning is will be available).
3. That depends on the priority you have configured for rebuilds/expansions on the controller.  It's a simple trade off -- performance of array versus speed of rebuild/expansion.  Since the array is online during this time, "how long" is virtually impossible to calculate.

However, the expansion can be done with both disks at the same time.  I would recommend starting this on Friday after business hours to minimize user impact which in turn gives the controller more cycles to work on the expansion versus serving user requests.  Of course, I'm presuming your low I/O points -- since you know your system better, start the process right before your normal low I/O periods.
Free Tool: ZipGrep

ZipGrep is a utility that can list and search zip (.war, .ear, .jar, etc) archives for text patterns, without the need to extract the archive's contents.

One of a set of tools we're offering as a way to say thank you for being a part of the community.

georgemasonAuthor Commented:
Hi both,

Thanks for your comments. The array is under pretty light load so am hopeful that it shouldn't take too long. Downside is that its a 7 day week office so there are no days when the machines are completely free, although the weekend is less busy.

The disks are all 15K SAS so should be as fast as they can be, the controllers are both HP devices (shipped with the array). At the moment the "utility priority" is set to high - assume this is the rebuild priority?

Can I change the rebuild priority on the fly? Perhaps then I can start it on high priority and lower if it impacts the user experience. I will try and move some VMs off the SAN before adding the disks if only to lower the load on the storage as it does the rebuild.

Finally, what is the situation re. disk redundancy whilst it's adding the disks?


georgemasonAuthor Commented:
Understood. So I guess the only question is, is there any risk in doing this at all? I'm in a position where I can move some of the more critical VMs off the array to local storage for the duration of the rebuild, which I think I might do in any case - but is this a risky process or run-of-the-mill?

(don't worry I won't sue if it goes wrong - just trying to get a feel for it!)

HossyConnect With a Mentor Commented:
There's always a risk and there is a slightly increased risk because you are drastically increasing the I/O on the disks themselves.  You maintain the same redundancy you currently have (presuming you're on RAID 5 currently) through the process.  If you have old drives, it wouldn't hurt to move critical systems off first.  Nonetheless, I would classify this as "run-of-the-mill" before I classify it as a "risky process".

Of course, a good rule of thumb: BACKUP before you start. :-)

Good luck.
georgemasonAuthor Commented:
All received and understood, thanks. Will move the sensitive machines off then run the upgrade.

Thanks to all for your contribution, will split the points. Will try to remember to update this with the results of the vdisk expansion to inform someone else in my situation another time!
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.