[Queue-developers] Interesting behaviour with queue-development

[Kai Harrekilde-P. <kh...@ex...>]

I've noticed lately that after restarting queued, I can submit
very few jobs before queued claims that someone else is running,
or the cookiefile is wrong.

This is the output from a simple test; I deleted the cookiefile,
restarted all daemons (on 5 machines, named cluster00-04):

cluster02:~$ queuedir/sbin/queued -D
cluster02:~$ queuedir/bin/queue -w -n -i -- hostname
cluster00
cluster02:~$ queuedir/bin/queue -w -n -i -- hostname
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
cluster03
cluster02:~$ queuedir/bin/queue -w -n -i -- hostname
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
cluster04
cluster02:~$ queuedir/bin/queue -w -n -i -- hostname
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
cluster01
        cluster02:~$=20
cluster02:~$ queuedir/bin/queue -w -n -i -- hostname
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Cookiefile authentication with server failed! Someone else is running
Queue on this cluster or the other side has the wrong cookiefile!
Queue: Failed to submit job to queue "now".
cluster02:~$=20

As ever, I compile/run with NO_ROOT.

The interesting part here is that each of the "remote" machines
(cluster00, cluster03, cluster04, cluster01) are able to execute
a command once, while the "local" machine never can/does.

Regards,


Kai
--=20
Kai Harrekilde-Petersen <kh...@ex...> Exbit Technology A/S