Hi, having thought I was past these issues with Unraid all of a sudden it's crashed and crashes on mounting array with error PANIC: zfs: adding existent segment to range tree (offset=1a74e203000 size=20000)

I don't know how to fix this! Really wish I'd never changed to zfs, it's been more trouble than it's worth!!

Hoping someone knows whats going on here?

Thanks,

Tim

tower-diagnostics-20250709-1424.zip

See if you can get the syslog from the CLI to see which filesystem has the issues:

cp /var/log/syslog /boot/syslog.txt

Ok, syslog attached

syslog.txt

Also disk3 is my zfs disk in the array, having disabled it, I can't add it back without erasing it!!!!

Syslog is before array start, so it doens't yet show the error.

As soon as error happens system crashes, so you can't do anything?

You should still be able to get the syslog from CLI, using SSH or a keyboard/monitor, if you can't, enable the syslog server and post that after start.

syslog server is enabled and set to save to flash drive. So should be in diagnostics?

Yes, but if you rebooted, it will be renamed to syslog-previous

Ok, there is both in the first diagnostics, but here is another in case it missed it somehow! :)tower-diagnostics-20250709-1817.zip

Still no array start logged.

Ok, strange! So I've just started the array and crashed system, I was able to copy syslog using keyboard and screen connected to server and it's attached. Array started at around 10:59

Thanks,

Tim

syslog.txt

The panic happens a few seconds after all zfs filesystem are imported so not sure which one is the issue, since you have one disk and two zfs pools, temporarily remove both pools, start the array, and see if it doesn't crash, If it does, it's the disk, If it doens't add the pools one at a time and retest.

Ok, so I removed both pools and had already removed disk3, but still the same problem!?

syslog.txt

That suggests disk3 is the problem. To confirm, with the array stopped, click on disk3 and change the filesystem to xfs for example, then start the array, disk3 won't mount, DON'T FORMAT IT, just see if it the server no longer crashes, and if not, re-add the pools and retest to confirm thay are OK, and we can then try to deal with disk3 data.

Ok, so the array has started! disk3 is the backups disk, so if necessary can be deleted and the backup recreated! It's my TimeMachine and appdata backups! Re-adding pools and will let you know!

Thanks,

Tim

Great, to re-add the pools:

create a new pool with the same name and number of slots

assign the pool device(s), leave the filesystem set to auto

start the array to import the pool

Ok that's worked, I did it before your answer and set the filesystem to zfs but it seems to have imported ok! disk3 is doing a data rebuild, how do I erase it and put it back to zfs?

Thanks,

Tim

If it's still set to zfs and it can be formatted, i.e., no important data format it now, then change back to zfs and format again.

Ok, done that and it's now re-building data! Do I let this happen? Or will it just put the corupted data back and cause same issue?

If it was formatted, it's rebuilding a new and good filesystem.

Ok, I'll leave it alone then! 😊

Thanks again for you help!