Having converted this from CORE using the progressive update method detailed in the documentation (Core 13.3 → Scale 24.04 → Scale 25.04), I’m having difficulty continuing to upgrade to 25.10:
[EFAULT] Error: Command ['chroot', '/tmp/tmpbosykf88', 'mkdosfs', '-F', '32', '-s', '1', '-n', 'EFI', '/dev/sdp1'] failed with exit code 1: mkdosfs: unable to open /dev/sdp1: Device or resource busy
It looks like /dev/sdp1 is currently mounted as /mnt/boot, which is probably the cause of the issue. Any idea how I should proceed?
Is everything you need in the 25.04 version? Make sure you have a current system configuration download. You can try deleting the 25.10 boot environment, if there is one, and trying to upgrade again. Another option could be a clean install of 25.10 and restoring the system configuration.
Are you attempting to upgrade 25.04.2.6 directly to 25.10.1 or did you do 25.10 and then upgrade to latest 25.10 version? Do you have enough freespace on your boot pool?
Kind of guessing since you didn’t post details for your system
I did grab a configuration download before attempting the upgrade. I am attempting to upgrade using the built-in upgrade in the WebUI, switching train from the current “TrueNAS SCALE Fangtooth 25.04 [release]” train to “TrueNAS SCALE Goldeneye 25.10”, which is saying it will Upgrade TrueNAS-25.04.2.6 -> TrueNAS-25.10.1
As to system specs:
IBM x3650 M3 chassis (16xSFF drive bays)
Dual Xeon X5677
72GiB of DDR3 ECC RAM
Boot Volume: 120GiB SATA SSD
Container/VM Storage: 6(RAIDZ1)x1TB SAS Drive (Seagate, IBM branded)
Storage: 2x5(RAIDZ1)x5TB (SMR) SATA Drive (Seagate)
DataDrives is showing at 96% full. Recommendation is below 80% usage. 50% if block storage. I don’t think that is the issue here but that is considered the ZFS danger zone leading to crashes and possible pool loss.
Did it create a boot environment for 25.10 or it’s stopping before that. Options could be attempting an upgrade by downloading the Manual Update and trying that or booting an install image of 25.10 and doing an upgrade. I think it should give you that option in the menus. Have the pools been scrubbed recently?
Yes, both pools are set to scrub weekly. I am aware that 96% is bad, and I am intending to move the data off to another machine/array, but currently 5TB is the largest capacity HDD you can get in a 2.5” size (Yes, I know about 8TB QVO SATA SSDs and/or 15/31TB Enterprise SAS SSDs, no I don’t have the budget for that).
There does not appear to be a dataset created for 25.10 in the boot drive (it looks like that’s what it was trying to do?), it hasn’t rebooted to start the install proper this is all within the webui at this point.
I would try using the manual download option first to see if it was just the automatic method that was getting stuck. If that doesn’t work, it would be a fresh install with 25.10 and reloading the system configuration.
I would go with that release and then do the upgrade to 25.10.1 if that is successful. You will lose the previous boot environments for 25.04 and earlier with fresh install, though.
As long as you were aware of the full pool, you should be okay. I’ve had a few users stuff their pool and then we couldn’t get back in or we had to work at deleting files. System gets unresponsive with much more data.
I should add that upgrading to 25.10 puts you in the Early Adopter category at this time. Once they release the next version and move 25.10 to General status on the software status page, then you could choose that. If you upgrade now, you might be stuck in Early Adopter. I don’t know if you can down grade to more conservative status. This is a recent change on the ‘upgrade trains’ shown.
Okay, attempting a reboot gets me to GRUB with a “25.10.1” revision, but trying to boot it gets the message “25.10.1 not found”
Loading Linux 6.12.33-production+truenas ...
error: file `25.10.1' not found.
Loading initial ramdisk ...
error: you need to load the kernel first.
Press any key to continue...
Pressing the key takes me back to the GRUB menu, where fortunately, there is also a menu option for 25.04, which I can use to boot into the earlier version.
It looks like I may be trying a manual update, or even booting from 25.10 install media and install/upgrading from there.
…And attempting a manual update gets me all the way up to:
[EFAULT] Error: Command ['chroot', '/tmp/tmpb_8ntmwm', 'efibootmgr', '-c', '-d', '/dev/sdp', '-p', '1', '-L', 'TrueNAS-0', '-l', '/EFI/debian/grubx64.efi'] failed with exit code 5: Could not prepare Boot variable: No space left on device
Do you have a lot of previous boot environments? I am confused because you show plenty of space with zpool list on the boot-pool. If you aren’t in a hurry, waiting for other users to post may help. I’m not sure where it could be having ‘space’ trouble.
We could try checking the partitions for the boot device but not sure what to really look at or where the GRUB is supposed to be.
Okay, so going into the (very early) EFI system for this server and deleting some defined boot options, along with booting directly from the 25.10 ISO and installing fresh seems to have worked. I’m now waiting for the system to finish restarting after importing my previous (25.04) configuration, which is currently sitting at ix-zfs.service/start and ix-netif.service/start, which have been running for the last 5 minutes.
