Increasing disks size when you have two vdevs?

[jdabb]

This will be the second time that I have run out of space and closing in on 98% of used space. I now will probably have to replace each disk wit larger storage. The problem is i am not sure if i will have to replaced all drives in both vdevs before seing the increase in space or will i get the increasing after replacing all in one of the vdevs?

 

[Samuel Tai]

Yes, you'll need to replace all members of the pool with larger disks before the pool expands.

 

[Basil Hendroff]

@jdabb Out of interest and to provide some context, what's the makeup of each VDev? Disk numbers and size? VDev RAID format?

I have an uneasy feeling that you're going to struggle to replace disks due to the lack of free working disk space. Search the forum for 80% rule. Start with this post https://www.truenas.com/community/threads/maximum-amount-of-space-useable.28579/post-187724.

 

[HoneyBadger]

All drives in a vdev need to be replaced with larger ones before the space will be available; it is not necessary to replace every drive across all vdevs. As suggested by @Basil Hendroff please post some details of your pool and vdev layout.

 

[jdabb]

6 in each vdev all 4TB. I should have bit the bullet and went all 10tb on the first vdev the first time, I wanted to go cheap at that time so i expanded to a second vdev. Now im stuck i guess with little options.

 

[jdabb]

raidz2 BTW.

 

[Basil Hendroff]

That's around 30TB of total (100%) usable ZFS storage capacity.

I'm just wondering whether the safest, risk-mitigating (and, in the long run, arguably cheapest) option is to build a second server with a pool based on a single RAID-Z2 VDev with at least seven 10TB NAS drives and then replicate data between the two servers? Seven disks will provide 36 TB of usable disk storage under the 80% limit. One advantage of the two server approach is that there is now a means, through replication, to backup 30TB+ of data rather than just relying on disk redundancy as the only safeguard. A second advantage of this approach is that you'll only be reading data from the existing pool and not manipulating its underlying structure. Given how full the pool is and how much data is at stake, I'd just be uncomfortable with replacing disks within a VDev, especially without a backup! A third advantage is that once your data is safely replicated you could rebuild the pool on the original server using a single RAID-Z2 VDev with the 12 x 4TB disks to give you a total (100%) usable backup capacity of around 35TB, or, at a pinch, a RAID-Z1 pool with a total capacity of around 41TB, which might be okay for a backup system.

@jdabb Is this a financially viable option for you? If not, I strongly encourage you to find some way to backup 30TB of data before you attempt to replace disks within a VDev.

 

[jdabb]

It’s an option didn’t want to spend a lot of money to fix this issue but it looks like I have no choice.

 

[Basil Hendroff]

One other option you might like to consider is to copy the existing data off onto three external 10TB external HDDs. You can build a new optimally configured pool, or at least try to upgrade the current pool, on the existing server and then copy the data back into the pool. This is a clumsier approach. The process will be slow and carries some risk, but it's a cheaper option. Your data won't be readily available during the process. You will still have an issue with the long-term backup of data as well. Replication trivialises the backup process.