[ovirt-devel] VDSM crashed while adding host to newly created cluster

[Christopher Pereira]

On 26-03-2015 18:16, Christopher Pereira wrote:
> Now, how can I fix the "invalid lockspace"?
hosted-engine start-pool + connect-storage + vm-start solved the invalid 
lockspace problem and made it possible to restart the engine VM, but the 
hosted-engine-setup is unable to resume:

    2015-03-26 21:56:42 DEBUG
    otopi.plugins.ovirt_hosted_engine_setup.engine.add_host
    add_host._wait_host_ready:189 VDSM host in non_responsive state
    2015-03-26 21:56:44 DEBUG
    otopi.plugins.ovirt_hosted_engine_setup.engine.add_host
    add_host._wait_host_ready:189 VDSM host in non_responsive state
    2015-03-26 21:56:46 DEBUG
    otopi.plugins.ovirt_hosted_engine_setup.engine.add_host
    add_host._wait_host_ready:189 VDSM host in non_responsive state
    [...]

Then, inside the web-manager, I tried to activate the host manually and 
it seems like the engine stops VDSM:

    Mar 26 22:01:10 h2 systemd: Stopping Virtual Desktop Server Manager...
    Mar 26 22:01:10 h2 journal: vdsm IOProcessClient ERROR IOProcess failure
    Traceback (most recent call last):
       File "/usr/lib/python2.7/site-packages/ioprocess/__init__.py",
    line 107, in _communicate
         raise Exception("FD closed")
    Exception: FD closed

And the engine-storage disappeared again.
Then, since QEMU doesn't reopen the (invalid) file descriptors, the VM 
can't be resumed (BZ 1058300).

Is it normal that 1) the engine stops VDSM during activation and 2) that 
VDSM stops or restarts the storage during shutdown?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/devel/attachments/20150326/811a4a2f/attachment-0001.html>