Author |
Message
|
nak1984 |
Posted: Mon Dec 12, 2011 10:00 pm Post subject: Channel sequence number issue in MQ v7 |
|
|
Newbie
Joined: 11 Dec 2011 Posts: 5
|
We are facing issues with one of our connection over MQ (ver. 7). The channel fails to remember sequence number after any restart event (at receiver end).
We have tried several fixes but no luck. Every time we have to reset the channel sequence number manually.
Any help in this regard would be highly appreciated. |
|
Back to top |
|
 |
bruce2359 |
Posted: Mon Dec 12, 2011 10:07 pm Post subject: Re: Channel sequence number issue in MQ v7 |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
nak1984 wrote: |
We have tried several fixes but no luck. Every time we have to reset the channel sequence number manually. |
Exactly what have you tried? _________________ I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
Back to top |
|
 |
Mr Butcher |
Posted: Mon Dec 12, 2011 11:10 pm Post subject: |
|
|
 Padawan
Joined: 23 May 2005 Posts: 1716
|
Also interesting - what are the channel numbers at which end when the out of sequence - condition occurs? one end is at 1, the other at xxx? if so, check the end that has "1" as sequence number. maybe there is something in the shutdown or startup procedure (you said it happens after restart) that is resetting channel, or that delete / defines mq objects , or uses define replace. ..... _________________ Regards, Butcher |
|
Back to top |
|
 |
nak1984 |
Posted: Tue Dec 13, 2011 1:55 am Post subject: |
|
|
Newbie
Joined: 11 Dec 2011 Posts: 5
|
Let me explain in bit more detail:
I’m facing problem at the receiver end of the channel always. Problem occurs in case, for example, n/w connection is down then the channel terminates. Now when the channel is established after a while, then it complains sequence number is not matching:
i.e. AMQ9526: Message sequence number error for channel ‘xx.xx.xx'.
EXPLANATION:
The local and remote queue managers do not agree on the next message sequence number. A message with sequence number 324246 has been sent when sequence number 194294 was expected. The remote host is '123.123.123.3'.
I have tried resetting the sequence number at receiver end of the channel to match the number at sender’s end. This solves the issue temporary but whenever next time if channel goes down (due to n/w issue or idle timeout); again error AMQ9526 occurs.
I have also tried granging ‘crt’ and ‘all’ authorization for the MCAUSER for SYSTEM.CHANNEL.SYNCQ queue (this queue remembers the channel numbers). |
|
Back to top |
|
 |
Mr Butcher |
Posted: Tue Dec 13, 2011 2:46 am Post subject: |
|
|
 Padawan
Joined: 23 May 2005 Posts: 1716
|
are these the real sequence numbers? they differ too much for just a batch being lost somehow..... do you have control over the remote end? looks like, thats a different queuemanager as it was before the connection interruption _________________ Regards, Butcher |
|
Back to top |
|
 |
gbaddeley |
Posted: Tue Dec 13, 2011 2:40 pm Post subject: |
|
|
 Jedi Knight
Joined: 25 Mar 2003 Posts: 2538 Location: Melbourne, Australia
|
nak1984 wrote: |
Let me explain in bit more detail:
I’m facing problem at the receiver end of the channel always. Problem occurs in case, for example, n/w connection is down then the channel terminates. Now when the channel is established after a while, then it complains sequence number is not matching: |
Does the CONNAME used by the sender channel resolve into multiple IP addresses, all going to the receiver host ?
Quote: |
I have also tried granging ‘crt’ and ‘all’ authorization for the MCAUSER for SYSTEM.CHANNEL.SYNCQ queue (this queue remembers the channel numbers). |
This would only help if the MCA was experiencing authorization errors. _________________ Glenn |
|
Back to top |
|
 |
nak1984 |
Posted: Wed Dec 14, 2011 12:45 am Post subject: |
|
|
Newbie
Joined: 11 Dec 2011 Posts: 5
|
Butcher,
The sequence numbers are real (I’ve only changed a few digits).
I can understand that you are thinking loss of those many messages but that’s not the situation. Message loss is 0.
I can explain the number thing. Our end of the channel (receiver end) is only able to remember the seq. number set by us. i.e. if during the last seq. number issue we reset our end to seq. number 194294 to match the sender end at that time, next time when channel disconnects due to some problem, although many messages have been exchanged before this, the receiver end will remember 194294!
If I use runmqsc and get current sequence number when channel is RUNNING it gives correct number but the issue is when channel goes down it fails to remember the current seq. number. Also I don’t have any control over remote end. But I can ask remote end to change any parameter if it can resolve this issue.
CONNAME is specified as IP(PORT) [123.123.1.2 (2341)] so there is no case of resolving in to multiple IP addresses. I tried ‘crt’ and ‘all’ authorization as I was not sure of the cause of the problem.
I got this from
http://www-01.ibm.com/support/docview.wss?uid=swg1IZ78326 |
|
Back to top |
|
 |
Mr Butcher |
Posted: Wed Dec 14, 2011 1:47 am Post subject: |
|
|
 Padawan
Joined: 23 May 2005 Posts: 1716
|
sounds wired. there is only one sender connecting to your receiver? no fdc files? no other messages in the amqerrxx.log files related tothis issue? what mq version do you run exactly (dspmqver) _________________ Regards, Butcher |
|
Back to top |
|
 |
mqjeff |
Posted: Wed Dec 14, 2011 2:39 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
... I would delete and redefine the rcvr channel. |
|
Back to top |
|
 |
bruce2359 |
Posted: Wed Dec 14, 2011 5:23 am Post subject: |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
mqjeff wrote: |
... I would delete and redefine the rcvr channel. |
By the way, this is the most common and likely cause of the sequence number mismatch in the first place. _________________ I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
Back to top |
|
 |
mqjeff |
Posted: Wed Dec 14, 2011 5:28 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
bruce2359 wrote: |
mqjeff wrote: |
... I would delete and redefine the rcvr channel. |
By the way, this is the most common and likely cause of the sequence number mismatch in the first place. |
Yes, but we've already eliminated that (at least mostly) through the discussion in this thread. Unless there's something being done as part of
Quote: |
if channel goes down (due to n/w issue or idle timeout); again error AMQ9526 occurs. |
|
|
Back to top |
|
 |
bruce2359 |
Posted: Wed Dec 14, 2011 5:38 am Post subject: |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
Memory leaks on a variety of WMQ versions have caused AMQ9526.
Found this via Mr. Google: http://www-01.ibm.com/support/docview.wss?uid=swg1IZ38767 _________________ I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
Back to top |
|
 |
gbaddeley |
Posted: Wed Dec 14, 2011 3:12 pm Post subject: |
|
|
 Jedi Knight
Joined: 25 Mar 2003 Posts: 2538 Location: Melbourne, Australia
|
nak1984 wrote: |
Let me explain in bit more detail:
I’m facing problem at the receiver end of the channel always. Problem occurs in case, for example, n/w connection is down then the channel terminates. Now when the channel is established after a while, then it complains sequence number is not matching:
i.e. AMQ9526: Message sequence number error for channel ‘xx.xx.xx'.
EXPLANATION:
The local and remote queue managers do not agree on the next message sequence number. A message with sequence number 324246 has been sent when sequence number 194294 was expected. The remote host is '123.123.123.3'.
I have tried resetting the sequence number at receiver end of the channel to match the number at sender’s end. This solves the issue temporary but whenever next time if channel goes down (due to n/w issue or idle timeout); again error AMQ9526 occurs. |
After the rcvr channel has terminated, what sequence number has MQ saved in its channel status?
Code: |
display chstatus('ChannelName') saved all |
_________________ Glenn |
|
Back to top |
|
 |
fjb_saper |
Posted: Wed Dec 14, 2011 9:28 pm Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
Is there an mcauser set on this channel at the receiver end?
Is the receiver a mainframe? (zOS).
You might also want to check the entries in SYSTEM.CHANNEL.SYNCQ and verify whether there are multiple entries for this channel?
If there are multiple entries, open a PMR with IBM to get it resolved.
Have fun  _________________ MQ & Broker admin |
|
Back to top |
|
 |
BBM |
Posted: Thu Jan 12, 2012 9:08 am Post subject: |
|
|
Master
Joined: 10 Nov 2005 Posts: 217 Location: London, UK
|
|
Back to top |
|
 |
|