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

Various ldm issues

Sitler Jeffrey L Civ AFIT/ENP wrote:
> 
> Hi Anne,
> That is fine, if we get things working properly, I can move it over to
> Nimbus at a later date and time, when I have more experience in such
> matters.
> Jeff
> 

Hi Jeff,

I have to backpedal.  I assumed that the ldm account was on blizzard and
was puzzled by how the queue could have been built on fujita.  Now I see
that the ldm account is remotely mounted on blizzard, and actually
exists on fujita, so the queue is indeed already local to the ldm
account.  So, I will *not* relocate the queue.  I just rlogged onto
fujita, and am in the process of looking around. 

And, adding the local0 facility to /etc/syslog.conf on blizzard was for
naught - it doesn't need to be there because the ldm is not running on
blizzard.  I see that the proper entry already exists on
/etc/syslog.conf on fujita and logging is successful.

In looking at the logs, I saw a major problem with space that may have
been resolved by now.  Still, to clean the slate, I removed all the
stats files in ~ldm/logs.  More will be generated.  See below for info
about how to keep those from piling up.

I wanted to see the log, but it was too large to view.  To handle this,
I split the log up into smaller files.  In ~ldm/logs you'll see several
smaller files called 'xaa' through 'xam'.

I just made a crontab for you.  If you do 'crontab -l' you will see the
jobs that I put in there.  Do 'man crontab' for more information about
how this works.  Of course, feel free to change these in any way that
better suits your needs.  These are intended just to get you started.

There first job listed in the crontab is to send in the statistics. 
This happens once per hour at 35 minutes after the hour.  Not only will
this allow us to see information about product latencies at your site,
it will clean up all but the most recent 24 *.stats files in ~ldm/logs,
which will be a help in saving space.

The second job is commented out.  It is to run 'scour' which will delete
old products.  I left it commented out so you could decide how long you
want to keep products and configure scour accordingly.  I'll help you
with that later.

The third job is to rotate the logs once per day, at 10 minutes after
midnight.  This will keep the log files from getting too large.

I need to leave for a meeting in a few minutes, but I wanted to let you
know what I found and what I've done so far.

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