Issue 5934: Forcing child activities of completed parents to fail is too restrictive (ots-structs-ftf) Source: International Business Machines (Dr. Alex Mulholland, nobody) Nature: Uncategorized Issue Severity: Summary: Section 2.2.9 of the specification, in the description of the compete method, includes the following text: "If the completion status is CompletionStatusFail, or CompletionStatusFailOnly, any encompassed active or suspended Activities will then have their completion status set to CompletionStatusFailOnly and transactions will be marked rollback_only." This behaviour is too restrictive for extended transaction models that may wish to allow child activities to complete successfully despite the fact that a parent activity is failed. I suggest that this restriction be removed. Resolution: Revised Text: Actions taken: May 7, 2003: received issue Discussion: End of Annotations:===== Subject: New issue for Additional Structures for OTS Revision Task Force To: issues@omg.org Cc: "Ian Robinson" X-Mailer: Lotus Notes Release 5.0.11 July 24, 2002 From: "Alex Mulholland" Date: Wed, 7 May 2003 14:33:06 +0100 X-MIMETrack: Serialize by Router on d06ml007/06/M/IBM(Release 5.0.9a |January 7, 2002) at 07/05/2003 14:32:47 Hello, I would like to submit the following issue to the Additional Structures for OTS Revision Task Force discussion list: Forcing child activities of completed parents to fail is too restrictive Section 2.2.9 of the specification, in the description of the compete method, includes the following text: "If the completion status is CompletionStatusFail, or CompletionStatusFailOnly, any encompassed active or suspended Activities will then have their completion status set to CompletionStatusFailOnly and transactions will be marked rollback_only." This behaviour is too restrictive for extended transaction models that may wish to allow child activities to complete successfully despite the fact that a parent activity is failed. I suggest that this restriction be removed. Dr Alex Mulholland IBM Hursley Laboratory, UK. amulholl@uk.ibm.com From: "Mark Little" To: , "Ian Robinson" Subject: Re: New issue for Additional Structures for OTS Revision Task Force Date: Fri, 9 May 2003 11:04:01 +0100 X-Mailer: Microsoft Outlook Express 6.00.2720.3000 X-Newcastle-MailScanner: Found to be clean X-Newcastle-MailScanner-SpamScore: s As a matter of interest, can you give me an example of a model where this isn't needed? Mark. ----- Original Message ----- From: "Ian Robinson" To: Sent: Thursday, May 08, 2003 10:15 PM Subject: New issue for Additional Structures for OTS Revision Task Force > > > Ian Robinson, > Senior Technical Staff Member, > WebSphere Transactions Architecture & Development, > IBM Hursley Lab, UK > Tel +44-(0)1962-818626 Tie: 724-8626 > ian_robinson@uk.ibm.com > > ----- Forwarded by Ian Robinson/UK/IBM on 08/05/2003 22:15 ----- > |---------+----------------------------> > | | Alex Mulholland | > | | | > | | 07/05/2003 14:33 | > | | | > |---------+----------------------------> > >--------------------------------------------------------------------------- -------------------------------------| > | | > | To: issues@omg.org | > | cc: Ian Robinson/UK/IBM | > | From: Alex Mulholland/UK/IBM@IBMGB | > | Subject: New issue for Additional Structures for OTS Revision Task Force | > | | > >--------------------------------------------------------------------------- -------------------------------------| > > > > Hello, > > I would like to submit the following issue to the Additional Structures for > OTS Revision Task Force discussion list: > > Forcing child activities of completed parents to fail is too restrictive > > Section 2.2.9 of the specification, in the description of the compete > method, includes the following text: > "If the completion status is CompletionStatusFail, or > CompletionStatusFailOnly, > any encompassed active or suspended Activities will then have their > completion status set > to CompletionStatusFailOnly and transactions will be marked rollback_only." > > This behaviour is too restrictive for extended transaction models that may > wish to allow child > activities to complete successfully despite the fact that a parent activity > is failed. I suggest that > this restriction be removed. > > Dr Alex Mulholland > IBM Hursley Laboratory, UK. > amulholl@uk.ibm.com > > > > Subject: Re: New issue for Additional Structures for OTS Revision Task Force To: "Mark Little" Cc: "Ian Robinson" , ots-structs-rtf@omg.org X-Mailer: Lotus Notes Release 5.0.11 July 24, 2002 From: "Alex Mulholland" Date: Fri, 9 May 2003 15:09:17 +0100 X-MIMETrack: Serialize by Router on d06ml007/06/M/IBM(Release 5.0.9a |January 7, 2002) at 09/05/2003 15:09:00 Hello Mark, I understand that in an open nested transaction model, failure of a parent will ultimately cause failure of all the nested work, so failing nested child activities is the correct behaviour. However, I am considering a model where, although nesting is used, the outcome of a child activity is not necessarily influenced by the outcome of its parent activity. In such a model, if the parent activity times out then the child activity should still be able to complete successfully; at present it cannot. Does that make sense? Regards, Alex. Dr Alex Mulholland Activity Service and Extended Transactions, WebSphere Distributed Development, IBM Hursley Laboratory, UK. Tel internal: 246084; external: 01962 816084; Tel international: +44 1962 816084 amulholl@uk.ibm.com |---------+----------------------------> | | "Mark Little" | | | | | | | | | 05/09/2003 11:04 | | | AM | | | | |---------+----------------------------> >----------------------------------------------------------------------------------------------------------------------| | | | To: , Ian Robinson/UK/IBM@IBMGB | | cc: | | Subject: Re: New issue for Additional Structures for OTS Revision Task Force | | | >----------------------------------------------------------------------------------------------------------------------| As a matter of interest, can you give me an example of a model where this isn't needed? Mark. ----- Original Message ----- From: "Ian Robinson" To: Sent: Thursday, May 08, 2003 10:15 PM Subject: New issue for Additional Structures for OTS Revision Task Force > > > Ian Robinson, > Senior Technical Staff Member, > WebSphere Transactions Architecture & Development, > IBM Hursley Lab, UK > Tel +44-(0)1962-818626 Tie: 724-8626 > ian_robinson@uk.ibm.com > > ----- Forwarded by Ian Robinson/UK/IBM on 08/05/2003 22:15 ----- > |---------+----------------------------> > | | Alex Mulholland | > | | | > | | 07/05/2003 14:33 | > | | | > |---------+----------------------------> > > --------------------------------------------------------------------------- -------------------------------------| > | | > | To: issues@omg.org | > | cc: Ian Robinson/UK/IBM | > | From: Alex Mulholland/UK/IBM@IBMGB | > | Subject: New issue for Additional Structures for OTS Revision Task Force | > | | > > --------------------------------------------------------------------------- -------------------------------------| > > > > Hello, > > I would like to submit the following issue to the Additional Structures for > OTS Revision Task Force discussion list: > > Forcing child activities of completed parents to fail is too restrictive > > Section 2.2.9 of the specification, in the description of the compete > method, includes the following text: > "If the completion status is CompletionStatusFail, or > CompletionStatusFailOnly, > any encompassed active or suspended Activities will then have their > completion status set > to CompletionStatusFailOnly and transactions will be marked rollback_only." > > This behaviour is too restrictive for extended transaction models that may > wish to allow child > activities to complete successfully despite the fact that a parent activity > is failed. I suggest that > this restriction be removed. > > Dr Alex Mulholland > IBM Hursley Laboratory, UK. > amulholl@uk.ibm.com > > > > From: "Mark Little" To: "Alex Mulholland" Cc: "Ian Robinson" , Subject: Re: New issue for Additional Structures for OTS Revision Task Force Date: Fri, 9 May 2003 15:12:39 +0100 X-Mailer: Microsoft Outlook Express 6.00.2720.3000 X-Newcastle-MailScanner: Found to be clean X-Newcastle-MailScanner-SpamScore: ss > I understand that in an open nested transaction model, failure of a parent > will ultimately cause failure > of all the nested work, so failing nested child activities is the correct > behaviour. However, I am > considering a model where, although nesting is used, the outcome of a child > activity is not necessarily > influenced by the outcome of its parent activity. In such a model, if the > parent activity times out then > the child activity should still be able to complete successfully; at > present it cannot. Does that make sense? I understand what you're trying to achieve, I was really after an idea of what type of model requires this? It's not a traditional parent-child relationship and I am more interested in understanding what the underlying requirements are first and foremost. It's almost a reverse nested-transaction approach, so I'm intrigued. Mark. > > Regards, Alex. > > Dr Alex Mulholland > Activity Service and Extended Transactions, > WebSphere Distributed Development, > IBM Hursley Laboratory, UK. > Tel internal: 246084; external: 01962 816084; > Tel international: +44 1962 816084 > amulholl@uk.ibm.com > > > > |---------+----------------------------> > | | "Mark Little" | > | | | | c.uk> | > | | | > | | 05/09/2003 11:04 | > | | AM | > | | | > |---------+----------------------------> > >--------------------------------------------------------------------------- -------------------------------------------| > | | > | To: , Ian Robinson/UK/IBM@IBMGB | > | cc: | > | Subject: Re: New issue for Additional Structures for OTS Revision Task Force | > | | > >--------------------------------------------------------------------------- -------------------------------------------| > > > > > As a matter of interest, can you give me an example of a model where this > isn't needed? > > Mark. > > ----- Original Message ----- > From: "Ian Robinson" > To: > Sent: Thursday, May 08, 2003 10:15 PM > Subject: New issue for Additional Structures for OTS Revision Task Force > > > > > > > > Ian Robinson, > > Senior Technical Staff Member, > > WebSphere Transactions Architecture & Development, > > IBM Hursley Lab, UK > > Tel +44-(0)1962-818626 Tie: 724-8626 > > ian_robinson@uk.ibm.com > > > > ----- Forwarded by Ian Robinson/UK/IBM on 08/05/2003 22:15 ----- > > |---------+----------------------------> > > | | Alex Mulholland | > > | | | > > | | 07/05/2003 14:33 | > > | | | > > |---------+----------------------------> > > > > > -------------------------------------------------------------------------- - > -------------------------------------| > > | > | > > | To: issues@omg.org > | > > | cc: Ian Robinson/UK/IBM > | > > | From: Alex Mulholland/UK/IBM@IBMGB > | > > | Subject: New issue for Additional Structures for OTS Revision > Task Force | > > | > | > > > > > -------------------------------------------------------------------------- - > -------------------------------------| > > > > > > > > Hello, > > > > I would like to submit the following issue to the Additional Structures > for > > OTS Revision Task Force discussion list: > > > > Forcing child activities of completed parents to fail is too restrictive > > > > Section 2.2.9 of the specification, in the description of the compete > > method, includes the following text: > > "If the completion status is CompletionStatusFail, or > > CompletionStatusFailOnly, > > any encompassed active or suspended Activities will then have their > > completion status set > > to CompletionStatusFailOnly and transactions will be marked > rollback_only." > > > > This behaviour is too restrictive for extended transaction models that > may > > wish to allow child > > activities to complete successfully despite the fact that a parent > activity > > is failed. I suggest that > > this restriction be removed. > > > > Dr Alex Mulholland > > IBM Hursley Laboratory, UK. > > amulholl@uk.ibm.com > > > > > > > > > > > > > >