Replacing two drives with one bigger drive

[GoChris]

I checked the wiki, but only seemed to find info about replacing one disk with a bigger one.

 

I have 2x500gb drives I want to replace with 1x2tb drive.

 

What would my process for this be? I've already pre-cleared the new drive.

 

Edit: Should mention I'm on 4.7 still.

[JonathanM]

I checked the wiki, but only seemed to find info about replacing one disk with a bigger one.

 

I have 2x500gb drives I want to replace with 1x2tb drive.

 

What would my process for this be? I've already pre-cleared the new drive.

 

Edit: Should mention I'm on 4.7 still.

Unraid has no built in function to reduce the number of slots used. It assumes you will always be adding slots, or upgrading existing slots. You could replace (upgrade) one of the disks and leave the other in place. If you really have to remove both disks, then after you have upgraded one, you could manually copy the contents of the other old disk onto the new disk, then remove it and rebuild parity after issuing the initconfig command. You would be without parity protection while it was being rebuilt, so a failure with another drive would not be recoverable.

[maxinc]

I've been in this situation recently when I had to replace 4 x 1TB drives with 2 x 2TB. I needed the 1TB drives for something else and I like running the tower with the minimum amount of drives but not more than I actually need. For me 2TB of free space is enough and I add more capacity once I'm well into the last TB.

 

There is information on the WiKi on how to remove drives and separate info on how to upgrade drives.

http://lime-technology.com/forum/index.php?topic=2088.0

http://lime-technology.com/wiki/index.php/Replacing_a_Data_Drive

 

As jonathanm mentioned, when you remove a drive which you don't intend to replace, unRAID will have to recompute the parity which will leave all your data unprotected for several hours. To minimise any potential problems I would suggest to perform a parity check before and after each procedure takes place and inspect the SMART report for the drives after each one of them.

 

Here'e what I'd do in you case.

 

1. Parity check + SMART report check. Make sure there are no errors in the array to start with.

2. Replace 1st 500G drive with 2TB drive. (see 2nd link) -> Wait for the upgrade several hours.

3. Parity check. SMART report check. Make sure all went ok and new drive is healthy.

4. Copy all data from the 2nd 500G disk to the upgraded disk.

5. Stop array. Unassign drive from the the array. Telnet into server > initconfig (see instructions in the 1st link).

6. Start array and wait patiently to rebuild. -> SMART report check.

7. Parity check. SMART report check. Before using the array again, make sure all went well.

 

It is a long process and it would be a lot shorted if you would have kept all 3 drives. On the other hand this is a good opportunity to check the array is in good shape and identify any potential weakness in the system.

 

One quicker possibility if you have a spare SATA port is to:

 

1. Add 2TB drive into the array. Being pre-cleared this should take only a few minutes.

2. Copy data from 2 x 500G drives.

3. Parity Check + SMART

4. Remove both 500G drives at the same time. Initiconfig + rebuild.

5. Parity Check + SMART

[dgaschk]

When checking SMART reports there are 2 things to check. The overall self-assement should be PASSED and the pending_sector_count RAW_VALUE should be zero.

[Johnm]

I would do it maxinc's first way.

 

It is time consuming, but your array is never "down" except while powered off to replace the drives at those particular steps.

I would take several evenings and let it do each build/check step overnight.

 

 

[GoChris]

Thanks for the info guys. I figured at one point I was going to have to rebuild parity.

 

I do have a slot in there right now as I've already pre-cleared the 2TB. So it's just sitting waiting. If I have to rebuild the parity at some point anyway, wouldn't maxinc's second method be just the same but much quicker? Then the array is also never down until the one rebuild. I'll start a parity check now anyway as it's been 80 days since the last one before I proceed with anyway. 

[maxinc]

I don't see why it wouldn't work. When transferring data between disks, I would copy it rather than moving it, this way I would still have a copy if the rebuild fails for any reason. Also, it is best if you don't use the server during the rebuild (like watching movies or copying data) to minimise stress on the busy disks. For me is always a time of great anxiety when such operations take place )

[GoChris]

Alright, I'm not sure what I'm doing wrong. I did a parity check, all SMART statuses looked good. Added the 2TB and copied all data from the 2 500GB drives. Did a parity and everything was good again. Unassigned the drives. Ran initconfig. Powered down, unplugged the 2 drives.

 

Now I don't see any rebuild button. I'm on 4.7. This is all I see.

[dgaschk]

Please show the entire page. And attach a syslog.

[GoChris]

 

Syslog is attached or available here: https://dl.dropbox.com/u/857800/syslog-2013-01-12.txt

syslog-2013-01-12.txt

[Joe L.]

Likely you did not answer

Yes

(Capital "Y", lower case "es") to the question asked by the initconfig command.

 

Joe L

[GoChris]

Well I thought I did, but no, I guess I typed yes instead of Yes. Doh.

 

So now I can just hit start correct? Also before I do that, can I just move all the drives up to fill in the two "Not installed" slots?

 

[GoChris]

Alright, I re-organized the drives, and hit start. Parity is rebuilding...eeeek. It's scary. Should be done by morning I hope.