Start a Conversation

Unsolved

This post is more than 5 years old

46996

April 21st, 2013 03:00

Dell PowerEdge T710 drive upgrades

Hi,

I'm looking for some help for a friend of mine. They have a Dell PowerEdge T710 server with 3x 146gb sas drives in a raid 5 configuration.

They have bought 4 new 600gb sas drives to replace these. 3x to replace the 146gb drives and 1 as a hot spare.

Can someone please advise me of the best way to replace these drives with the minimum downtime.

I looked at using "replace member disk" in the server administrator and wondered if I used this 1 by 1 on each drive it would get the out come he wants.

It has a PERC 6/I raid controller. Thanks for any comments in advance.

Moderator

 • 

6.2K Posts

April 21st, 2013 11:00

Hello youservices

The recommended method and best practice is to backup data, delete/recreate array with new disks, reinstall OS, and restore data from backup.

The unsupported method is to offline a drive, replace it with a larger drive, and then allow the replacement to rebuild. Once the rebuild completes and the array is in an optimal state rinse and repeat with the remaining drives(Important! One drive at a time). After all of the drives are replaced and rebuilt you will need to retag the array to allow the use of the extra space. You will need to write down the current array settings(Stripe size, disks in array, RAID level, etc) and then delete the array. After you delete the array you will need to recreate it using the same settings as before(Important! Do not initialize the array. This will delete the data). This will rewrite the meta data and allow use of the new drive space. Alternatively, you can create a sliced array using the additional drive space by just creating a new array across the drives.

The second method has a risk of data loss, so it is still advisable to have a valid tested backup before attempting the rebuild-retag. If you need more details on how to accomplish this then I suggest searching the forums. This question is asked quite frequently.

Thanks

7 Technologist

 • 

16.3K Posts

April 22nd, 2013 10:00

"The unsupported method is to offline a drive, replace it with a larger drive, and then allow the replacement to rebuild. Once the rebuild completes and the array is in an optimal state rinse and repeat with the remaining drives(Important! One drive at a time). After all of the drives are replaced and rebuilt you will need to retag the array to allow the use of the extra space. You will need to write down the current array settings(Stripe size, disks in array, RAID level, etc) and then delete the array. After you delete the array you will need to recreate it using the same settings as before(Important! Do not initialize the array. This will delete the data). This will rewrite the meta data and allow use of the new drive space."

This is unsupported, because it is dangerous.  It will create an unstable array that will be subject to many types of future failures.  If this server is providing critical data or services, I'd strongly recommend against a "retag" to regain your space.  Especially where this is not even your server, I'd make sure you do it right ... backup/delete/replace/create/restore or create a new/additional array with the larger drives to move data too, or grow your RAID by adding disks to the RAID 5 (3-disk RAID 5 -> 5-disk RAID 5, for example - which obviously depends on how many bays you have).

No Events found!

Top