Custom Query (894 matches)

/etc/bcfg2.conf is readable by root only and when importscript.py isn't ran by root or with sudo, we get this:

 python /usr/share/pycentral/bcfg2-server/site-packages/Bcfg2/Server/Reports/importscript.py 
Traceback (most recent call last):
  File "/usr/share/pycentral/bcfg2-server/site-packages/Bcfg2/Server/Reports/importscript.py", line 7, in ?
    import settings
  File "/usr/share/pycentral/bcfg2-server/site-packages/Bcfg2/Server/Reports/settings.py", line 6, in ?
    if c.get('statistics', 'web_debug') == "True":
  File "/usr/lib/python2.4/ConfigParser.py", line 511, in get
    raise NoSectionError(section)
ConfigParser.NoSectionError: No section: 'statistics'

Better error message probably can just be "/etc/bcfg2.conf not readable".

When the master runs both bcfg2-server and bcfg2 client, the IP of the client is retrieved as the public IP of the master, not 127.0.0.1 (localhost).

It works well after fixing clients.xml to use public IP for the master, but exception handling should inform the user with useful information, not just saying

Failed to download probes from bcfg2
<type 'exceptions.TypeError'>:__init__() takes exactly 3 arguments (1 given)

• Testing environment: Ubuntu 9.04 (jaunty) with bcfg2 1.0.0 from PPA.
• The syslog says:

  Jan 26 12:43:41 host0 bcfg2-server[15314]: Got request for 192.168.1.191 from incorrect address ('192.168.2.191', 47499)
  Jan 26 12:43:42 host0 bcfg2-server[15314]: Asserting client host0.local profile to slave
  Jan 26 12:43:42 host0 bcfg2-server[15314]: Failed to set client host0.local to private group slave
  Jan 26 12:43:42 host0 bcfg2-server[15314]: Client metadata resolution error for 192.168.2.191; check server log
  Jan 26 12:43:42 host0 bcfg2-server[15314]: Unexpected handler error Traceback (most recent call last):   File "/var/lib/python-support/python2.6/Bcfg2/SSLServer.py", line 46, in _marshaled_dispatch     response = self.instance._dispatch(method, params, self.funcs)   File "/var/lib/python-support/python2.6/Bcfg2/Component.py", line 227, in _dispatch     self.logger.error(e, exc_info=True)   File "/usr/lib/python2.6/logging/__init__.py", line 1056, in error     self._log(ERROR, msg, args, **kwargs)   File "/usr/lib/python2.6/logging/__init__.py", line 1143, in _log     self.handle(record)   File "/usr/lib/python2.6/logging/__init__.py", line 1153, in handle     self.callHandlers(record)   File "/usr/lib/python2.6/logging/__init__.py", line 1190, in callHandlers     hdlr.handle(record)   File "/usr/lib/python2.6/logging/__init__.py", line 669, in handle     self.emit(record)   File "/var/lib/python-support/python2.6/Bcfg2/Logger.py", line 114, in emit     newrec = copy.deepcopy(record)   File "/usr/lib/python2.6/copy.py", line
  

1. Client removes only one excess backup copy at every configfile change, even if there are more backup copies eligible for deletion. Moreover, lowering the max_copies parameter could cause client to never remove any excess files from there.

2. Space character in backup file names is mildly annoying - datetime.isoformat() will put a 'T' there instead.

3. More robust handling of 'paranoid' attribute values from info.xml (paranoid='True' should also be allowed - currently only paranoid='true' works as expected).

A simple patch (attached) addresses all these issues.

/mkd