XenAPI Resize Fails with Migration Error: Failed to transfer vhd to new host
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Won't Fix
|
Wishlist
|
Unassigned |
Bug Description
Setup:
OpenStack Grizzly and XCP on Debian Wheezy from the repositories (Kronos).
Shared storage based on NFS as the default SR configured for live migration.
When pressing the "More->Migrate Instance" on Horizon web interface, a resize is initiated (Migrate here is misleading since it's basically resize but this is another thing) which always fails with the following errors:
2013-07-16 17:49:35.684 ERROR nova.compute.
2013-07-16 17:49:41.862 ERROR nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
2013-07-16 17:49:41.862 5363 TRACE nova.openstack.
Exactly the same happens when I try to resize an instance from the command line (either up or down)
~# nova resize <image-id> m1.small --poll
Instance resizing... 40% complete
Error resizing instance
I assume it might be because of the shared storage?
Anyway it's a bug which puts the VM in ERROR state so it should be handled properly.
Changed in nova: | |
importance: | Low → Wishlist |
Changed in nova: | |
status: | Triaged → Confirmed |
Agreed the shared storage is the likely cause. The resize code (as opposed to the migrate) is specifically intended for local storage.
VM migrate (from the command line) should work in this scenario