George,
We will give this a try.
Do you know of anyone that we could pay to help us. This is way beyond our skills and now our clients are getting really upset. I have just 8 VMs that we need to restore somehow.
Anyone you can refer we are open to talking to.
Thanks.
Juan
From: George Shuklin [mailto:george.shuklin@xxxxxxxxx]
Sent: Friday, November 09, 2012 10:28 AM
To: Juan Lorenzana
Cc: xen-api@xxxxxxxxxxxxx
Subject: Re: [Xen-API] XCP 1.5 VM images will not start
How did you saw difference between snapshots from vdi's? The only (actual) difference is xapi opinion, for VHD there is base copy and leafs, any leaf can be a disk or a snapshot.
If you done any snapshots (I think I see you done), you gonna have bunch of VHDs, some of them are 'disks', others 'snapshots', some 'base copy'. You can get your disk content if only every VHD in chain is 'in place'. If you lost any VHD file - whole chain is broken.
1st: Check /var/log/SMlog for exceptions.
2nd: shutdown every VM and be sure no stray tapdisk processes is running. I saw them sometimes been running even after VM shutdowned.
3rd: try to create new SR on separate disk/storage and copy disks (xe vdi copy).
I think traces in SMlog will say reasons about all bad stuff happens around.
On 09.11.2012 14:40, Juan Lorenzana wrote:
George,
I tried. Here is what I know. The .vhd files that show up in Xen Center all appear to be snapshots. The ones that were running at the time of the failure are no where to be seen in the storage repository.
We also can find the last snapshot which was early yesterday morning and all the snapshots include the base. Those we get to directly through dom0.
However we have no way to tell them to coalesce.
So we can not find the VDIs that were running in Xen Center. We only see what appears to be snapshots. Not sure if the base is able to be seen now that I think about it.
Any other ideas?
I was hoping someone knew how to get the current ones to show up or to help coalesce from the base to the snapshot.
Even if we do that, I am still not sure that XenCenter will show it to me. My CLI command are poor and being that we have been up for almost 24 hours, I am little tired and not thinking my best.
If you have any ideas I would appreciate it.
Thanks.
Juan
Can you access to those VDI from dom0? (/var/run/sr-mount) If no - work with nfs and dom0 settings. If 'yes' - try to create vm, add vdi to it as secondary device (device=1,2,etc) and see what will happens.
On 09.11.2012 02:33, Juan Lorenzana wrote:
Okay. So this is pretty serious and I know it. Any help is appreciated and I am willing to pay for it at this point.
We had a XCP 1.5 pool with 2 machines as hosts. The storage pool which as NFS mounted lost its ability to export the filesystem with the VM images on it.
We contacted the vendor and we were able to restore it. However, they had recommended that we delete the XCP Pool, recreate it, add the hosts back in, and then mount the storage.
That is where we are at now. We can see the storage and the VM images. However in XCP Xen Center when looking at the images is clear that the metadata is lost. So we can not see which image or snap shot belongs to which VM. The names all are blank when we look at the storage.
Also, if I randomly mount an image to a VM, it does not boot. It produces an error that says:
Traceback (most recent call last): - File â/usr/binpygrubâ, line 808, in ? â fs = fsimage.open(file,part_offs[0[, bootfsoptions) â IOError: [Errno 95] Operation not supported.
I know it is pretty bad.
Can anyone help? We have about 10 VMs for 8 clients on these two hosts.
Thanks.
I can be reached at by emailing help@xxxxxxxxxxxxxx or calling 602-738-3220.
Juan
_______________________________________________
Xen-api mailing list
Xen-api@xxxxxxxxxxxxx
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api
