| Author |
Message
|
| George Carey |
 Posted: Thu Sep 08, 2011 11:49 am Post subject: Assured delivery |
 |
|
Knight
Joined: 29 Jan 2007
Posts: 500
Location: DC
|
Whew!! Stumbled on this looking for something else but was anything resolved in this disussion ?
The original question was about whether the new Client connection capabilities now had the IBM assured delivery moniker associated with them.
As was noted by mvic concerning 'assured delivery' :
| Quote: |
| I don't see this phrase as having any specific technical meaning in the software world. |
Likely because it doesn't. If you recall this was an IBM legal department change from the previous term 'guaranteed delivery' which unconditioned is an impossibity of course. Leaving IBM open to a whole lot of tort actions ..
So suspenders assure your pants will stay up but not guarantee it and MQI Clients calls have assured delivery of messages but just not as assured as MQS calls but neither are guaranteed delivery. A better use of time might be what MQ assured delivery offers over and above other types of message delivery paradigms.
GTC
_________________
"Truth is ... grasping the virtually unconditioned",
Bernard F. Lonergan S.J.
(from book titled "Insight" subtitled "A Study of Human Understanding") |
|
| Back to top |
|
 |
| mvic |
| Posted: Thu Sep 08, 2011 12:10 pm Post subject: Re: Assured delivery |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| George Carey wrote: |
As was noted by mvic concerning 'assured delivery' :
| Quote: |
| I don't see this phrase as having any specific technical meaning in the software world. |
Likely because it doesn't. |
That's one of the things I think was established in the discussion.
Was there anything besides this that you need to be discussed / resolved for a situation you are facing? |
|
| Back to top |
|
|
| George Carey |
| Posted: Thu Sep 08, 2011 12:24 pm Post subject: handle states |
|
|
Knight
Joined: 29 Jan 2007
Posts: 500
Location: DC
|
I was doing a google seach on MQ client 'handle states' and trying to find a good source for discussion/explanation of them ... in my on going/never ending search for a complete understanding of Client connection behavior and their conditioners and contributors.
If you have a good link to such a source I would be glad to follow it.
Also, to your comment
| Quote: |
| That's one of the things I think was established in the discussion. |
I didn't think it was an established point to all ... but feel it may now be.
GTC
_________________
"Truth is ... grasping the virtually unconditioned",
Bernard F. Lonergan S.J.
(from book titled "Insight" subtitled "A Study of Human Understanding") |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Sep 08, 2011 1:22 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| mvic wrote: |
I don't see this phrase as having any specific technical meaning ... |
It does, in WMQ. A search of Google for "assured delivery"+wmq leads to a small fleet of meanings - all of them relating ONLY to the message delivery.
IBMs definitions (including those in the APG and APR) include hardware and o/s failures, and recovery of the message at qmgr restart for UofWs in-flight; and messages transmitted across channels in UofWs; and whether or not a message is committed/backed out if the app fails, and so on.
In the definitions I've read, none address re-instantiating an application to continue its processing should the connection should become broken. The only platform (that I know of) that can address a failed application is WMQ in a z/OS Parallel Sysplex.
_________________
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 |
|
|
| mvic |
| Posted: Thu Sep 08, 2011 1:23 pm Post subject: Re: handle states |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| George Carey wrote: |
| If you have a good link to such a source I would be glad to follow it. |
Have you seen: http://publib.boulder.ibm.com/infocenter/wmqv7/v7r0/topic/com.ibm.mq.amqzag.doc/fa70190_.htm
| Quote: |
Also, to your comment
| Quote: |
| That's one of the things I think was established in the discussion. |
I didn't think it was an established point to all ... but feel it may now be. |
To be fair, it was me who was saying it mostly. But others, such as bruce, were agreeing, I think. |
|
| Back to top |
|
|
| mvic |
| Posted: Thu Sep 08, 2011 4:08 pm Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| bruce2359 wrote: |
| mvic wrote: |
I don't see this phrase as having any specific technical meaning ... |
It does, in WMQ. A search of Google for "assured delivery"+wmq leads to a small fleet of meanings |
I like your phrase "a small fleet of meanings". That rather makes my point, I think. 
Among those search results I did not see any I would consider as specific or technical, but admittedly I could have overlooked the one that was.
I know what people tend to mean by the phrase... but it's not a technical phrase, used by technical people, in my experience. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Sep 08, 2011 4:25 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| mvic wrote: |
| If any of those fleet of meanings are specific or technical, please post a link. |
None of them are concise, in the sense that a single sentence or paragraph adequately describes what it takes to assure delivery."
The shortest and sweetest definition that I've seen is "messages are delivered once and once only, with no loss or duplication;" but there a large fleet of presumptions and prerequisites. As we've discussed here, there are an infinite (-1) number of reasons that a particular message could not be delivered. The specific and technical definition that you seek must encompass all of the exceptions, mustn't it?
MQ assured delivery is somewhat parallel to the US Postal Service version. The PO will attempt to deliver if the destination is known, and if the destination is valid. Valid in PO-speak means that the destination address exists (isn't a vacant lot), has a suitable mail-box or person who will accept the mail.
Optionally, both MQ and PO will notify the sender if the message can or cannot be delivered.
Both MQ and PO presume that the infrastructure exists to transport the message, and that the infrastructure is working perfectly.
The APR and APG offer ReasonCodes to deal with exceptions; and developers are charged with dealing appropriately with all of them. In addition, the WMQ Messages manual offers informational AMQ/CSQ messages.
Obvious understatement: It's complicated.
_________________
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 |
|
|
| mvic |
| Posted: Thu Sep 08, 2011 4:30 pm Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| Bruce, I amended my post while you were replying.. I apologise. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Sep 08, 2011 4:53 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Absolutely no need to apologize (or apologise). Our lot is plagued by words and phrases with similar, opposite, and overlapping meanings. ("cluster" being one of my favorites.)
We do our best work when we strive for precision; and we get to annoy others in the process of doing so.
The quest is endless. Search next for the definition of precision. How does it differ from accuracy? Seemingly, words that imply the same meaning. Not.
_________________
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: Thu Sep 08, 2011 5:21 pm Post subject: |
|
|
Jedi Knight
Joined: 25 Mar 2003
Posts: 2538
Location: Melbourne, Australia
|
I'd suggest that "assured delivery" only applies to the transactional MQ Client. The ordinary MQ Client can't have "assured delivery" because the network transport session could be lost in the middle of an MQI call, and the Client would have no way of knowing if the message was accepted or not by the Queue Manager.
Handles have an internal state in MQ. Unless its a value defined in cmqc.h, eg. MQHC_DEF_HCONN (0), MQHC_UNUSABLE_HCONN (-1), contact admin (-3), MQHO_UNUSABLE_HOBJ (-1), MQHO_NONE 0, the only way I know to test the state is to use it as an argument to a MQI call.
_________________
Glenn |
|
| Back to top |
|
|
| George Carey |
| Posted: Fri Sep 09, 2011 2:54 pm Post subject: assured delivery |
|
|
Knight
Joined: 29 Jan 2007
Posts: 500
Location: DC
|
As I said perhaps:
| Quote: |
| A better use of time might be what MQ assured delivery offers over and above other types of message delivery paradigms. |
Maybe an itemization of assumptions paired with the set of connection anamolies that could occur, the following behavior is exhibited for MQI(non-etc) connections, MQI (etc) connections, and MQS connections.
...
-----------------------------------------------------
Also on the handle states mentioned:
| Quote: |
| Handles have an internal state in MQ. Unless its a value defined in cmqc.h, eg. MQHC_DEF_HCONN (0), MQHC_UNUSABLE_HCONN (-1), contact admin (-3), MQHO_UNUSABLE_HOBJ (-1), MQHO_NONE 0, the only way I know to test the state is to use it as an argument to a MQI call. |
how do these map to the MQ Explorer screen (I believe queue status) column labeled 'handle state' ?
_________________
"Truth is ... grasping the virtually unconditioned",
Bernard F. Lonergan S.J.
(from book titled "Insight" subtitled "A Study of Human Understanding") |
|
| Back to top |
|
|
| gbaddeley |
| Posted: Sun Sep 11, 2011 5:20 pm Post subject: Re: assured delivery |
|
|
Jedi Knight
Joined: 25 Mar 2003
Posts: 2538
Location: Melbourne, Australia
|
| George Carey wrote: |
Also on the handle states mentioned:
| Quote: |
| Handles have an internal state in MQ. Unless its a value defined in cmqc.h, eg. MQHC_DEF_HCONN (0), MQHC_UNUSABLE_HCONN (-1), MQHC_U N A S S O C I A T E D_HCONN (-3), MQHO_UNUSABLE_HOBJ (-1), MQHO_NONE 0, the only way I know to test the state is to use it as an argument to a MQI call. |
how do these map to the MQ Explorer screen (I believe queue status) column labeled 'handle state' ? |
Ahh, that's different to my initial impression of 'handle state'.
The WMQ PCF manual says MQIACF_HANDLE_STATE indicates "Whether an API call is in progress."
Header file cmqcfc.h contains:
| Code: |
/* Handle States */
#define MQHSTATE_INACTIVE 0
#define MQHSTATE_ACTIVE 1 |
_________________
Glenn |
|
| Back to top |
|
|
| aydink |
| Posted: Wed Dec 02, 2015 1:22 am Post subject: |
|
|
Newbie
Joined: 28 Aug 2015
Posts: 2
|
| sunny_30 wrote: |
This is what i understood from the discussion so far:
Im not a MQ expert.. So, please correct me if Im wrong.
MQserver-binding-conn + XA = 2PC
MQserver-binding-conn w/o XA = 1PC
MQclient -> always 1PC, XA not possible
MQtxClient w/o XA = 1PC
MQtxClient + XA = 2PC
2PCs provide 'once-only' assured messaging (if msg set to persistent)
1PCs dont provide assurity as they cannot support XA
MQCMIT call possible with MQclient, but doesnt help as XA not an option
Only XA supports MQBACK option
|
As I did not find any good technical summary related to this e2e once-and-only-once delivery topic better than your summary on the net, I decided to reactivate this thread 
I made one update on your summary though, client bindings mode and XA is possible. Can somebody confirm the other assumptions by sunny_30?
MQserver-binding-conn + XA = 2PC
MQserver-binding-conn w/o XA = 1PC
MQclient -> always 1PC, XA possible with extension – free since v7 (http://www-01.ibm.com/support/docview.wss?uid=swg21605388)
MQtxClient w/o XA = 1PC
MQtxClient + XA = 2PC
2PCs provide 'once-only' assured messaging (if msg set to persistent)
1PCs dont provide assurity as they cannot support XA
MQCMIT call possible with MQclient, but doesnt help as XA not an option
Only XA supports MQBACK option |
|
| Back to top |
|
|
| zpat |
| Posted: Wed Dec 02, 2015 2:59 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
File Transfer or Web Services - no protection at all from failed connection, partial or total loss, duplication, out of sequence etc.
WSRM - no-one uses it! 
_________________
Well, I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Dec 02, 2015 3:32 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| aydink wrote: |
1PCs dont provide assurity as they cannot support XA
MQCMIT call possible with MQclient, but doesnt help as XA not an option
Only XA supports MQBACK option |
This is not how it works.
Any message with persistency set to yes and no expiry participates in assured delivery. For this participation to become effective the message needs to be committed if it is in a Unit of Work. There is no difference in whether the message was part of a single phase commit or a multiple phase commit. The message needs to have been committed or put successfully outside of a Unit of Work. 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
|
|
