Hola Arnau,
> I've a questions about wmproxy-server daemon.
>
> When querying wmproxy status it always show 2 parts:
>
> WMProxy httpd listening on port 7443
> ===
> WMProxy Server running instances:
>
>
> In our WMS the server part is not always running and, from its status
> function, seems that server is not important for service to return ok:
Exactly: the WMProxy Server instances reflect job submission,
cancellation and output activities. No activity, no processes.
(They linger for a while.)
> [...]
>
> But service ref card says that "glite-wms-wmproxy" should be running,
> but does not make difference between part.
Right: the /opt/glite/etc/init.d/glite-wms-wmproxy status command
must return success, that is all that matters.
I normally use this script to check if the WMS daemons look OK:
----------------------------------------------------------------------
#!/bin/sh
export LANG=C
ps -u glite | sort -k 4 | uniq -f 3 | perl -pe '
BEGIN {
@daemons = (
"condor_collecto",
"condor_master",
"condor_negotiat",
"condor_schedd",
"glite-lb-interl",
"glite-lb-logd",
"glite-lb-proxy",
"glite-proxy-ren",
"glite-wms-ice",
"glite-wms-job_c",
"glite-wms-log_m",
"glite-wms-workl",
"httpd",
);
}
for my $d (@daemons) {
if (/ $d/) {
$seen{$d}++;
break;
}
}
END {
for my $d (@daemons) {
print "\nABSENT: $d\n" unless $seen{$d};
}
print "\n";
}
'
echo === JC ===
tail -3 /var/log/glite/jobcont*oller_events.log
echo === LM ===
tail -3 /var/log/glite/logmonitor_events.log
echo === WP ===
tail -3 /var/log/glite/wmproxy.log
echo === WM ===
tail -3 /var/log/glite/workload_manager_events.log
----------------------------------------------------------------------
|