On Sat, 16 Dec 2000, Eric W. Ribble wrote:
> We are in the process of transitioning to a new LDM machine here at Purdue.
> For the most part, things have gone well.
>
> The new server address is anvil.eas.purdue.edu. For those of you who send
> us data, would it be possible for you to continue sending data to
> rainbow.eas.purdue.edu and to our new machine 'anvil' while I get the kinks
> worked out? Please let me know if this will be problem.
Instead of requesting separate feeds from your upstream sites for rainbow
and anvil, you may want to just make upstream requests for rainbow and
then have rainbow feed anvil until you have everything running OK on
anvil. This will reduce network bandwidth consumption.
> I've noticed that we're receiving mcidas and nowrad data from our providers
> (on anvil) yet the data is not finding its way to their specified
> directories. Things are going along just fine on our old LDM server,
> rainbow.
>
> For example the log shows (on anvil):
>
> Dec 16 23:48:33 pqutil: 182044 20001216234829.379 WSI 345
> NOW/MASTER/200012162345
> Dec 16 23:48:33 pqutil: 22595 20001216234832.702 WSI 346
> NOW/USRAD/200012162345
>
> Yet when I check in the nowrad directory, nothing is there. I didn't change
> anything in the pqact.conf file when I moved it over from rainbow.
>
> ############################################################################
> # WSI feed
> #
> # Master sector
> WSI NOW/MASTER/..(.*) FILE -overwrite data/nowrad/\1.nowrad
> WSI NEX/NOW/MASTER/..(.*) FILE -overwrite data/nowrad/\1.nowrad
> #
> # US sector
> WSI NOW/USRAD/..(.*) FILE -overwrite data/nowrad/\1.usrad
> WSI NEX/NOW/USRAD/..(.*) FILE -overwrite data/nowrad/\1.usrad
>
You should check the setting of the LDMHOME and LDMDATA env. variables on
anvil and make sure the pathnames are correct for anvil, since the
pathname 'data/nowrad/\1.usrad' is relative to to LDMHOME. If I were to
hazard a guess, you may have copied the variable settings from rainbow to
anvil, but the pathnames are different on the two hosts. If necessary,
you can use absolute pathnames in your pattern-action to make sure it gets
put where you want it.
> The same goes for the mcidas feed; it's not ending up in its local directory
> on anvil's disk. Here's a snip of the mcidas block from our pqact.conf
> file.
>
> # GOES West Visible
> MCIDAS ^pnga2area Q1 U9
> PIPE -close /project/ldm/decoders/pnga2area
> /project/ldm/data/mcidas/%y%m%d%H_svw.mca
Here you would want to make sure that the directory tree
'/project/ldm/data/mcidas' exists on anvil, i.e., those directories should
already exist.
> Also, on the FOS Topology page:
>
> (http://www.unidata.ucar.edu/projects/idd/status/idd/fosTopo.html)
>
> anvil is showing up on the page as "unknown" and rainbow is no where to be
> found. Any ideas on that one? Before I brought anvil online, rainbow did
> show up and was reporting its latency.
Well, rainbow is showing up there now, also as 'unknown'. (The FOS
topology page has been a little flaky lately. It's showing the time as
16Z when it's only 15Z currently, and the branch of the tree rooted at
snow.cit.cornell.edu isn't showing up at all.) I'm not sure if anything
you did caused rainbow to stop reporting stats. To make sure stats get
sent from anvil, check to see that there is a crontab entry on anvil under
user 'ldm' similar to this (your pathname to ldmadmin will be different):
#######################
### LDM Statistics and Check Scripts
#
35 * * * * /weather/ldm/bin/ldmadmin dostats
Tom
------------------------------------------------------------------------------
Tom McDermott Email: tmcderm@xxxxxxxxxxxxxxxxxxxxx
System Administrator Phone: (716) 395-5718
Earth Sciences Dept. Fax: (716) 395-2416
SUNY College at Brockport