Reset From Image issue

Submitted by gnilebein on Mon, 09/26/2011 - 09:54
  1. I tried to create an Xen from ""-Image
  2. i tried to overwrite this xen with "Debian Squeeze Xen instance with base OS"
  3. I got the following output:

Starting reset of Xen system image.cloudmin.cluster101 from image Debian Squeeze Xen instance with base OS ..

Shutting down system .. .. done

Copying 446.35 MB image file to host system node02.cluster101 ..

.. already in host cache

Re-creating filesystem from original image .. .. creation started.

Waiting for creation to complete .. .. creation has completed successfully.

Removing missing disks from fstab file .. .. cleanup failed : Failed to read fstab file : Failed to mount /xenstore/image.img on /mnt/xen-image : mount: you must specify the filesystem type

Expanding filesystem to 2 GB .. .. failed to find primary disk

Mounting new instance's filesystem .. .. failed : Failed to mount /xenstore/image.img on /mnt/xen-image : mount: you must specify the filesystem type

Configuring kernel for Xen instance .. .. setup to boot Xen system's kernel with PyGrub

Configuring Xen instance for VNC console access .. .. added on dynamic port

Starting up new Xen instance .. .. failed to start : Using config file "/xenstore/image.cfg". Error: (2, 'Invalid kernel', 'elf_xen_note_check: ERROR: Will only load images built for the generic loader or Linux images')

The "Reset From Image" works fine with the "Debian Lenny 64-bit Xen instance with base OS"-Image.

Status: 
Active

Comments

Submitted by JamieCameron on Mon, 09/26/2011 - 16:26

Have you upgraded to Cloudmin 5.8 ? Older versions had a bug that could cause this error if reseting a VM that used a single-partition image with a whole-disk image.

Submitted by gnilebein on Tue, 09/27/2011 - 06:02

Yes, we are using the newest release of cloudmin. What i have fogotten. I tried this with images shipped with cloudmin. Debian Lenny 64-bit Xen instance with base OS and Debian Squeeze Xen instance with base OS...

Submitted by JamieCameron on Tue, 09/27/2011 - 13:51

If you create a complete new VM using the image that you are resetting to, does that work?

Submitted by gnilebein on Tue, 09/27/2011 - 17:13

Yes, the reset to the inital image works...

But an reset from Debian Lenny 64-bit Xen instance with base OS to Debian Squeeze Xen instance with base OS does not work :(

Submitted by JamieCameron on Tue, 09/27/2011 - 17:41

What about creating a completely new VM using the "Debian Squeeze Xen instance" image?

Submitted by JamieCameron on Wed, 09/28/2011 - 16:04

So does not setting the maxmem parameter fix this problem?

Submitted by gnilebein on Thu, 09/29/2011 - 03:53

No, this won't fix the issue... And i tried it without a samba share. I still get this error message:

Removing missing disks from fstab file .. .. cleanup failed : Failed to read fstab file : Failed to mount /xenstore/test.img on /mnt/xen-test : mount: you must specify the filesystem type

Mounting new instance's filesystem .. .. failed : Failed to mount /xenstore/test.img on /mnt/xen-test : mount: you must specify the filesystem type

Submitted by JamieCameron on Thu, 09/29/2011 - 19:30

Can I login to your system again with the credentials you provided and try a reset or VM creation from this image?

If so, let me know which VM to reset ..

Submitted by gnilebein on Fri, 09/30/2011 - 03:40

The credentials are sill correct... Please try it with test.cloudmin.cluster101... :)

Submitted by JamieCameron on Fri, 09/30/2011 - 15:36

I see you still have that CIFS filesystem mounted as /xenstore .. so that I can eliminate one possible cause of the problem, can we try a local disk location for your Xen VMs?

Submitted by gnilebein on Tue, 10/04/2011 - 04:00

We did a switch to NFS. You can try it

Submitted by JamieCameron on Tue, 10/04/2011 - 15:51

Thanks .. doing some more testing now.

Submitted by JamieCameron on Tue, 10/04/2011 - 15:59

So I still saw the same problem even with the filesystem being on NFS.

As a further test, I instead had Cloudmin create a VM called jamietest.cloudmin.cluster101.net-lab.net with it's disk images on /tmp instead of /storage , and it worked fine!

What kind of system is exporting that filesystem? It seems to be behaving oddly..