Error starting rosterd

[Kevin Brooks]

This is a clean install of CentOS 6.2, python 2.6.6, and roster 0.16.
When I try starting rosterd it fails with the following error:

---------------------------------------------
[root@cns1 RosterServer-0.16]# /etc/init.d/rosterd start
Starting Roster Server...
[root@cns1 RosterServer-0.16]# Exception in thread Thread-1:
Traceback (most recent call last):
File "/usr/lib64/python2.6/threading.py", line 532, in
__bootstrap_inner
self.run()
File "/usr/lib/python2.6/site-packages/RosterServer-0.16-py2.6.egg/
EGG-INFO/scripts/rosterd", line 76, in run
self.cert_file, unit_test=self.unit_test)
File "/usr/lib/python2.6/site-packages/RosterServer-0.16-py2.6.egg/
roster_server/server.py", line 127, in __init__
self.LogMessage('Roster server started on port %s' % self.port,
'rosterd')
File "/usr/lib/python2.6/site-packages/RosterServer-0.16-py2.6.egg/
roster_server/server.py", line 188, in LogMessage
log_file_handle.close()
UnboundLocalError: local variable 'log_file_handle' referenced before
assignment
---------------------------------------------

Any help on how to debug the problem would be appreciated.

[Stephen Harrell]

Hi Kevin,
I had thought we had fixed this bug but apparently not. This means
that the place where the log file should exist (where you set it in
the config file) can't be opened. For instance the user you are
running the daemon under doesn't have permissions or the directory
doesn't exist.

Let us know if this doesn't fix your problem. I will open a new bug
about this.

Thanks
-stephen

[Stephen Harrell]

Upon further investigation we have fixed this in trunk and this fix
will be included in the next release. The fix will allow you to see
the actual error instead of the error you received which doesn't
really say anything useful.

-stephen

On May 18, 10:29 pm, Stephen Harrell <stephen.lien.harr...@gmail.com>
wrote: