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

20040806: New Aeolus May Not Happen

>From: Larry Riddle <address@hidden>
>Organization: UCSD
>Keywords: 200408061512.i76FCWaW025256 LDM pqact

Hi Larry,

I tried calling you last week or the week before (can't quite remember
the date right now) after Anne asked me to field the request for help
she got from you.

When Jeff Weber and I setup the new aeolus for ingest testing, we
were working under the assumption that you would attend to getting
the data filing and decoding that you want going.  We did not, therefore,
pay any attention the pqact.conf entries in the ~ldm/etc/directory,
nor did we bother to make sure that the running of pqact in
~ldm/etc/ldmd.conf was uncommented.

After my attempted call to your cell phone, I logged onto the new
aeolus and saw that pqact was not running, so we would expect that none
of the data being received would get processed.  In addition, we saw
that the DNS setup for the new aeolus was not complete -- we have been
feeding the new aeolus in a test mode from emo.unidata.ucar.edu by
allowing feed requests by IP address, not fully qualified hostname.
We were left with the impression that you were happy to exercise
the new aeolus without worrying about transitioning from the old
one.

If you would like Jeff or I to logon to aeolus and see if data decoding
is not occurring simply by the non-running of pqact, we will be happy
to do so.  What we wanted to avoid was getting everything specified in
your pqact.conf running without you being intimately involved in the
process.  The reason for this is that we don't know your setup/intended
setup, so it would be all too easy for us to not do what you want/need.

Please let us know how you would like us to proceed!

Cheers,

Tom

>From address@hidden  Fri Aug  6 11:00:33 2004

It's the simple things that get you.  Tom Yoksas diagnosed the problem from 
what I said in my email.

The LDM was working perfectly.  It was doing exactly what I told it to 
do.  Since I neglected to remove the "#" from in front of the "exec pqact", 
the LDM knew I wanted nothing and that's what it gave me.

It is now happily writing files to disk.  Thanks Tom.

Larry

       ---===---=-=-=-=-=-=-=-=-=-=-=====[\/]=====-=-=-=-=-=-=-=-=-=-=---===---
     -----===(*  Climate's what we expect, but weather's what we 
get.  *)===-----
    Larry Riddle : Climate Research Division : Scripps Institution of 
Oceanography
        University of California, San Diego : La Jolla, California  92093-0224
        Phone: (858) 534-1869 : Fax: (858) 534-8561 : E-Mail: address@hidden 


>UCSD may have to bow out as a tier-one node.  I can't get the new "aeolus" 
>to work properly and the old aeolus is scheduled to be picked up by Surplus 
>Sales soon (possibly today).  All of my downstream nodes should switch over 
>to their alternate sites.  When, if ever, you can switch back to UCSD 
>remains to be seen.
>
>The problem with the new aeolus is that it does not write any files to 
>disk.  I was told that, when we started the ldm on the new machine, I 
>didn't have to recreate the directory trees for the files.  If they didn't 
>exist, the LDM would create them automatically as it needed them.  After 
>running smoothly for about two weeks, it has not created any directories.
>
>I went in and mkdir'd several of the missing directories.  I made sure that 
>ownership (ldm and users) and privileges (rwxr-xr-x) were the same as on 
>the old machine.  I then shutdown and restarted the LDM.  After about 
>sixteen additional hours of operation, still no joy.  No files have been 
>written to disk.
>
>The old machine is a DEC Alpha running OSF-1.  The new machine is a 
><something> running Red Hat Linux.  Part of the problem is that, while I 
>have a lot of Unix experience, this is the first Linux machine I've ever 
>had to work with.  I guess I may be clueless when it comes to Linux.  I 
>attempted to get assistance from the UPC, but was told that configuration 
>and operation was my responsibility and not theirs.
>
>My apologies if the loss of UCSD as a tier-one node causes you any 
>inconvenience.  But, as we have zero support for running the LDM (no 
>funding and no FTE since we have no undergraduate program and the data do 
>not contribute directly to our research), I can only devote a minimum of 
>time to solving this problem.  Our systems administrator has given more 
>time to this problem than he was supposed to, and has reached the limit to 
>what he can do.
>
>So, in the interest of making one last try at getting the system up and 
>running properly, can anyone give me any suggestions about what I'm doing 
>wrong?
>
>Larry
>
>
>       ---===---=-=-=-=-=-=-=-=-=-=-=====[\/]=====-=-=-=-=-=-=-=-=-=-=---===--
> -
>     -----===(*  Climate's what we expect, but weather's what we 
>get.  *)===-----
>    Larry Riddle : Climate Research Division : Scripps Institution of 
>Oceanography
>        University of California, San Diego : La Jolla, California  92093-0224
>        Phone: (858) 534-1869 : Fax: (858) 534-8561 : E-Mail: address@hidden
>
>
>
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publically available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.