| Author |
Message
|
| Northern Virginia Coder |
 Posted: Mon Jan 27, 2014 7:11 am Post subject: Testing in Migration of mainframes |
 |
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
Hi - It would be so easy to port MQ over without name changes, can I do this? When I test can I set up a parallel system with duplicate names in the QA LPAR? There is no clustering on the system. Thanks!  |
|
| Back to top |
|
 |
| Vitor |
| Posted: Mon Jan 27, 2014 7:17 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
What do you need to port that has queue manager names in it?
No application should ever know what queue manager it's connecting to. If you have queue names which vary from environment to environment:
- why?
- why are these values not parameterized?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Mr Butcher |
| Posted: Mon Jan 27, 2014 7:22 am Post subject: |
|
|
Padawan
Joined: 23 May 2005
Posts: 1716
|
You can not have 2 qmgrs of the same name in an LPAR at the same time.
you can set up 2 or more with the same name, by using different datasets and jcl, but you can start only one. however, thats not a good setup, just think of security, batch processing, users getting confused, ..... .......
_________________
Regards, Butcher |
|
| Back to top |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 7:51 am Post subject: Duplicate Queue Manager names |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
| we are migrating from mainframe 1 to mainframe 2. there will be a period of time when both mainframes are active and connected to each other and to the distributed environment. we want to conserve naming from mainframe 1 to mainframe 2, queue naming should stay the same. So what queue managers does the new mainframe need have to allow routing from (old) mainframe 1 to (new) mainframe 2 to distributed and still function without code change as (old) mainframe 1 did after shutdown of (old) mainframe 1 occurs? |
|
| Back to top |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 7:56 am Post subject: restated |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
if A->B originally as mainframe 1 (A) goes to distributed system (B)
then
A->C and D->B in mainframe 2 at the same time A->B
so in mainframe 2, the new mainframe, C and D queue managers could be clustered, but we are trying to avoid the use of clustering on the zOS for the time being |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 27, 2014 8:17 am Post subject: Re: Duplicate Queue Manager names |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Northern Virginia Coder wrote: |
| So what queue managers does the new mainframe need have to allow routing from (old) mainframe 1 to (new) mainframe 2 to distributed and still function without code change as (old) mainframe 1 did after shutdown of (old) mainframe 1 occurs? |
I repeat my question - what in the code knows what environment it's running in?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 9:22 am Post subject: messaging |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
the message would contain the queue and host and some other uninteresting attributes, so....
A->B and A->C->A1->B where A is old mainframe and C and A1 are queue managers on new mainframe and B is distributed system. A and A1 would be named the same but on different hosts
is that reasonable? |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 27, 2014 9:37 am Post subject: Re: messaging |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Northern Virginia Coder wrote: |
| the message would contain the queue and host |
Why?
| Northern Virginia Coder wrote: |
| is that reasonable? |
No
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 9:55 am Post subject: not sure I know why I am wrong |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
so since there are 2 different mainframes, the queue manager names can be conserved
A->B
A->A1->B A and A1 can be exact but on different hosts
just set up MQ on the new server with all the naming and attributes as old system and route those messages through? |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 27, 2014 9:58 am Post subject: Re: not sure I know why I am wrong |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Northern Virginia Coder wrote: |
| so since there are 2 different mainframes, the queue manager names can be conserved |
No.
| Northern Virginia Coder wrote: |
| just set up MQ on the new server with all the naming and attributes as old system and route those messages through? |
Of course not.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 27, 2014 10:54 am Post subject: Re: Duplicate Queue Manager names |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| Northern Virginia Coder wrote: |
| we are migrating from mainframe 1 to mainframe 2. there will be a period of time when both mainframes are active and connected to each other |
You cannot create duplicate qmgr names in the same o/s instance (LPAR).
For clarity, when you say you will have the old m/f LPAR and new m/f LPAR connected, you don't literally mean that you will have QMGR1 (old) connected to QMGR1 (new) - identically named qmgrs connected to teach other - do you?
The best-practice is to not have the same qmgr name replicated in your network. In you network means interconnected in any way. As long as old QMGR1 and new QMGR1 don't connect to each other, there should be no problems.
_________________
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 |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 12:13 pm Post subject: testing scenerio |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
Thanks, ok. I was thinking that the queue managers with the same names but on different mainframes would not connect. I want to keep the intermediate-temporary shared environment as simple as possible but without clustered queue managers. So if I set up a temporary queue manager that can connect to the old server and another queue manager on the new server with the same naming as the old system I should be ok.
A->B and A->C C->D D->B where A is the old mainframe, B is distributed system, and C and D are queue managers on the new system. A and D can have the same names to keep things simple. When A is shut down I can remove C.
Then all that is required is some fancy MQ admin work.
Is this type of thing ever done that you have seen? or do people just create clusters? Thanks for all help! |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 27, 2014 12:30 pm Post subject: Re: testing scenerio |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Northern Virginia Coder wrote: |
| Then all that is required is some fancy MQ admin work. |
Very fancy & error prone.
| Northern Virginia Coder wrote: |
| Is this type of thing ever done that you have seen? |
Typically you don't need this because neither the application code nor the topology require it.
| Northern Virginia Coder wrote: |
| do people just create clusters? |
In some instances, especially when you're trying to preserve topology, a cluster could make things worse not better.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Northern Virginia Coder |
| Posted: Mon Jan 27, 2014 12:59 pm Post subject: clean scenerio with less rework |
|
|
Newbie
Joined: 27 Jan 2014
Posts: 8
|
| so you propose that the old and new systems work in parallel so that as a line of data and messaging is tested and comes on line in the new system it is shut down in the old system if possible, no? then any connection between the old and new system would only be for verification of data transfer. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 27, 2014 1:34 pm Post subject: Re: clean scenerio with less rework |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| Northern Virginia Coder wrote: |
| ... then any connection between the old and new system would only be for verification of data transfer. |
Please be a bit more technically precise when you post. Servers don't connect, systems don't connect, mainframes don't connect, LPARs don't connect. WMQ channels connect to other WMQ channels.
You cannot create a duplicate qmgr name in a single o/s instance. You may have any number of identically named qmgrs across multiple platforms, LPARs and o/s images.
What do you mean by connecting one system to another? Are you thinking of moving messages from a queue on an old qmgr to a queue on a new qmgr? Something like that? While rarely done, it is possible to have two identically named qmgrs connected to each other.
Again, please be precise as to your requirement and intent.
_________________
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 |
|
|
|
|
