20030922: WSI feed to RAP LDM host(s)

>From: Celia Chen <address@hidden>
>Organization: NCAR/RAP
>Keywords: 200309221631.h8MGVak1029670

Celia,

>Currently we get the WSI feed on iita at RAP.  The iita machine
>is getting old and having some issues over the weekend. We did 
>not get any data as the result.  Could you please feed the WSI data 
>to our main LDM host chisel, in addition to iita?

chisel should be able to get the WSI feed from wsihscn.  You will have
to add the request to chisel's ~ldm/etc/ldmd.conf file and stop and
restart the LDM to start the data flowing.

>Also, do you have the WSI data archived for the weekend?

I just checked and see that we have not received any WSI data since
1245Z on September 19.  We are investigating this now.

While I have you "on the line" (so to speak), I want to ask you to
upgrade the LDM on all RAP machines to the current release, 6.0.14.

Here are the machines currently running LDM-6.x and reporting LDM-6
statistics:

clamp.rap.ucar.edu   - ldm-6.0.14

chisel.rap.ucar.edu  - ldm-6.0.10
level.rap.ucar.edu   - ldm-6.0.10
moonbow.rap.ucar.edu - ldm-6.0.13

Here are RAP machines still running LDM-5.x:

ofour.rap.ucar.edu   - ldm-5.1.4

Along with the upgrade, please have all machines report real time
statistics back to us (mostly a comment about ofour).  As a reminder,
reporting of real time stats is accomplished by adding the line:

exec    "rtstats -h rtstats.unidata.ucar.edu"

to the machine's ~ldm/etc/ldmd.conf file (as the last 'exec' line),
and then stopping and restarting the LDM.

The reason we want sites to upgrade to 6.0.14 from 6.0.10 is that the
strategy used to connect to an upstream feed site changed between the
two releases.  6.0.10 first tries to connect using LDM-6 protocols; if
this attempt times out for whatever reason, it then tries using LDM-5
protocols.  The problem comes in when an upstream feeder goes away and
then comes back (due to a network glitch, an LDM stop and restart,
etc.): 6.0.10 will connect up using LDM-5 protocols.  The data transfer
is a lot less efficient using LDM-5 protocols for both the downstream
host (you) and the upstream host (whoever you are feeding from).
Thanks.

>Thanks in advance.
>Celia

Tom