Red Hat Bugzilla – Bug 836263
Error message for VM remove, when host non operational, not informative enough
Last modified: 2012-12-04 15:12:13 EST
Description of problem:
For VM removal in data center with hosts which are non operational,
failure messages are not informative enough:
1) User get error event: "Failed to remove VM <name>"
Can it be more informative? Why does it fail(no up/spm host on data center), or what should be done to allow it's removal (turn hosts to up).
2) On engine log (attached), error is not so clear or informative on what the root cause is:
2012-06-27 15:37:21,946 ERROR [org.ovirt.engine.core.bll.RemoveVmCommand] (pool-3-thread-48) [6ad336a] Command org.ovirt.engi
ne.core.bll.RemoveVmCommand throw Vdc Bll exception. With error message VdcBLLException: Cannot allocate IRS server
2012-06-27 15:37:21,951 INFO [org.ovirt.engine.core.bll.RemoveVmCommand] (pool-3-thread-48) [6ad336a] Command [id=ccc6005d-5
022-48bc-be90-9e910ef29c1c]: Compensating CHANGED_STATUS_ONLY of org.ovirt.engine.core.common.businessentities.VmDynamic; sna
pshot: EntityStatusSnapshot [id=64f0a3ea-2bfa-435a-bd0b-d9e3b39786fb, status=Down].
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Add host with wrong cpu type, in order to have it non-operational
2. add (iscsi) storage domain
3. add vm (with no disk)
4. try to remove the VM
Give user information of actions he should take to be able to remove the VM,
or explain why VM can't be removed.
Created attachment 595030 [details]
engine log (to find failure search word IRS at 2012-06-27 15:32:30)
*** Bug 836261 has been marked as a duplicate of this bug. ***
The line in the engine log specifically say "Cannot allocate IRS server",
What is not clear ?
(In reply to comment #3)
> The line in the engine log specifically say "Cannot allocate IRS server",
> What is not clear ?
We are not using this term ( IRS ) anymore . Most of the messages which are using this term rephrased already .
Also, A simple user, who has only the events, see "Failed to remove VM <VM name>",
without explanation on why, or what can be done to overcome this.
The problem is that if a VM doesn't have disks, then when requesting to remove it, the canDoAction doesn't test whether the DC is valid. However, it does fail on execution.
Fix will include testing that the DC is valid even when removing VMs without disks.
Error while executing action: Cannot remove VM: Storage Domain cannot be accessed.
-Please check that at least one Host is operational and Data Center state is up.
Still not sure why we need to test DC status when VM has no disks