XE -Multiple queues in Unix


Hi guys,

We have a problem setting up multiple queues in Unix with XE. We amended the RunOneWorld.sh and EndOneWorld.sh to include the new queues. The queues start, then when we run UBEs it seems that they get stuck and we have to end the queues manually (EndOneWorld.sh does not kill them). We have the correct number of processes in the server ini file. When we restart the services it works fine for a while, then suddenly they start playing up again and we have to kill them.
Can anyone help please?


Here are a couple of things I have seen in the past on our system.

1. If the queues go to the top of your top listing (i.e you see them hitting
high 90's for CPU usage)this has generally meant that a particular spec
file has gotten blown. Generally one of the DD's but could be any one.

2. If the queues just hang or possibly die I have found that our DB wasn't
"totally" up before I brought up One World.

3. If the queues seem to have sporadic issues I have also upped some of the
parameters in the JDE.INI file on the server to compensate for system
resource issues.

4. If all else has failed (and it has) I have restored the system directory
from my last good day and then done a server refresh (i.e tamftp - We are
still on B733 base.)

We are currently upgrading to XE and only have 2 or 3 queues but since I
need to test all of my special queue maintenance programs I plan to add
about 30 queues today and run some tests. If I have any similar weirdness
and find any answers to said weirdness I'll let you know. Hope this helps

Mark Siebenschuh
HP9000/Oracle 8.0.5/B733 Base (XE on the way)
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

Content-Type: text/plain;

We have a similar problem, although we are on 7332. Try running the
EndOneWorld.sh script with 'now' after it, that should kill your queues
quickly. As for what's causing our problem here with it hanging (or yours
there), we have a few ideas, but nothing absolute yet. I would like to hear
if anyone else has encountered this before.


Oracle 7.3
Unix 11.0