| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Xmitq and channel problems |
« View previous topic :: View next topic » |
| Author |
Message
|
| vigsai |
 Posted: Mon Dec 31, 2007 12:10 pm Post subject: Xmitq and channel problems |
 |
|
Newbie
Joined: 31 Dec 2007
Posts: 4
|
Since these queries are discussing about a full XMITQ, I thought I will ask my question in this chain.
[moderator]Ref to This discussion chain[/moderator]
Background:
We have a Queue manager on mainframe zos(QM1) talking to many other Queue managers like QM2, QM3, QM4. A Batch application on zos drops datagram messages from QM1 to a Queue(LQ1) in QM2(Unix). Applications from QM3(NT), QM4(Unix) send/receive critical(persistent) messages to QM1.
Problem:
LQ1 in QM2 reached it's maximum depth and hence QM2 stopped accepting messages from QM1. The messages that were sent after LQ1 reached it's limit were sitting in the XMITQ(QM1_TO_QM2) of QM1. It seamt like that this issue between MQ1 and MQ2 started affecting communications between QM3/QM4 and QM1 causing delay in responses and failures.
Although we fixed everything in QM2 and made everything back again, we lost a lot of messages from QM3/4 to QM1. I still don't understand why the XMITQ between QM1 and QM2 will affect communication between QM1 and other queue managers. 
Please reply.. |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Mon Dec 31, 2007 12:43 pm Post subject: Re: XMITQ Full |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| vigsai wrote: |
Since these queries are discussing about a full XMITQ, I thought I will ask my question in this chain.
Background:
We have a Queue manager on mainframe zos(QM1) talking to many other Queue managers like QM2, QM3, QM4. A Batch application on zos drops datagram messages from QM1 to a Queue(LQ1) in QM2(Unix). Applications from QM3(NT), QM4(Unix) send/receive critical(persistent) messages to QM1.
Problem:
LQ1 in QM2 reached it's maximum depth and hence QM2 stopped accepting messages from QM1. The messages that were sent after LQ1 reached it's limit were sitting in the XMITQ(QM1_TO_QM2) of QM1. It seamt like that this issue between MQ1 and MQ2 started affecting communications between QM3/QM4 and QM1 causing delay in responses and failures.
Although we fixed everything in QM2 and made everything back again, we lost a lot of messages from QM3/4 to QM1. I still don't understand why the XMITQ between QM1 and QM2 will affect communication between QM1 and other queue managers.
Please reply.. |
Well the messages to QM3 and QM4 should only be affected if you multihop through QM2. There are no messages lost. What probably happened is that you had messages expire.
On the other hand it should not have affected at all messages from QM3/QM4 to QM1. You did not specify if at any time during the interval or immediately before you have made any changes to a firewall, or suffered any hickups to the network however brief...
What is the channel status on your different qmgrs...?
Is sender/receiver channel status consistent?
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| vigsai |
| Posted: Mon Dec 31, 2007 1:00 pm Post subject: XMITQ Full |
|
|
Newbie
Joined: 31 Dec 2007
Posts: 4
|
Not all messages from QM3/QM4 to QM1 failed. The channels were coming up and working.
Initially, it seamt like a network delay as some of them were working while some were not. However, there were no network issues reported at that time. I just wanted to see if anybody sees a reason why QM1-QM2 communication will affect other Queue managers. One theory is that QM1 got busy servicing QM2 and became resource intensive/slow. The message depth in QM1 and QM2 was between 3K and 5K. The Queue max depth in QM2 is 5000. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Mon Dec 31, 2007 1:11 pm Post subject: Re: XMITQ Full |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| vigsai wrote: |
Not all messages from QM3/QM4 to QM1 failed. The channels were coming up and working.
Initially, it seamt like a network delay as some of them were working while some were not. However, there were no network issues reported at that time. I just wanted to see if anybody sees a reason why QM1-QM2 communication will affect other Queue managers. One theory is that QM1 got busy servicing QM2 and became resource intensive/slow. The message depth in QM1 and QM2 was between 3K and 5K. The Queue max depth in QM2 is 5000. |
Well a few things you can do is increase the maxqdepth (no need to recycle the qmgr).
On the other hand no reported network incident doesn't mean there wasn't any.
We had all of a sudden trouble with maxchannels and the only thing that changed had been a new firewall replacing the old one. No network incident had been reported. When we finally switched back to the old firewall everything worked as expected again.
Both old and new firewalls had exactly the same settings... go figure...
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Dec 31, 2007 3:14 pm Post subject: |
|
|
Guest
|
| Any messages in SYSLOG for this qmgr? Did the pageset where the transmit queue resides fill up? |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
