Why different logging destinations
Hello all!
Just a thought: what's the reason for mlmmj executables using both syslog (via the log_error() function) and a file in the list's spool directory (mlmmj.operation.log, via the log_oper() function)? The latter method also uses hardcoded log rotation logic. Also mlmmj-maintd logs to a still different file, with its own rotation logic.
On top of that, some log_error()s look like debugs. I "love" this one particularly:
Nov 2 18:15:02 hostname /usr/bin/mlmmj-process[16632]: mlmmj-process.c:839: Found Cc: [email protected]: Success
IMHO it would make more sense to log everything to syslog (falling back to a file if syslog is not available) with different priorities, so admins could apply filtering, routing and rotation to their tastes.
Currently I have to play tricks with syslog-ng to get all info into one place, and it fails sometimes.
Greetings,
P.
Signature · 1 line
Piotr "Malgond" Auksztulewicz [email protected]