Like to add additional Storage space to RAID Z

Status
Not open for further replies.

Freesnofla

Patron
Joined
Oct 27, 2013
Messages
216
Hi Guys, actually I run FreeNas with RAID Z. The installation runs on 2 HDD 2 TB of each. If I like to increase the capacity to 8 or 12 TB what steps do I have to take care about? Could I always increase by 2 x 2 TB or could I also use 2 x 6 TB?

best regards
Freesnofla
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,194

enemy85

Guru
Joined
Jun 10, 2011
Messages
757
You can add more vdevs (2 disks each time if u have just a 2 disks pool) or you can replace the 2tb disks with greater ones (one each time doing a resilver)
 

Freesnofla

Patron
Joined
Oct 27, 2013
Messages
216
Well just wrote the guide from Cyberjock and still in doubt what exactly I should do. Vdev is in my understanding a summarize of HDD, which all together give me the possibility to have more Space available. If I would increase with 2 x 2 TB I just can install the HDD's and then format them? Then finally copy the data to the new devices. Or do I need to do some additional formatting ? What does resilver exactly mean?
 

danb35

Hall of Famer
Joined
Aug 16, 2011
Messages
15,504
RAIDZ has 3 or more disks; if you have two disks, you have a mirror. Your two disks comprise one mirrored vdev, and your pool consists of that one vdev. To increase the capacity of your pool, you can (1) replace your disks with larger disks, one at a time, following the manual's instructions for replacing failed disks--once you've replace both disks, your vdev capacity (and thus your pool capacity) will increase to reflect the larger disks; (2) add more disks, which will comprise one or more vdevs and will be striped with your existing vdev to expand the capacity of your pool; or (3) some combination of (1) and (2).

If you have two more 2 TB disks, and want to add them to your pool and maintain redundancy, here's how you'd do it:
  • Install the two new disks in your server
  • Log into the web GUI
  • Click the Storage button
  • Click the ZFS Volume Manager button
  • In the "Volume to extend" dropdown, choose your pool
  • Click to add the disks
  • Make sure the dropdown is set to "mirror" (which it should default to)
  • Click the "Extend Volume" button
The system will create a mirror vdev using the two new disks, and add that vdev to your existing pool. The capacity of your pool will be increased by the capacity of the new vdev, about 1.8 TB. This process is irreversible--once you've added that vdev, you can't remove it. You will still have redundancy in your pool--any single disk can fail, and you won't lose any data. Two disks can fail, but only if they're the "right" two disks--they would need to be one from the old mirror, and one from the new mirror. If both disks in either mirror failed, you would lose all your data.
 

RegularJoe

Patron
Joined
Aug 19, 2013
Messages
330
It is called OCE = Online Capacity Expansion.

Apparently there are some ZFS gods that do not like this feature. I think it could work, redistribute all blocks to the end of the disk starting at the front, once that is done let the file system place new files or frequently used files at the front of the disk as this is the faster part of the mechanical anyway.

Here is a discussion on the subject :

http://zfsguru.com/forum/zfsgurusupport/608

Hoping ZFS version 5001 has this feature. ;)
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,680
I think it could work, redistribute all blocks to the end of the disk starting at the front, once that is done let the file system place new files or frequently used files at the front of the disk as this is the faster part of the mechanical anyway.

What benefit do you think there would be, differentiating faster from slower transfer areas on disk, when a NAS is fundamentally limited by network speed (a speed which is typically slower than a single disk)?

I have no idea what this has to do with what was being discussed in any case.
 

RegularJoe

Patron
Joined
Aug 19, 2013
Messages
330
I believe the ZFS cache algorithm will not use the LOG device if the IO is large and linear, i.e. do no blow out the cache on a large file write. I am not sure how large a large write is.
 
Status
Not open for further replies.
Top