20010220: ldmfail not killing McIDAS XCD processes cleanly (cont.)

>From: 10 <address@hidden>
>Organization: NMSU/NSBF
>Keywords: 200101090229.f092TWo04390 LDM ldmfail

Robert,

>Okay, thanks Tom.  I am still a little confused..ldmfail works
>about 70% of the time, the other 30% it fails over, and the GEMPAK
>processes start fine, but the McIDAS processes fail to write any data
>and the culprit is the hanging inge processes it seems.

I can't explain this. I guess it would depend on where you put things like
xcd_run (if they are in the PATH created by the cron invocation then
XCD restart should always work).

>Since
>xcd_run is referenced in pqact.conf, does it matter that it can
>be found in the path?

Yes, it should matter unless you were explicit about where to find xcd_run
in both the ldmd.conf exec line that runs it and in the pqact.conf lines
that start the XCD ingesters.  For example:

exec    "xcd_run MONITOR"

will rely on PATH to find xcd_run

exec    "/usr/local/ldm/decoders/xcd_run MONITOR"

won't need PATH.

You might think that the answer is as simple as prefixing xcd_run with the
directory in which it is found, but if PATH does not contain this directory,
then the ROUTE PostProcess BATCH files will not run since batch.k will not
get found.

Tom