What redundancy do I have? Created a raid-z first, then added another disk.

  • Thread starter Miguel Sequeira
  • Start date
Status
Not open for further replies.
M

Miguel Sequeira

Guest
Hello! The scenario is as follows:

I've created a raid-z with the following disks: 4x500GB+1,5TB. The total available space became 1.8TB.

After that I added a new 2.0TB disk, with the same volume name as the raid-z. The space available became 3.6TB.

While the description for the disks in raid-z is "Member of volume1 raidz", the new disk is "Member of volume1 stripe". All disks belong to volume1.

The question is: what redundancy do I have? Initially with raid-z, one of the disks could fail without losing data; after adding this new disk, did I loose all redundancy?

In general terms, if we start playing around with disk adding, how can we know what redundancy our system has?

Thanks for any feedback anyone may have.

Miguel.
 
Joined
May 27, 2011
Messages
566
you have zero redundancy. if your 2 TB drives dies, you're SOL.

to figure out redundancy, just figure what is the worst case scenario if you can loose any 1 disk, you have 1 disk of redundancy, if you can loose any 4 drives, you have 4 disks. if you can loose any disk except 1, you have no redundancy. to keep redundancy, you must add a redundant vdev. if you have a raidz, then you must add a raidz to the pool


to comment on your original raidz, you're only using 500GB from each disk, you're wasting 1TB from the 1.5TB drive. you'd be better served with a raidz of the 4 0.5TB drives and then a mirror of the 1.5 and 2 TB drives. then stripe them together. you'd have 3 TB of space and 1 disk of redundancy. and only waste half a TB.

However, I'd highly recommend you get a set of disks that are the same size.
 
M

Miguel Sequeira

Guest
Thanks for the advice, it makes sense to reconfigure. This all started with the 500GB disks. Then 2x1.5TB (one of the 1.5TB is faulty and will be returned to WD for replacement) and recently I purchased 3x2.0TB disks. But since I'm migrating from a linux software raid, I temporarily need the 2x2.0TB to backup all the stuff from the old raid. But in the end I will have 4x0.5TB + 3x2.0TB + 2x1.5TB. So would you say I should stripe a 4x0.5TB raidz, 3x2.0TB raidz and a 1.5TB mirror?

Regarding the redundancy calculation, thanks for clarifying it. If we add a vdev that isn't redundant in itself, it compromises the whole volume. It would be nice for freenas to warn the (unexperienced) user about it.
 
Joined
May 27, 2011
Messages
566
did you do it through the CLI or the GUI?

if you did it through the cli it did warn you.

Code:
  pool: test1
 state: ONLINE
 scrub: none requested
config:

	NAME                 STATE     READ WRITE CKSUM
	test1                ONLINE       0     0     0
	  raidz1             ONLINE       0     0     0
	    /mnt/storage/d1  ONLINE       0     0     0
	    /mnt/storage/d2  ONLINE       0     0     0
	    /mnt/storage/d3  ONLINE       0     0     0

freenas# zpool add test1 /mnt/storage/d4
invalid vdev specification
use '-f' to override the following errors:
mismatched replication level: pool uses raidz and new vdev is file

 
M

Miguel Sequeira

Guest
I used the GUI. Didn't know the CLI had these warnings, thanks for the information. I think it makes sense to request this information to appear in the GUI. Can I use your example to post in the features request area?
 
Status
Not open for further replies.
Top