My unraid system crashed one day and I've been trying to figure out why I cannot start the array anymore without restart. It appears that the server crashes once the 2 NVME ZFS cache pool is mounted. My issue seems to be similar to this one but I cannot confirm. I also am not sure how to back up my 2 cache disks and restore it as it is a zfs pool.

Thanks in advance!

lancenas-diagnostics-20260510-0007.zip

See if the pool mounts read-only

zpool import -o readonly=on cache

1 minute ago, JorgeB said:

See if the pool mounts read-only

zpool import -o readonly=on cache

It does mount as read only

The diags posted don't have the normal mount attempt, so can't see the crash, but if it's crashing in read/write mode, you can backup the pool and reformat, data will be under /mnt/cache. You can also try a normal import and post a new syslog to see the actual crash

zpool export cache

zpool import cache

35 minutes ago, JorgeB said:

The diags posted don't have the normal mount attempt, so can't see the crash, but if it's crashing in read/write mode, you can backup the pool and reformat, data will be under /mnt/cache. You can also try a normal import and post a new syslog to see the actual crash

zpool export cache

zpool import cache

How do I backup a pool? My plan (if it was only one NVME) was to clone it to my PC and then then restore it onto the reformatted drive.

Here's what I get after running the commands:

I can't access the server anymore

Just now, LanceM26 said:

How do I backup a pool? My plan (if it was only one NVME) was to clone it to my PC and then then restore it onto the reformatted drive.

Here's what I get after running the commands:

I can't access the server anymore

Actually, I was trying to access using HTTPS, I can access using HTTP. Here are the new logs

lancenas-diagnostics-20260510-0217.zip

I don't see any crashes logged, did the pool fail to import in normal read/write?

1 hour ago, LanceM26 said:

How do I backup a pool?

If it only imports read-only, after doing the import, start the array and copy from /mnt/cache to the array (or another pool), you can use, for example, Midnight Commander (mc on the console)

3 hours ago, JorgeB said:

I don't see any crashes logged, did the pool fail to import in normal read/write?

If it only imports read-only, after doing the import, start the array and copy from /mnt/cache to the array (or another pool), you can use, for example, Midnight Commander (mc on the console)

This should still be ok right?

Yes, it's normal for the WebGUI to show unmountable when it's imported manually.

11 hours ago, JorgeB said:

Yes, it's normal for the WebGUI to show unmountable when it's imported manually.

Data has been successfully backed up to another drive. Is the format process just a simple stopping the array and pressing format on the GUI and then copying the /mnt/cache backup?

If all the data is copied export the pool first

zpool export cache

Then stop the array, click on the first pool device, and "erase pool", then start the array and format it.

16 minutes ago, JorgeB said:

If all the data is copied export the pool first

zpool export cache

Then stop the array, click on the first pool device, and "erase pool", then start the array and format it.

I'm getting the above error

I waited a bit and was able to run the command. It's now formatted and when I try to copy the backup into the new /mnt/cache, unraid just restarts. I've attached a new diagnosticslancenas-diagnostics-20260512-0021.zip

Server restating on its own is usually a hardware issue (or bad power); enable the syslog server and try again.

1 hour ago, JorgeB said:

Server restating on its own is usually a hardware issue (or bad power); enable the syslog server and try again.

I fear it may be the Samsung NVMe I have in the cache but am not sure. Smaller directories copy fine but when I try to copy everything in, unraid crashes. Syslog turned on and diagnostics attached.

lancenas-diagnostics-20260512-0206.zip

Nothing relevant is logged in syslog-previous, if you can, try using a different device, but typically, even if it's bad, that should not make the server restart

4 hours ago, JorgeB said:

Nothing relevant is logged in syslog-previous, if you can, try using a different device, but typically, even if it's bad, that should not make the server restart

I am not sure how to proceed if even the syslog isn't showing what's wrong with the system...

6 hours ago, LanceM26 said:

Samsung NVMe

Maybe something here for you:

6 minutes ago, trurl said:

Maybe something here for you:

I believe mine is already running the latest firmware

24 minutes ago, trurl said:

Maybe something here for you:

18 minutes ago, LanceM26 said:

I believe mine is already running the latest firmware

Also adding to this, Unraid 7 doesn't have nerd tools plugin so I can't get 7zip on it

1 hour ago, LanceM26 said:

I am not sure how to proceed if even the syslog isn't showing what's wrong with the system...

Like mentioned, server rebooting by itself is almost always a hardware problem, and those types of issues don't typically leave anything logged, since it just reboots, but if you have another device that you can use for cache to rule out the NVMe it may be worth a try.

2 minutes ago, JorgeB said:

Like mentioned, server rebooting by itself is almost always a hardware problem, and those types of issues don't typically leave anything logged, since it just reboots, but if you have another device that you can use for cache to rule out the NVMe it may be worth a try.

I have removed the samsung drive from the cache pool and will try with the T-FORCE NVMe as the only device in the cache pool

6 minutes ago, JorgeB said:

Like mentioned, server rebooting by itself is almost always a hardware problem, and those types of issues don't typically leave anything logged, since it just reboots, but if you have another device that you can use for cache to rule out the NVMe it may be worth a try.

2 minutes ago, LanceM26 said:

I have removed the samsung drive from the cache pool and will try with the T-FORCE NVMe as the only device in the cache pool

Yup lol, it is copying the files without restart when the samsung drive is removed from the pool.