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

20030130: ldm issues?

>From: William C Klein <address@hidden>
>Organization: Valparaiso
>Keywords: 200301301738.h0UHci615298 LDM McIDAS-XCD

Bill,

>Here's a part of our ldmd.log from this morning:
>
>Jan 30 17:35:22 aeolus pqact[25917]: child 11471 exited with status 1
>Jan 30 17:35:23 aeolus pqact[25917]: child 11473 exited with status 1
>Jan 30 17:35:24 aeolus pqact[25917]: pbuf_flush (4) write: Broken pipe
>Jan 30 17:35:24 aeolus pqact[25917]: pipe_dbufput: xcd_runDDS write error
>Jan 30 17:35:24 aeolus pqact[25917]: pipe_prodput: trying again
>Jan 30 17:35:24 aeolus pqact[25917]: child 11475 exited with status 1

This says that an action in ~ldm/etc/pqact.conf is trying a product
to run 'xcd_run DDS' and the write is failing.

>[ aeolus : ldm : ~/logs ]
>[ 42 ] > ps -u ldm
>   PID TTY      TIME CMD
> 11397 ?        0:00 sh
> 11398 ?        0:00 ldmadmin
> 25916 ?        0:00 rpc.ldmd
> 25919 ?        0:03 pqact
> 25917 ?        0:12 pqact
> 25921 ?        0:03 rtstats
> 25920 ?        0:08 pqbinsta
> 25922 ?        0:02 rpc.ldmd
> 25930 ?        0:01 rpc.ldmd
> 25925 ?        0:04 rpc.ldmd
> 11400 ?        1:05 perl
> 22925 pts/5    0:00 tcsh
> 25931 ?        0:00 rpc.ldmd

This listing shows that your LDM is not running the McIDAS-XCD stuff
for some reason.

The line in ~ldm/etc/ldmd.conf that starts up the McIDAS-XCD data monitor
supervisor is:

exec    "xcd_run MONITOR"

For this entry to run, 'xcd_run' will have to be findable in the PATH
of the user running the LDM.  So the first troubleshooting step is
to:

<login as 'ldm'>
which xcd_run

If 'xcd_run' can't be found, then the PATH for the user running the LDM
is incorrect.  Since this used to run, the PATH would have had to been
changed.

If this is OK, then I would try shutting down and restarting the LDM:

<still as 'ldm'>
ldmadmin stop
<wait for all LDM processes to exit>
ldmadmin start

At this point, redo your 'ps -u ldm' and see if you see McIDAS-XCD
processes running.  You should see 'startxcd.k' and one entry for each
data monitor that you have configured McIDAS to run.  This will include
a subset of:  dmsfc.k, dmraob.k, dmmisc.k, dmsyn.k, and dmgrid.k.

>And it looks like I'm getting stuff through:
>
>[ aeolus : ldm : ~/logs ]
>[ 43 ] > ldmadmin watch
>(Type ^D or ^C when finished)
>Jan 30 17:37:55 pqutil:      105 20030130163750.960 IDS|DDPLUS 29358624
>SAUS80 KWBC 301628
>Jan 30 17:37:55 pqutil:      117 20030130163750.961 IDS|DDPLUS 29358625
>SAMX54 MMMD 301626
>Jan 30 17:37:56 pqutil:      152 20030130163750.962 IDS|DDPLUS 29358626
>SPCN43 CWAO 301627
>Jan 30 17:37:56 pqutil:      972 20030130163752.091 IDS|DDPLUS 29358665
>UBUS01 KNKA 301629
>Jan 30 17:37:57 pqutil:      143 20030130163752.112 IDS|DDPLUS 29358669
>SNVD17 CWTO 301600
>
>Looks like old stuff in /var/data as well.

The LDM is running, but the McIDAS-XCD data decoding processes are not.

Tom