Disk Enclosures
1748216 Members
3424 Online
108759 Solutions
New Discussion юеВ

Re: Hp Surestore Disk Array 12

 
Darren Murray
Advisor

Hp Surestore Disk Array 12


Hi guys,

I have a Autoraid box which has 12 x 9gig disks. I have now available from another server
6x 18gig disks.

The server is currently in a mc/service guard
environment.

Whats the best way to upgrade the disks? Can I just take one 9gig disk out at a time and will the auto include being on automatically include the disk into the array?

Any help would be appreciated!

Thanks
Darren
Can you imagine life without beer?
5 REPLIES 5
Bill McNAMARA_1
Honored Contributor

Re: Hp Surestore Disk Array 12

make sure that autorebuild is on too.

arraydsp -i
to get the serial number

and arraydsp
to ensure that the State = Ready

If the state is anything other than READY... do not pull out anything.

But yes, what you suggest doing is fine.. one disk at a time, wait for rebuild, insert other disk.

See the storage forum for more on the autoraid.
http://forums.itrc.hp.com/cm/CategoryHome/1,1147,195,00.html


http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0x71e4f9beca68d511abcd0090277a778c,00.html

Later,
Bill
It works for me (tm)
David Allen
Frequent Advisor

Re: Hp Surestore Disk Array 12

Hi Darren,

Firstly, are you replacing all of the 9GB disks with the 18GB disks? If so, you will probably end up with much less available disk space in the array to allocate to LUNS. Assuming you will have the active hot spare turned on (recommended), the Auto Raid will allocate 18GB to the hot spare, where as with the 9GB disks the Auto Raid will allocate only 9GB. More space will be allocated to redundancy as well.

For example, one of our customer sites has an Auto Raid that has 6x 18GB disks (formatted capacity is 17366MB each disk). 17366MB is allocated as the hot spare, 22057MB is allocated for redundancy, which leaves 64776MB to allocate to LUNS. Another customer site has an Auto Raid with 12x 9GB disks. 8683MB is allocated as the hot spare, 17061MB for redundancy and 78454MB available to allocate to LUNS.

We recently upgraded the site with the 6x 18GBs from 12x 4GB disks and we debated about replacing them the way you are suggesting. We decided against it and went for destroying the LUNS, removing the 4GB disk and installing the 18GBs and then recreating LUNS, volume groups, logical volumes and restoring data from backup tapes (we had duplicate backup tapes as a safe guard).

What you need to consider with taking the approach of removing a 9GB disk and replacing it with an 18GB, is that the Auto Raid would need to rebuild it's raid set and the new hot spare size would be 18GB. If your Auto Raid is struggling for space at the moment, pulling a disk and replacing it with a larger disk could really upset the apple cart. Once the Auto Raid begins rebuilding it could take several hours to complete (for each disk replaced).

This is merely what I would suggest from personal experience, others in the HP support world may have a different opinion. If you don't get a clear answer from this forum, place a service call with HP and get it clarified from them.

Regards,
Dave

BTW, Coopers Pale Ale is a much better beer than VB.
Insu Kim
Honored Contributor

Re: Hp Surestore Disk Array 12

Backing up the whole data and restoring them later to the AutoRAID populated with 18gig disk modules would be the best, quickest and safest.

If you take a module out, AutoRAID will perform "rebuilding and balancing" and then a replacement goes in, the array will do the same thing.

You have to think about the entire 9gig disk modules, it will take tremendous time to replace every single 9 gig disk step by step.

Hope this helps,

Never say "no" first.
Dan Eastman
Occasional Contributor

Re: Hp Surestore Disk Array 12

Ours worked fairly well. We swapped six 9's for 18's It will take some time. Make sure your firmware is up to date and of course your backup.
I signed it?
A. Clay Stephenson
Acclaimed Contributor

Re: Hp Surestore Disk Array 12

Some months ago, I swapped out 12 9 GB drives for 12 36GB several months ago and the entire process worked just as expected. It was an MC/SG production cluster so that I could not take the array down. As mentioned earlier, take your time; wait until the array is 'Ready' before replacing the next drive; and all will go well. The first two drives took several hours but the remaining drives take less and less time. I think I replaced 4 drives the first day and the remaining 8 drives the next day.

The good news was that I then had 12 9GB drives which I then installed in a fully loaded (3 power supplies; 2 controllers) acquired from a used equpment dealer for well under $4K. It made an excellent addition to my sandbox environment.

Regards, Clay
If it ain't broke, I can fix that.