Dear colleagues, Today is the 3rd day I'm fighting this issue, so I any hints will be greatly appreciated. :)
I'm deploying ACS-4.2.1 on a small setup of 9 servers: 1 manager, 2 storages accessible via iSCSI and 6 hosts. All servers are running CentOS-6.5 with all updates. Primary storages are accessible via multipathed iSCSI, the filesystem is GFS2. The secondary storage is accessible via NFS and it works okay (at least I can download an ISO-image and boot from it, so I assume it works). Everything is okay, but when I'm trying to take a snapshot of a volume, I'm getting an error message after several munutes of "Snapshotting". And, according to tcpdump between iSCSI-initiator and iSCSI-target, it really does it. :) The "Snapshotting" phase goes okay, I even can see that snapshot has been done (by "qemu-img info" or by "virsh snapshot-list"), but there is no "Backing Up" phase after that. When it's supposed to copy the snapshot to the secondary storage, it fails. That's what I see in cloudstack's logfile, when it happens: http://pastebin.com/vfAVHm0B. Okay, I'm going to the relevant host and looking in the libvirtd's logfile. As it's pretty big, I'm pasting only contents between successful mounting and unmounting of the secondary storage: http://pastebin.com/0FQKFKv9. And what I really can't get is where are any error messages? I don't see any attempts to run "qemu-img convert" that is supposed to be runned, I don't see any attempts to do anything between "mount" and "umount". Maybe I should watch more careful? When I'm mounting it manually (as root) and running "qemu-img convert" (as root), it works fine and I can see the converted QCOW2-image on my secondary storage after that. So I can't understand why cloudstack can't do it and why I can't see any attempts to do it. Please, help to find the root of the problem. Thank you so much! -- With best regards, wishes and vibes :) Vladimir Melnik P.S. I also have a small lab with ACS-4.2.0, there are no issues with snapshots in the lab.