[ 
https://issues.apache.org/jira/browse/QPID-8091?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16358416#comment-16358416
 ] 

ASF subversion and git services commented on QPID-8091:
-------------------------------------------------------

Commit 63c315f07553dcdf32e2de1888f1cb9749e15d5c in qpid-broker-j's branch 
refs/heads/master from [~k-wall]
[ https://git-wip-us.apache.org/repos/asf?p=qpid-broker-j.git;h=63c315f ]

QPID-8091: [Broker-J] Update transaction timeout chapter in docbook

* Remove note regarding AMQP 1.0
* Generalise from 'producer transaction timeout' to 'transaction timeout'.  The 
former was only true when using the Qpid JMS AMQP 0-x client
  (which delayed acking the messages until the application called commit).
* Update the operational log messages


> [Broker-J] [AMQP 1.0] Store transaction timeout feature
> -------------------------------------------------------
>
>                 Key: QPID-8091
>                 URL: https://issues.apache.org/jira/browse/QPID-8091
>             Project: Qpid
>          Issue Type: Improvement
>          Components: Broker-J
>            Reporter: Keith Wall
>            Assignee: Keith Wall
>            Priority: Major
>             Fix For: qpid-java-broker-7.1.0
>
>
> Berkeley JE's design means that once a transaction has begun, its internal 
> cleaner is unable to clean beyond the point the transaction started in the 
> transaction log.  Other transaction work continues as normal, but the disk 
> space utilisation can't shrink until the long running transaction is 
> completed.  In an extreme case, disk space can be exhausted.
> This has consequence for long running store transactions in Qpid.  For 0-x, 
> the transaction timeout features allows the the length of time a transaction 
> may be open or idle to be constrained, thus limiting the harmful effects of a 
> long running store transaction.
> For robustness, that features needs to be implemented on AMQP 1.0 too.   A 
> decision needs to be made about the correct course of action to be taken when 
> a long running transaction exceeds the threshold.   The transaction 
> coordinator link could be closed with an error or the entire connection 
> closed.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org

Reply via email to