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

Re: today/tomorrow

Sitler Jeffrey L Civ AFIT/ENP wrote:
> 
> Hi Anne,
> I noticed you sent me an email at 630 our time. Our system was down for work
> from 6-7 so I am guessing that was the problem. I just tried to log in as
> ldm, and I had no problems.
> Hope you had a good evening.
> Thanks
> Jeff
> 

Hi Jeff,

Yes, that must've been it.

I think logging is straightened out now.  The ldm is logging to
~ldm/logs/ldmd.log and those logs are being rotated.  Also, syslogd is
logging to /var/adm/messages and ldm messages are not going there.  That
log has not yet been rotated.  If it gets too big we'll need to look
into it's rotation.  The file /var/tmp/wsconAAAp.aG3b:0.0 is still
growing, but ldm messages are not being put there.  In looking at the
head of that file, it looks like a log from lmgrd, a license manager
deamon.  That file is bloated in large part because of ldm messages that
are now moot.  It's possibile that you could simply remove the file, and
it will be restarted when the deamon runs again.  (I think that deamon
runs when someone starts up an application that requires a license, but
I know very little about it.  I'll defer to your own sys admin for that
one.)

> I did leave things in the data directory/file for a week.  I just cleaned
> that out yesterday.
> 7 days of data will sit without bringing the drive to 100%. So once we get
> to scour we are going to hold for 7 days then delete.
> 

Yes, space seems fine, judging from the past week or so.

It looks like you went through a period of bad connectivity from 1:40Z
this AM to 9:40Z - see the logs.  You lost some products that were too
old by just a minute.  If this gets to be a problem, we can increase the
default latency value.

Also, it looks like for a while your name server was failing - the ldm
couldn't contact striker.atmos.albany.edu, but that seems to be resolved
now.

I'm not seeing any more errors from dcgrib - I wonder what changed?  In
fact, the logs from the decoders are looking pretty good for the last 24
hours.

Also, it is worthwhile to check ldm's mail periodically, if not
regularly.  That's where cron will report problems. If you look there
now, you'll see the problem regarding the mailhost.  

Do you want to turn off stats until the mailhost problem is fixed?  You
don't have to.  One benefit of having the stats files is that you can
see information about the data your getting and its timeliness.

I also meant to tell you that I modified the 'path' variable in ldm's
.cshrc script.  (The old script is still available in .cshrc.save.) 
This fixes the path to the various ldm executables. 

In general, I think your ldm is in pretty good shape at this point.  I
will stop checking it now.  Please let me know if you see any further
problems.  And, if you have problems pertaining to gempak and/or
decoders, send them to support@unidata, where they will most likely be
forwarded to Steve Chiswell. 

Anne
-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************