| Author |
Message
|
| RichA |
 Posted: Wed Feb 11, 2009 8:49 am Post subject: MQRC 2009 |
 |
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
Recently we've started getting a bunch of 2009 errors from a java app that was last updated October last year, since January we've had irregular (but growing in frequency) occurances of 2009 errors it seems to happen when the app tries to connect to a queue which it's cached but hasn't used for a while.
Personally I want to change the code, but given the environment it's unlikely that'll happen any time soon.
We're out of support so PMR's no good, it's running on MQ 5.3 CSD 7, also unlikely to change, there don't appear to be any MQ exceptions (other than the 2009 thrown by the app) or FFST's. Nothing sticks out in the OS logs (SLES 8 )
It's a batch application so it's not hurting us too much, we can kill the app, restart it and it works no problem (also without touching the QM) it's more out of curiosity I'm asking about this one than anything. The whole lot's due to be upgraded easter so I'm not expecting to have to fix it.
Oh and here's the rub, it's a bindings connection and we've failed over the application and queue manager to a second server ruling out hardware.
I wouldn't be asking if it wasn't bindings.
<edit>
I just realised I posted this in the wrong forum, it should probably go in the Java/JMS forum instead, hopefully a friendly moderator will shift it
</edit>
Last edited by RichA on Wed Feb 11, 2009 8:55 am; edited 1 time in total |
|
| Back to top |
|
 |
| mqjeff |
| Posted: Wed Feb 11, 2009 8:54 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
2009's from java applications is exactly why CSD7 of MQ v5.3 was withdrawn.
If your business (or your customer's business) believes that running software without staying current on maintenance is a valid practice, then suggest to them that they must also run their automobiles without changing the oil. |
|
| Back to top |
|
|
| RichA |
| Posted: Wed Feb 11, 2009 9:00 am Post subject: |
|
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
lol
just
lol |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Feb 11, 2009 10:13 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
Moved to JMS section
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Feb 11, 2009 10:18 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| mqjeff wrote: |
2009's from java applications is exactly why CSD7 of MQ v5.3 was withdrawn.
If your business (or your customer's business) believes that running software without staying current on maintenance is a valid practice, then suggest to them that they must also run their automobiles without changing the oil. |
 slightly
If they're using a version which is not just out of support but withdrawn by IBM, they probably changed the oil for water years back.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| RichA |
| Posted: Wed Feb 11, 2009 12:09 pm Post subject: |
|
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
The app's been working in basically the same way (MQ wise) since the dawn of time, it's been on CSD 7 since there was a CSD 7, it's been tested extensively in 3 different environments.
I accept what you're saying, but the approach has been if it's not broken why change it?
We already pointed out the flaw in that argument and there's a project underway to move to SLES 10 and MQ 6 (7 came too late for the project inception) that's due to roll out Easter but knowing the way these things have gone previously that's going to drop back a way.
I'll push to go to a later CSD, but I'm not going to hold my breath. Most likely thing to happen is it'll get absorbed as a support process, check the app logs for 2009's if they're there restart app.
I could try to push for MQ 6 in, but as there's already a big project around that and it's be running on an unsupported OS there's not a lot of point, it'll be a big fat NO. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Feb 11, 2009 1:47 pm Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| RichA wrote: |
| the approach has been if it's not broken why change it? |
Because...
| RichA wrote: |
| there's a project underway to move to SLES 10 and MQ 6 |
....rather than a small, incremental change you (or the powers that be) now have a big bang. Also with such a back level you can never be sure what the synergistic change will be of any OS or other patch.
Another refutation of the"why change it" question is this post. Assume for a second that a simple restart didn't fix the problem. Or doesn't continue to fix the problem after the next Java patch. Or something. You call IBM, who point out the software is out of support. You post here, and we say "yes, CSD07 does that". You need to do an urgent patch to your system, all the while your application (and the business processes that rely on it) are sitting in the bottom of the server, twitching.
How much money does that cost the business?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| zpat |
| Posted: Wed Feb 11, 2009 1:57 pm Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
| This problem can be exposed when the queue manager is upgraded and the client app is not (in terms of MQ levels). |
|
| Back to top |
|
|
| mqjeff |
| Posted: Wed Feb 11, 2009 2:26 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| RichA wrote: |
The app's been working in basically the same way (MQ wise) since the dawn of time, it's been on CSD 7 since there was a CSD 7, it's been tested extensively in 3 different environments.
I accept what you're saying, but the approach has been if it's not broken why change it? |
But if it's been tested in 3 different working environments, and it doesn't work "all of a sudden", then obviously *something* has changed. So someone did something.
Or the "new" environment (that is "the same" as the tested environments) is not actually "the same", but different in some way. So, again, someone did something. |
|
| Back to top |
|
|
| RichA |
| Posted: Thu Feb 12, 2009 1:40 am Post subject: |
|
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
| Quote: |
| Assume for a second that a simple restart didn't fix the problem. Or doesn't continue to fix the problem after the next Java patch. Or something. |
There's not been a java patch for an age either, it's 1.4.2 SR 2, it's not likely to change.
| Quote: |
| How much money does that cost the business? |
Money is not the main concern of the business. I won't say any more than that.
| Quote: |
| But if it's been tested in 3 different working environments, and it doesn't work "all of a sudden", then obviously *something* has changed. So someone did something. |
There may have been changes to the OS during the period we're talking about, but it's out of my control so I couldn't say.
My current working theory is the volume of traffic has changed and there's something in CSD 7 which was (& should be) thread safe, but isn't.
It would really help if IBM had some kind of announcement (maybe they did and it just didn't reach the right people) the CSD was withdrawn. If someone could direct me to an announcement, some kind of IBM document revealing the fact the CSD is withdrawn (even better if the reasons for withdrawl are documented) it would be really helpful.
I'm a bit disappointed (understatement) this didn't happen when I speculatively opened a PMR. Instead they ask for us to take out an extended support agreement, great policy, take more money and tell the customer what they should have been told to start with. |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Feb 12, 2009 1:58 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| RichA wrote: |
| Money is not the main concern of the business. I won't say any more than that. |
An organisation that's not principally interested in money? Now that's something you don't see everyday. Even charities don't like losing money (with better reason than most businesses).
| RichA wrote: |
| My current working theory is the volume of traffic has changed and there's something in CSD 7 which was (& should be) thread safe, but isn't. |
It's not a bad theory. Java support was highly dubious back in the day.
| RichA wrote: |
| It would really help if IBM had some kind of announcement (maybe they did and it just didn't reach the right people) the CSD was withdrawn. |
http://www-01.ibm.com/support/docview.wss?uid=swg27007825
Interestingly CSD07 is still shown as available: it's 8 & 9 that are withdrawn (along with 3). I thought all 3 of the notorious Java patches had been pulled.
| RichA wrote: |
| I'm a bit disappointed (understatement) this didn't happen when I speculatively opened a PMR. |
Which part of "WMQv5.3 is out of support" was not clear to you? Why would the support desk provide support for an out of support product without an extended support agreement?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| RichA |
| Posted: Thu Feb 12, 2009 2:11 am Post subject: |
|
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
IC45588 JMS client application fail with 2009 (mqrc_connection_broken) at receive
This might be the one, although it does say client, if I show it to the people who make decisions they probably wont have a clue anyway.
| Quote: |
| Which part of "WMQv5.3 is out of support" was not clear to you? Why would the support desk provide support for an out of support product without an extended support agreement? |
Yeah because telling someone there's a known problem with the version is a real function of support that someone should pay extra for. A better response may have been to suggest we move to a supported version, how does that sound?
Vitor you come across as really confrontational.  |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Feb 12, 2009 2:17 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| RichA wrote: |
| Yeah because telling someone there's a known problem with the version is a real function of support that someone should pay extra for. A better response may have been to suggest we move to a supported version, how does that sound? |
Before Western capitalism melted they probably would have said that. These days you chase every penny. Well, your organisation doesn't, but IBM has always taken a different view of the accumulation of wealth.
| RichA wrote: |
| Vitor you come across as really confrontational. |
Thank you! 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| zpat |
| Posted: Thu Feb 12, 2009 2:23 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
I've seen something like this happen when the queue manager was upgraded and the solution was to upgrade the MQ JAR files for the JMS application to the latest CSD for MQ v5.3 or for MQ v6.
No program changes were needed. Just replace the IBM MQ Jar files. Done. |
|
| Back to top |
|
|
| RichA |
| Posted: Thu Feb 12, 2009 2:35 am Post subject: |
|
|
Centurion
Joined: 14 Mar 2002
Posts: 102
|
| Quote: |
| No program changes were needed. Just replace the IBM MQ Jar files. Done. |
The JAR files at a different CSD level to the QM?
That would probably be easier to get approved, was that a client app or server bindings? |
|
| Back to top |
|
|
|
|
