[SOLVED] Failed to import pool rpool

[matthew]

I installed Proxmox 6.0-1 on an older Supermicro server using ISO image. Used ZFS and several drives in ZFS raid array configured from the proxmox install GUI. On boot I get this every time.

Command: /sbin/zpool import -N 'rpool'
Message: cannot import 'rpool' : no such pool available
Error: 1

Failed to import pool 'rpool'.
Manually import the pool and exit.

If I manually import the pool and exit the server then boots normally. Is there any fix so I do not have to manually import it on every boot?

 

[wolfgang]

Hi,

this sounds like the initializing timeout problem of the disk.
Look at this here https://pve.proxmox.com/wiki/ZFS:_Tips_and_Tricks#Grub_boot_ZFS_problem
it might help you.

 

[Stark-Johan]

I can confirm that adding the ZFS_INITRD_PRE_MOUNTROOT_SLEEP as described in the posted link solved my issue on 6.1-2. An important note is that updating the initramfs has to be done using the current kernel. I just wanted to clarify this for others that might be struggling with this as I did.

Check this by

pveversion -v

and look for the line similar to:

proxmox-ve: 6.1-2 (running kernel: 5.3.13-2-pve)

In my case the command used was:

update-initramfs -k 5.3.13-2-pve -u

 

[kamzata]

How can you do that from Busybox or a Rescue system?

 

[wolfgang]

Hi kamzata,

you can't set this at the rescue system.
You have to boot into the rootfs or set the kernel parameters.
Please read the manual at the link provided.

Look at this here https://pve.proxmox.com/wiki/ZFS:_Tips_and_Tricks#Grub_boot_ZFS_problem

 

[willybong]

I can confirm that adding the ZFS_INITRD_PRE_MOUNTROOT_SLEEP as described in the posted link solved my issue on 6.1-2. An important note is that updating the initramfs has to be done using the current kernel. I just wanted to clarify this for others that might be struggling with this as I did.

Check this by

pveversion -v

and look for the line similar to:

proxmox-ve: 6.1-2 (running kernel: 5.3.13-2-pve)

In my case the command used was:

update-initramfs -k 5.3.13-2-pve -u

Thanks a lot!

 

[SomeoneElseIsTalking]

Hi kamzata,

you can't set this at the rescue system.
You have to boot into the rootfs or set the kernel parameters.
Please read the manual at the link provided.

Hello Wolfang,

Would you mind explaining a little more than what the manual says? Below seems to be what I need to do :

To prevent it there have been some suggestions in the forum. Try to boot following the suggestions of busybox or searching the forum, and try [...] of the following: [...]
edit /etc/default/zfs, set ZFS_INITRD_PRE_MOUNTROOT_SLEEP='4', and then issue a "update-initramfs -k 4.2.6-1-pve -u"

How, in details, do I edit this file? In Busybox (initramfscommand prompt) directly? In Busybox, I'm able to modify and save this file using vi /etc/default/zfs, but after the command prompt does not recognize the update-initramfs. Upon a reboot, this file is back to its original state... Sorry, if I'm being so incompetent... Any help would be appreciated!

 

[SomeoneElseIsTalking]

So, I now feel like an idiot. The answer was fairly explicit, I must have got a typo or I was expecting something harder than that. Anyhow, ServeTheHome made a video on that subject, showing how to fix this issue : https://www.youtube.com/watch?v=kJB6BOtKKNU, as well as an article.

So, for future me, let me lay out the steps:
On the Busybox/initramfs command prompt (exactly where you should land if you are having this issue), simply type:
zpool import -N "rpool"
No output means it went OK.
Then, type:
exit
At this point, your system should boot as expected. Log on to your server using the root account with the password set during your setup.

Then to issue the correction proposed by the manual, edit your files, using the nano code editor:
nano /etc/default/zfs
and then, to know the kernel version to use, run
pveversion
you should get an output like this:
pve-manager/6.3-2/22f57405 (running kernel: 5.4.73-1-pve)
Then, finally, issue the final command using the kernel version of the previous output:
update-initramfs -k 5.4.73-1-pve -u

I don't know if it will help someone, one day, but if it does, then I wouldn't have written all this for nothing.

 

[Caledrith]

I don't know if it will help someone, one day, but if it does, then I wouldn't have written all this for nothing.

I made a forum account just to say that you helped me. Thank you so much for being one of the good souls that documents what they figure out!

 

[Andrii.B]

Nothing helps me. I can boot into kernel-6.5.13, but can't into kernel-6.8.8-2 or kernel-6.8.8-4. Tried to reinstall the kernel, added limits into /etc/default/zfs. It does not help.

 

[leesteken]

Nothing helps me. I can boot into kernel-6.5.13, but can't into kernel-6.8.8-2 or kernel-6.8.8-4. Tried to reinstall the kernel, added limits into /etc/default/zfs. It does not help.

Maybe it's because intel_iommu is on by default since 6.8, and your drive controller does not work with IOMMU? Maybe try intel_iommu=off (or maybe iommu=pt is enough)?

 

[Andrii.B]

intel_iommu=off was helpful for me. Thank you.