Malformed month names in syslog messages cause logengine crash

[Morten Brekkevold]

A customer keeps getting these error message from the logengine in their cronmail:

ERROR:logengine:Unhandled exception during message parse: pr 18 05:12:59.716 CEST: %SISF-6-ENTRY_CHANGED: Entry changed A=FE80::10F1:F7E9:6EDF:2129 V=204 I=Gi0/8 P=0005 M=
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/nav/logengine.py", line 344, in parse_and_insert
    message = create_message(line, database)
  File "/usr/lib/python2.7/dist-packages/nav/logengine.py", line 154, in create_message
    timestamp = datetime.datetime(year, month, day, hour, minute, second)
TypeError: an integer is required

Presumably, the start of this message has been truncated by the previous run of the logengine (which is bad by itself, but has to be expected due to the braindead way logengine works).

The message is basically unparseable, and can easily be thrown away. But instead, it crashes the logengine,, which potentially prevents further processing of log lines.

[Morten Brekkevold (mbrekkevold)]

fix here: https://nav.uninett.no/hg/stable/rev/175bd2b89bf2