BTRFS: open_ctree failed after ubuntu image deploy

[Tom Elliott]

I feel I should at least kind of chime in a little bit.

The issue here is not in any way, shape, or form, related to the message as described in the title. The “random: nonblocking pool is initialized” is simply a kernel debug statement just telling you the pool to randomize elements in a non-blocking form has been initiated. This is NOT what is causing the failure to boot after upload, nor is it impeding with BTRFS in anyway.

I think @Quazz is right, at least in that we can perform a btrfs filesystem check. I doubt it will fix anything though. See, @Sebastian-Roth has successfully imaged the system using similar layout as yours, and while there are a few concerning error messages, the system is still operation. Maybe something else is causing issues?

[Sebastian Roth]

@Tom-Elliott said in Upload image "random: nonblocking pool is initialized":

See, @Sebastian-Roth has successfully imaged the system using similar layout as yours, and while there are a few concerning error messages, the system is still operation. Maybe something else is causing issues?

While you’re right that my system seemed to boot up properly after cloning I am still very concerned about those messages I posted. I just started up the system again. Booted ok, but I get a couple of these bad tree block start messages every minute now. I am trying to get in contact with the clonezilla developers about this as I think this is not a very special case and will hit is from time to time. I don’t think we should do raw imaging with btrfs filesystems just to circumnavigate this issue.

PS: Tom is right about the title. @Oleg would you mind changing the title to something appropriate?

[Quazz]

@Sebastian-Roth Have you tried a fsck? There might be some useful info coming out of that if anything, although this seem like a partclone issue, more info is always nice.

[Oleg]

@Sebastian-Roth
no problem - change the topic in something more related to the discussed issue.

today I tried to clone that image with another fstab-mount-options. I remove discard because I read that this option should not be used with BTRFS and also tried with clear_cache and nospace_cache but with no success.

[Sebastian Roth]

@Quazz said in Upload image "random: nonblocking pool is initialized":

Have you tried a fsck? There might be some useful info coming out of that if anything, although this seem like a partclone issue, more info is always nice.

There are about a dozen or so of btrfs-tools (like btrfs, btrfs-find-root, btrfs-debug-tree and some more) to examine and fix those kind of filesystems. Unfortunately I haven’t played with those tools before and don’t really know enough about btrfs to find out what is causing this BTRFS: open_ctree failed.

[Tom Elliott]

@Oleg what if you use those added arguments to the host kernel args?

[Sebastian Roth]

@Oleg

Hi,

Yes, Thomas Tsai is working on that.
BTRFS code base is under heavy development, so it changes very fast. Of
course we have to catch that.
Once Thomas has improved that, we will release another release.
Thanks.

Steven

On 6/17/2016 PM 03:48, sebastian.roth wrote:

Hi Steven,

I am wondering about if you have heard about any known issues with
partclone (0.2.88) and BTRFS?

After cloning a fresh installed Ubuntu system with btrfs the root
partition seems crooked.

https://forums.fogproject.org/uploads/files/1465402138541-img_20160608_175505.jpg

More Information:
https://forums.fogproject.org/topic/7700/btrfs-open_ctree-failed-after-ubuntu-image-deploy/6?page=2

Thanks in advance!

Cheers,
Sebastian

[Oleg]

@Sebastian-Roth
ok. That means we need to wait now until it’s solved in partclone.

Thanks for the information!

[Sebastian Roth]

@Oleg said:

ok. That means we need to wait now until it’s solved in partclone.

Yes. As soon as there is a new partclone version out (please inform us if you hear about it first) Tom can include it into the init’s and hopefully we should be fine then. But I don’t know when that will be.