Error starting instance from custom published image

I have found the issue with the automated test, it does not start v1 after increasing the disk size to 11GB. Although it does copy it to v2 and starts v2 to check the size, the original v1 has not been started with the new size, and so the cloud-init inside the focal/cloud image won’t have grown the filesystem into the larger partition.

This meant that when it was published as an image, and during unpack truncated back to 10GB, nothing but empty space was removed, and the original filesystem was untouched.

Starting v1 before publishing it also caused the same error when it came to using the image.

This should fix it:

Awesome! When would this arrive in the lxd snap? So I know to look for the fix and give it a try.

I’m hoping to do a round of cherry-pick later today so we can do one last refresh of 4.11 before we release 4.12 next week.

Sounds good. I appreciate the prompt response and triaging. Thanks!

1 Like