20030123: ldmadmin pqactHUP

>From: "Benjamin Cotton" <address@hidden>
>Organization: Purdue
>Keywords: 200301231449.h0NEnwx02003 LDM FreeBSD ldmadmin pqactHUP

Ben,

>Another day, another thing that doesn't quite work right.I'm guessing
>this is again going to be a FreeBSD issue, but I'm having problems with
>'ldmadmin pqactHUP'.

This _is_ a FreeBSD issue.  The good news is that the problem has been
fixed in LDM 6, which is now in testing at the UPC.

>I thought it might be fun to try and get the POES
>data that the nws-changes list was informed of this morning, so I
>changed my pqact.conf file.  Unfortunately, when I ran 'ldmadmin
>pqactHUP' I got a message saying "pqact: process not found, can't HUP
>pqact".  So I figured why not try forcing it.  I typed '% hup
>pqact.conf' from the ~/etc directory and it told me permission was
>denied.

The HUP signal needs to get sent to the process ID of the LDM program
pqact.  The procedure to be followed is:

<login as 'ldm'>
ps -eafx | grep pqact             <- to the the PID of pqact
kill -HUP pid_of_pqact

>I'm not a UNIX expert, but I'd guess that's just because I'm
>not the root user.

pqact runs as the 'ldm' user, not root, so you don't have to be 'root'
to send it a HUP.

>The question is, then, is there any way to have LDM
>re-read pqact.conf without having to delete my queue?

You don't have to delete the queue.  If manually sending a HUP signal
to pqact does not do the trick, all you have to do is stop and then
restart the LDM:

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

Stopping and restarting the LDM does not require the deletion and
remaking of the queue.  Remaking the queue should only need to be done
if/when it gets corrupted.  The typical way the LDM queue gets
corrupted is if the machine crashes while the LDM is running, or if the
machine runs out of resources.

Tom Yoksas