I was attempting to migrate a root filesystem from NFS to Ceph. During the 
process, Cloudstack came back and told me the process took too long and it was 
canceling the job. This put the state of the filesystem in limbo as the there 
no longer the NFS filesystem and the Ceph image is stuck in “Creating”.

I was able to export the image from Ceph using "rbd export --pool=cloudstack 
31c8d8d5-9dde-4512-ab1e-dcce8dbaf6f3 rootfs.img”.

I’m able to mount the /boot filesystem on this image using proper offsets, so 
this gives me the indication that this image is probably healthy, but how do I 
get this back in to Cloudstack and how do I tell the VM to use this new image 
for its root filesystem? This image has an LVM partition and it needs to boot 
in conjunction with the additional storage I provisioned for this instance 
which makes up the lvm volume being used.

I have the image, I just need to get it back in to cloudstack and I need the 
instance config to use this new image. Or may be even simpler, how to I 
re-establish the relationship with the image which now exists in Ceph but has 
no mapping within Cloudstack since the job bailed in the middle?

This is Cloudstack 4.16.0.0 and Ceph Pacific 16.2.4.

Thanks
-jeremy

Attachment: signature.asc
Description: PGP signature

Reply via email to