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

Re: 20040920: Possible pqact issue in LDM? (cont.)



Steven,

If you are logging pqact to a file, eg "-l file", instead of using
syslog, then the newlog won't be able to rotate your output stream.

If you are logging pqact to a file to avoid sending the log to the
rpc.ldmd LOCAL0 log, then you can make pqact use another LOCAL# facility
and use newlog to rotate that independently of ldmd.log (using Local0).

For example, if you want pqact to use LOCAL4 logging, and set that up in
syslog.conf, then you can create a pqact startup script that defines
ULOG_FACILITY_OVERRIDE as 4 before starting pqact

eg:

ULOG_FACILITY_OVERRIDE=4
export ULOG_FACILITY_OVERRIDE
pqact


Steve Chiswell
Unidata User Support

On Mon, 2004-09-20 at 13:40, Steven Danz wrote:
> Tom Yoksas wrote:
> 
> >>From:  "Steven Danz" <address@hidden>
> >>Organization: Aviation Weather Center
> >>Keywords: 200409091803.i89I3pnJ023109 LDM pqact processing
> >>    
> >>
> >
> >Steven,
> >
> >re:
> >
> >  
> >
> >>True, but that is for the rpc.ldmd log, I'm looking to rotate the pqact 
> >>log.... doesn't look like that is part of the pqact right now.
> >>    
> >>
> >
> >The default log file used by pqact is ~ldm/logs/ldmd.log.  If you
> >are specifying the output log file in your pqact.conf action, then
> >you can use the LDM utility 'newlog' to rotate that log file; we
> >do this routinely.  Here is an crontab example:
> >
> ># Rotate the ldm-mcidas log files
> >15 18 * * * bin/newlog logs/ldm-mcidas.log 7
> >
> >The newlog man page will give you an explanation of the command line
> >parameters it takes.
> >  
> >
> What 'kicks' pqact into writing to the new file?  Looks like newlog is 
> just a script that renames the log
> file, but I don't see what tells pqact (or anything else) to close the 
> old file and open the new one. 
> Moving pqact.log to pqact.log.1 works, and it creates a new file 
> pqact.log, but the old file is what
> is written to.  Will I have to stop/start ldm to get it to use the new 
> log file?
> 
> >re: So, pqact(1) doesn't miss data-products if run manually on a saved
> >product-queue but does in the context of a executing LDM system.  Is
> >this true?
> >
> >  
> >
> >>This is true.
> >>    
> >>
> >
> >The problem would then seem to be related to interrupt processing.
> >Just a thought...
> >
> >  
> >
> Seems like a possibility given the nature of the problem.
> 
> >Cheers,
> >
> >Tom
> >--
> >+-----------------------------------------------------------------------------+
> >* Tom Yoksas                                             UCAR Unidata 
> >Program *
> >* (303) 497-8642 (last resort)                                  P.O. Box 
> >3000 *
> >* address@hidden                                   Boulder, CO 80307 *
> >* Unidata WWW Service                             
> >http://www.unidata.ucar.edu/*
> >+-----------------------------------------------------------------------------+
> >
> >  
> >