[Pkg-ganeti-devel] Bug#510965: Acknowledgement (ganeti: Ganeti does not work with python-twisted 8.1)

[Iustin Pop]

On Mon, Jan 26, 2009 at 02:28:51PM +0100, Daniel Schreiber wrote:
> Iustin Pop schrieb:
>
>> Attached to this mail is a python script; can you run it from any node
>> as: "python call_version.py $node"
>>
>> where $node should be any of your ganeti nodes. For me it looks like:
>> # python call_version.py localhost
>> [twisted, version 8.1.0]
>> Call 0, result: 14
>> Call 1, result: 14
>> Call 2, result: 14
>> Call 3, result: 14
>> Call 4, result: 14
>> Call 5, result: 14
>> Call 6, result: 14
>> Call 7, result: 14
>> Call 8, result: 14
>> Call 9, result: 14
>
> Unfortunately, it does not work.
>
> [twisted, version 8.1.0]
> ^CLogging system not initialized while processing message:
> caller_connect: could not connect to remote host edv05.esda.local,  
> reason [Failure instance: Traceback (failure with no frames): <class  
> 'twisted.internet.error.ConnectionLost'>: Connection to the other side  
> was lost in a non-clean fashion: Connection lost.
> ]
> caller_connect: could not connect to remote host edv05.esda.local,  
> reason [Failure instance: Traceback (failure with no frames): <class  
> 'twisted.internet.error.ConnectionLost'>: Connection to the other side  
> was lost in a non-clean fashion: Connection lost.
> ]
> Unhandled error in Deferred:
> Traceback (most recent call last):
>   File "/usr/lib/python2.5/site-packages/twisted/spread/pb.py", line  
> 1086, in clientConnectionLost
>     self._failAll(reason)
>   File "/usr/lib/python2.5/site-packages/twisted/spread/pb.py", line  
> 1073, in _failAll
>     d.errback(reason)
>   File "/usr/lib/python2.5/site-packages/twisted/internet/defer.py",  
> line 269, in errback
>     self._startRunCallbacks(fail)
>   File "/usr/lib/python2.5/site-packages/twisted/internet/defer.py",  
> line 312, in _startRunCallbacks
>     self._runCallbacks()
> --- <exception caught here> ---
>   File "/usr/lib/python2.5/site-packages/twisted/internet/defer.py",  
> line 328, in _runCallbacks
>     self.result = callback(self.result, *args, **kw)
>   File "/var/lib/python-support/python2.5/ganeti/rpc.py", line 103, in  
> cb_err1
>     self._check_end()
>   File "/var/lib/python-support/python2.5/ganeti/rpc.py", line 64, in  
> _check_end
>     reactor.stop()
>   File "/usr/lib/python2.5/site-packages/twisted/internet/base.py", line 
> 495, in stop
>     "Can't stop reactor that isn't running.")
> twisted.internet.error.ReactorNotRunning: Can't stop reactor that isn't  
> running.
> Call 0, result: False
>
> It blocks forever after printing the twisted version.

This is strange then. The breakage occurs even before printin the first result,
which means it does not work at all (so it's different from the bug that we
fixed in 1.2.5).

Just to be sure, can you confirm this also happens when trying against
localhost? And do you get anything in /var/log/ganeti/node-daemon.log?

I would be interested at this point in:
  - the output of "python -v call_version.py ...."
  - an strace of "python call_version.py ..."
  - an strace of ganeti-noded while doing the above

This is very very strange, and I still can't reproduce it, even with your
package list (but not running the xen kernel, and only 32 bit) when running the
normal ganeti-noded.

The only way to reproduce this problem is if I don't start ganeti-noded, but a
simple nc or similar, then indeed I get the same backtrace... this is why I'd
like to see the strace.

iustin, really confused.