Pool degraded/write errors after adding a new drive.

[HoneyBadger]

It is indeed possible (on TrueNAS 12+) to remove top-level vdevs from pools that contain only mirrors or stripes; it does consume some small amount of additional RAM, but if the goal is to rebuild eventually then it might be acceptable.

 

[Etorix]

any way of just removing the drive and even lose that 73GB?

A single drive vdev is a 1-way mirror, so you should be able to remove the drive from the GUI, and have these 73 GB transferred to the 8 TB drive.

 

[Whattteva]

A single drive vdev is a 1-way mirror, so you should be able to remove the drive from the GUI, and have these 73 GB transferred to the 8 TB drive.

It's not though. He's basically added that second drive and striped the pool to include that drive. So now the vdev that used to be 1 disk now has 2 disks.

 

[Etorix]

…making a pool of two vdevs which are 1-way mirrors. There's no such thing as a "stripe vdev".

 

[HoneyBadger]

…making a pool of two vdevs which are 1-way mirrors. There's no such thing as a "stripe vdev".

In a sense, they're all vdevs of one device each.

It's supported, but it seems like the UI is missing the option for the single-drive removal option.

@bar1 You can run the following command from SSH to

zpool remove -n Bar1-8TB de364a5e-87b8-4b39-8d76-cecc219c7bb8

This will let you know how much RAM will be consumed after the migration. If it's acceptable, remove the -n and re-run the command to do it for real.

 

[bar1]

In a sense, they're all vdevs of one device each.

It's supported, but it seems like the UI is missing the option for the single-drive removal option.

@bar1 You can run the following command from SSH to

zpool remove -n Bar1-8TB de364a5e-87b8-4b39-8d76-cecc219c7bb8

This will let you know how much RAM will be consumed after the migration. If it's acceptable, remove the -n and re-run the command to do it for real.

Thanks , confirm my cloud backup is up to date and will attempt.

 

[bar1]

zpool remove -n Bar1-8TB de364a5e-87b8-4b39-8d76-cecc219c7bb8
Memory that will be used after removing de364a5e-87b8-4b39-8d76-cecc219c7bb8: 398K

This is a little unclear to me....what does it actually mean?
"Memory that will be used"

 

[HoneyBadger]

zpool remove -n Bar1-8TB de364a5e-87b8-4b39-8d76-cecc219c7bb8
Memory that will be used after removing de364a5e-87b8-4b39-8d76-cecc219c7bb8: 398K

This is a little unclear to me....what does it actually mean?
"Memory that will be used"

It means that ZFS will need another 398K of memory - about a third of a megabyte - to keep track of the location of data that gets moved off of that 18T disk.

Go ahead and re-run the command without the -n, and then you can issue zpool status -v Bar1-8TB to track the process of device removal.

 

[bar1]

It means that ZFS will need another 398K of memory - about a third of a megabyte - to keep track of the location of data that gets moved off of that 18T disk.

Go ahead and re-run the command without the -n, and then you can issue zpool status -v Bar1-8TB to track the process of device removal.

So will that move the files from the 18TB on my 8TB, or not?
sorry about all the questions.....

 

[Etorix]

Yes. -n is for a dry-run; without this option the remove command will do what its name says.

 

[bar1]

Hi,
So it started promising:
remove: Evacuation of /dev/disk/by-partuuid/de364a5e-87b8-4b39-8d76-cecc219c7bb8 in progress since Sun Jul 2 03:10:00 2023
22.4G copied out of 71.1G at 92.5M/s, 31.49% done, 0h8m to go

and then:
Removal of /dev/disk/by-partuuid/de364a5e-87b8-4b39-8d76-cecc219c7bb8 canceled on Sun Jul 2 03:18:50 2023

Should i remove the files with errors, zfs clear and zfs scrub first?
because i do have some errors again on this disk