[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20021125: pqing and pqact not writing to ldmd.log



Jeff,

I'm one of the LDM experts at Unidata.  Unfortunately, I'm also on
vacation until next week -- so my reponse will be brief.

> To: address@hidden
> cc: Scott Jacobs <address@hidden>,
> cc: Mark Shirey <address@hidden>
> From: Jeff Ator <address@hidden>
> Subject: pqing and pqact not writing to ldmd.log
> Organization: NOAA/NWS/NCEP

The above message contained the following:

> I work for the National Weather Service, and we are attempting to run
> LDM 5.1.4 on our new AIX 5.1 mainframe.  We compiled using CFLAGS=-q64
> (64 bit objects), and we also had to add the line:
> 
> #define _XOPEN_SOURCE_EXTENDED 1
> 
> to ulog.c in order to get it to compile.  We had to do this because,
> under AIX 5.1 and the inherent xlc C compiler, the directive:
> 
> #define _POSIX_SOURCE
> 
> which exists within ulog.c causes certain definitions that are
> subsequently required within <syslog.h> to not be initially defined
> within <stdio.h>.  This occurs because the syslog package is part of
> XOPEN, but XOPEN is not part of the POSIX standard (in fact, it is an
> extension to POSIX).  I don't believe this source change is causing the
> problem below, but I figured I should mention it just in case.
> 
> Anyway, the problem we are having is that, even though we are giving the
> -v option to pqing and pqact, we cannot get either of them to write to
> ldmd.log.  They are running OK otherwise at ingesting into and reading
> data from the LDM queue, but they just aren't logging to ldmd.log.  We
> have /etc/syslog.conf set up properly to direct local0.debug output to
> /usr/ldm/logs/ldmd.log (where /usr/ldm is actually a symbolic link to
> the LDM home account), and, in fact, if we manually run the following
> command from $LDMHOME/src/ulog:
> 
> echo jefftest | ./ulogger  -i -t test -p local0.debug
> 
> we do indeed get the following output within the ldmd.log file:
> 
> Nov 25 21:20:33 loopback test[38310]: jefftest
> 
> so it certainly appears as though everything is set up correctly and is
> running through the syslogd daemon.  However, again, pqing and pqact
> output is not showing up in ldmd.log!?

Very odd.  If the above test works and the log file exists, then the
ulog(3) facility of the LDM should also work.  What about the LDM
itself?  Are you running it?  Does it log correctly?

I have seen this behavior before, but only if the log file doesn't
exist.

> Any ideas?  We do realize that we could override the use of syslogd by
> specifying the -l option to pqing and pqact, but then we wouldn't be
> able to use hupsyslog within "ldmadmin newlog" to rotate these LDM logs,
> which I'm assuming means that we would then have to figure out some
> other way to rotate the LDM logs on our own.

I'm copying this to someone at the office who has experience in these
matters (and more resources that I at the moment :-).  Hopefully, you'll
hear from them soon.

> I sincerely appreciate any help that you can provide to us, and I thank
> you for your consideration!

You're most welcome.  That's why we're here.

> Best regards,
> -Jeff
> 
> --
> __________________________________
> 
> Jeffrey B. Ator
> 
> National Centers for
>    Environmental Prediction
> Central Operations
> 5200 Auth Road, Rm. 307
> Camp Springs, MD  20746
> 
> Phone:  (301) 763-8000, Ext. 7104
> Fax:    (301) 763-8381
> 
> E-Mail:  address@hidden

Regards,
Steve Emmerson