VM Deployment Fails with NFS drivers
|Assignee:||Javi Fontan||% Done:|
|Category:||Drivers - Auth||Estimated time:||1.00 hour|
Symptoms: oned creates the image and the context in /var/lib/one instead of the shared directory /srv/cloud/one/var/images/ after upgrading to 2.1-1
After seeing all my VM failing to boot, I started watching the log wandering under which absurd hypotesis (i.e. "The Vogon already demolished the planet" :) ) an
echo "/one/var/images/[vmid]/images/disk.0" | $SED -e 's/^[^:]*:\(.*\)$/\1/'
(core of arg_path called in tm_clone.sh) would change /srv/cloud/one/var/images/[vmid] into /var/lib/one/[vmid] when I saw a little
about in line 35... This function is defined in tm_common.sh and does that change :) !
Therefore i commented it and did the same to the
inline 41 of tm_context.sh and to the
in line 34 of tm_delete.sh
Could it be that these instructions slipped in from the SSH driver ?
#5 Updated by Javi Fontan over 9 years ago
- Target version deleted (
I'll leave it open then but without version as it is working for us in both 2.0.x and 2.2. Check the link provided before. Anyway "instead of the shared directory /srv/cloud/one/var/images/" tells me the problem is you are trying to use VM_DIR to set frontend's VM image dir to /srv/cloud/one/var and that is not its semantic.
#6 Updated by Gian Uberto Lauri over 9 years ago
O.K., I checked the #445 bug.
Mounting the NFS share as /var/lib/one is a bad solution, since you have to put the DB on the network disk, something that must be avoided.
I had the reported problem after updating to 2.1.
I will open a request for upgrade. Thank you.