2:08 a.m.
Hi,
I have a ovirt "cluster" with one manamgement node and a single
virtualization node.
After fiddling with the database manually, I had to restore a database
backup. The database backup was from a time when the host hat 7 virtual
machines - it now has 9.
So there are now two machines "missing" from the web-interface. They're
still running on the node itself but ovirt-engine keeps logging:
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo] (DefaultQuartzScheduler_Worker-8)
VDS::UpdateVmRunTimeInfo Error: found VM on a VDS that is not in the database!
Is there a way to "re-import" the machine into the database so it can
be managed from the web-interface again?
--
sven === jabber/xmpp: sven(a)lankes.net
2:30 a.m.
----- Original Message -----
From: "Sven Lankes" <sven(a)lank.es>
To: users(a)ovirt.org
Sent: Tuesday, June 4, 2013 10:08:14 AM
Subject: [Users] I accidentally the whole database. Or: Can I re-import a running vm into
the engine-db?
Hi,
I have a ovirt "cluster" with one manamgement node and a single
virtualization node.
After fiddling with the database manually, I had to restore a database
backup. The database backup was from a time when the host hat 7 virtual
machines - it now has 9.
So there are now two machines "missing" from the web-interface. They're
still running on the node itself but ovirt-engine keeps logging:
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo]
(DefaultQuartzScheduler_Worker-8) VDS::UpdateVmRunTimeInfo Error: found VM
on a VDS that is not in the database!
Is there a way to "re-import" the machine into the database so it can
be managed from the web-interface again?
well im not sure if there is a tool that does it automatically,
but you can try the following (never tried this before, but should work):
on the host that run the vms:
* run vdsClient 0 getAllVmStats (or vdsClient -s 0 getAllVmStats if you are using secure
connection)
* find in the result the vm id (first line) and the image id for each disk of the vms
* find under /rhev/data-center/{dc-id}/mastersd/master/vms/{vm-d}/ the ovf file which
contains the vm info
* find the vm disk images on the storage domains
* use ovirt-image-uploader to upload the vm to an export domain (you need to look how
exactly to do this one.. :) )
* import the vm.
--
sven === jabber/xmpp: sven(a)lankes.net
_______________________________________________
Users mailing list
Users(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/users
2:47 a.m.
On Tue, 4 Jun 2013 03:30:42 -0400 (EDT)
Omer Frenkel <ofrenkel(a)redhat.com> wrote:
well im not sure if there is a tool that does it automatically,
but you can try the following (never tried this before, but should work):
on the host that run the vms:
* run vdsClient 0 getAllVmStats (or vdsClient -s 0 getAllVmStats if you are using secure
connection)
* find in the result the vm id (first line) and the image id for each disk of the vms
* find under /rhev/data-center/{dc-id}/mastersd/master/vms/{vm-d}/ the ovf file which
contains the vm info
* find the vm disk images on the storage domains
* use ovirt-image-uploader to upload the vm to an export domain (you need to look how
exactly to do this one.. :) )
* import the vm.
UUIDs can be discovered from qemu-kvm process args too.
jbelka
3987
days inactive
3987
days old
2 comments
3 participants
participants (3)
-
Jiri Belka -
Omer Frenkel -
Sven Lankes