Re: [hpoj-devel] lots of ptal-mlcd fixes

[<pa...@rc...>]

Obi-Wan wrote:
> With the stock syslog, splitting into different files is done based
> on the facility and the log level, as specified in your /etc/syslog.conf
> file.
> 
> I guess I feel like I'd rather have all my printer-related messages
> going to the same spot.  LOG_DAEMON is more of a generic catch-all
> facility for use by miscellaneous daemons that don't have their own
> pre-defined facility and don't want to use one of the eight local
> facilities.  Basically, everything except mail, printing, uucp,
> usenet, and cron.  Lprng & its cousins all log to LOG_LPR, so it
> makes sense that the low-level printer drivers log there as well.

Your syslog.conf must split things out more than mine does, because LOG_LPR
for me still ends up in /var/log/messages, which is fine with me.  I changed
the facility setting to LOG_LPR for both daemons, and added a few syslog-only
messages that display when the daemons initialize and when ptal-mlcd
activates.  For ptal-mlcd, I also joined the two lines into one for syslog,
to prevent log messages from separate instances from getting interleaved.
Hopefully everything is to your liking now.  I'm not particularly fond of
having to join the lines, but it needed to be done.

Joe Piolunek wrote:
> This appeared in /var/log/messages, I don't seem to recall any failures 
> occurring at the time, though. Unfortunately, I don't recall exactly what
I 
> was doing that led to the errors being logged. I tried to recreate the
errors 
> with no luck.
> 
> Jun 12 19:05:05 tumbleweed ptal-mlcd:  ptal-mlcd: ERROR at ExMgr.cpp:708, 
> dev=<par:0>, pid=1185, errno=11
> Jun 12 19:05:05 tumbleweed ptal-mlcd: fdRegister(invalid fd=-1,r=1,w=0) 
> called from ExMgr.cpp:2056!
I haven't yet determined with 100% certainty why this happens, but I added
an extra error message after this particular failed call that prints out
some interesting variables.  I somehow managed to hit this on the first
try after adding the extra error message, but only once.  In my case, the
extra information was "exContext=4, scd=8, state=5".  Let me know if you
or anybody else sees this again, especially if becomes reproducible.  I'll
try to get to root cause instead of applying a superficial fix that makes
the error go away but doesn't actually fix the underlying problem.

David