[ovirt-users] Adding another host to my cluster

[Nir Soffer]

On Fri, May 13, 2016 at 3:53 PM, Charles Tassell <ctassell at gmail.com> wrote:
> Hi Gervais,
>
>   Okay, I see two problems: there are some leftover direcyories causing
> issues and for some reason VDSM seems to be trying to bind to a port
> something is already running on (probably an older version of VDSM.)  Try
> removing the duplicate dirs (rmdir
> /var/run/vdsm/storage/248f46f0-d793-4581-9810-c9d965e2f286 and
> /rhev/data-center/mnt - if they aren't empty don't rm -rf them because they
> might be mounted from your production servers.  Just mv -i them to /root or
> somewhere.)

You should not touch directories under /run/vdsm/storage or /rhev/data-center,
they should not have any negative effect on the system.

>   Next shutdown the vdsm service with "service vdsm stop" (I think, might be
> service stop vdsm, I don't use CentOS much) and kill any running vdsm
> processes (ps ax |grep vdsm)  The error that I saw was:
>
> MainThread::ERROR::2016-05-13 08:58:38,262::clientIF::128::vds::(__init__)
> failed to init clientIF, shutting down storage dispatcher
> MainThread::ERROR::2016-05-13 08:58:38,289::vdsm::171::vds::(run) Exception
> raised
> Traceback (most recent call last):
>   File "/usr/share/vdsm/vdsm", line 169, in run
>     serve_clients(log)
>   File "/usr/share/vdsm/vdsm", line 102, in serve_clients
>     cif = clientIF.getInstance(irs, log, scheduler)
>   File "/usr/share/vdsm/clientIF.py", line 193, in getInstance
>     cls._instance = clientIF(irs, log, scheduler)
>   File "/usr/share/vdsm/clientIF.py", line 123, in __init__
>     self._createAcceptor(host, port)
>   File "/usr/share/vdsm/clientIF.py", line 201, in _createAcceptor
>     port, sslctx)
>   File "/usr/share/vdsm/protocoldetector.py", line 170, in __init__
>     sock = _create_socket(host, port)
>   File "/usr/share/vdsm/protocoldetector.py", line 40, in _create_socket
>     server_socket.bind(addr[0][4])
>   File "/usr/lib64/python2.7/socket.py", line 224, in meth
>     return getattr(self._sock,name)(*args)
> error: [Errno 98] Address already in use

Please open a bug - this should never happen.
https://bugzilla.redhat.com/enter_bug.cgi?product=vdsm
oVirt team: Infra
Severity: High

Vdsm must set the socket option socket.SO_REUSEADDR
before binding.

Ensuring that only one vdsm instance is running must be
done elsewhere if needed, for example using a lock file.

Nir