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

Re: 20010201: ldm problems at UVa

Unidata Support wrote:
> 
> ------- Forwarded Message
> 
> >Subject: ldm problems (Anne and Tom)
> >Organization: UVa
> >Keywords: 200102012128.f11LSdX24812 LDM McIDAS ROUTE PP
> 
> Anne, and maybe Tom too,
> 
> There still seems to be some problems with the ldm.
> The log is once again filling up with error messages,
> and the mcidas log is showing that batch.k cannot be
> found.  I think we still need to have a correction to
> the PATH in the /usr/local/ldm/util/batch.k script,
> but maybe I am wrong.
> 
> I am also puzzled by the fact that once again the ROUTEPP.LOG
> file is owned by ldma, although Anne indicated she had
> changed it.
> 

Hello Jennie,

The comment about ROUTEPP.LOG made me realize my mistake.  Although I
set up the crontab table for ldm, I neglected to turn off the cron jobs
for ldma.  I apologize for this.  Of course, this caused significant
problems and confusion.  I have now commented out everything in ldma's
crontab.

This is why ROUTEPP.LOG got removed and recreated by ldma.  I had also
forgotten to change some paths in the ldmd.* configuration files for the
failover, which also caused some confusion.

I think this fixed the problem that appeared to be that batch.k wasn't
found.  Really, instead of batch.k not being found, I think that batch.k
wasn't finding something it needed, somehow related to ROUTEPP.LOG.  I
don't fully understand that, but it seems to be working now.


> Another weird thing, that Anne had mentioned, is that
> when I start the ldm diagnostic notifyme, there is no way
> to stop it short of killing the process?  And, for navier,
> as Anne noted, it just shows NEXRAD-related stuff.
> 

I'll need to look into both of these issues further.

> If PennState went down and then came back up, we have
> had problems in the past where we resume getting IDD/DDPLUS
> data but nothing from the McIDAS feed- I don't think that
> has ever been resolved.  But given the errors in the logs,
> it seems something else is wrong on my end anyway.
> I note that the last imagery I seem to have received
> was from 17Z yesterday, and yet the log reset (occurred
> at 18:50Z yesterday (with automatic failover, what
> initiates a log-reset? )
> 

Again, I suspect this is a problem initiated by ldmfail running out of
the ldma crontab at a time when navier was down...  This is a guess. 
Let's see what happens.

Also, Tom and I thought it would be better if mcidas.log was rotated
rather than being simply removed, so we modified ldm's crontab to rotate
2 logs.  Of course, you can change this if it's not what you want.

> Once again, I am uncertain of what to do, I could change
> the ownership of ROUTEPP.LOG to ldm, and I could add
> to the PATH in the _script_ batch.k to include
> /usr/local/ldm/bin/ldm-mcidas, but I am not sure if these
> would solve whatever is going wrong.  I am not sure how to
> see if navier is getting/sending mcidas data.
> 

I think things are ok now.  I will check again later.    

> Sorry to continue to be the source of problems, especially
> when I can see from the mail that there has been a flurry
> of questions today.
> 

Nope, you're not the source of the problem, I was.  I apologize.  I'm
embarrassed about the crontab problem...

> befuddled,
> 
> Jennie
> 
> ------- End of Forwarded Message

Hope this eases your fuddlement.

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