| Author |
Message
|
| dragos.musat |
 Posted: Mon Dec 24, 2018 6:27 am Post subject: Message segments have different encodings |
 |
|
Newbie
Joined: 18 Dec 2018
Posts: 3
|
Hi all,
I recently had some tests on IIB 9.0.0.5 and MQ 7.5 (07050001) on Windows Server 2008 Std R2.
The problem is when i work with large messages that require segmentation, the las segment has different encoding that the first one, and that gets me on a mq get with an warning 2244 
Error.ErrorCode=2651
Error.ErrorMessage=Warning on MQGET. Propagating a message to the failure terminal. 2244
MQRC_INCONSISTENT_ENCODINGS
(2244, X'8C4') Message segments have differing encodings.
The flow is very simple . i put a blob message on a MQ in queue and without any compute in between i wired to an MQ output with segmentation allowed checked flag.
I then watched the output queue with MQ explorer. i see that there are 2 messages there but the first message has encoding 546 and ccsid 1208 and the second one has encoding 1208 and ccsid 1208.
Why does this happen ? why the second message has a different encoding than the first one ? i didn't change the input msg...i let the MQ decide how to split. Is any configuration that i am missing?
Thank you,
Merry Christams
Dragos |
|
| Back to top |
|
 |
| timber |
| Posted: Mon Dec 24, 2018 12:22 pm Post subject: |
|
|
Grand Master
Joined: 25 Aug 2015
Posts: 1292
|
That certainly is rather surprising behaviour because the MQMD.Encoding field is nothing to do with code pages - it controls how an MQ application reads numbers.
| Quote: |
| the first message has encoding 546 and ccsid 1208 and the second one has encoding 1208 and ccsid 1208. |
Very strange. So strange that I wonder whether you are reading the values incorrectly. I'm not sure that 1208 is even a meaningful value for MQMD.Encoding. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Dec 27, 2018 3:32 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| timber wrote: |
That certainly is rather surprising behaviour because the MQMD.Encoding field is nothing to do with code pages - it controls how an MQ application reads numbers.
| Quote: |
| the first message has encoding 546 and ccsid 1208 and the second one has encoding 1208 and ccsid 1208. |
Very strange. So strange that I wonder whether you are reading the values incorrectly. I'm not sure that 1208 is even a meaningful value for MQMD.Encoding. |
Perhaps somebody meant to set MQMD.CCSID and set MQMD.ENCODING instead??? 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| timber |
| Posted: Sat Dec 29, 2018 8:51 am Post subject: |
|
|
Grand Master
Joined: 25 Aug 2015
Posts: 1292
|
I agree that's a possibility. But in this case, the 'somebody' might be IBM because the OP said:
| Quote: |
| The flow is very simple . i put a blob message on a MQ in queue and without any compute in between i wired to an MQ output with segmentation allowed checked flag. |
That's why I suggested a re-check. |
|
| Back to top |
|
|
| dragos.musat |
| Posted: Thu Jan 03, 2019 1:23 am Post subject: |
|
|
Newbie
Joined: 18 Dec 2018
Posts: 3
|
i've checked again, with rfhutil and mq explorer....the first message has encoding 546 and ccsid 1208 as it should, then after segmentation is done the explained behavior is present. So ...i am reading the values correctly for sure. No one is trying to change the input mqmd headers, they are just passed thru. there is no computation done between the mq in and mq out nodes.
i will check this behavior on IBM App Connect 11.0.0.3 as the fix pack was released recently and IBM MQ 9.1 and hopefully this won't happen  |
|
| Back to top |
|
|
| rekarm01 |
| Posted: Mon Jan 07, 2019 4:35 pm Post subject: Re: Message segments have different encodings |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 1415
|
| timber wrote: |
| I'm not sure that 1208 is even a meaningful value for MQMD.Encoding. |
It is not meaningful. Ignoring the unrecognized bits, it roughly means: MQENC_FLOAT_TNS + MQENC_DECIMAL_UNDEFINED + MQENC_INTEGER_UNDEFINED
| dragos.musat wrote: |
| i will check this behavior on IBM App Connect [Enterprise] 11.0.0.3 as the fix pack was released recently and IBM MQ 9.1 and hopefully this won't happen |
Hopefully. |
|
| Back to top |
|
|
| hughson |
| Posted: Mon Jan 07, 2019 7:39 pm Post subject: |
|
|
Padawan
Joined: 09 May 2013
Posts: 1959
Location: Bay of Plenty, New Zealand
|
Given that the encoding is set to the same value as the CCSID (1208) it sounds very like a defect somewhere. 1208 is not a normally seen encoding value (even if you pick out the valid bits). It is just too coincidental that it has the same value as the CCSID.
Cheers,
Morag
_________________
Morag Hughson @MoragHughson
IBM MQ Technical Education Specialist
Get your IBM MQ training here!
MQGem Software |
|
| Back to top |
|
|
| dragos.musat |
| Posted: Tue Jan 08, 2019 12:01 am Post subject: |
|
|
Newbie
Joined: 18 Dec 2018
Posts: 3
|
I've made the same test with IBM ACE 11.0.0.2 and MQ 9.1 on a linux platform, and the results are not the same ! the defect does not reproduce on that system. Thats more to me that this is a product defect...but we managed to find a workaround untill we migrate to ACE 11.
Thanks a lot for all the help ! |
|
| Back to top |
|
|
|
|
