This is still part of ARTEMIS-1776 fix, which still part of the same release as we are on now.
Hence I'm not opening a new JIRA for this one.
Cherry-picked from e5bce13316f7e81bb15a12592622df2ea2632a35
The cluster connection bridge hard codes its producerWindowSize to -1
(meaning no producer flow control) even if you configure it otherwise.
Cherry-picked from 98ce31bf588a314a0408216e4ad4b85597a15186
* initialize ActiveMQActivationSpec#shareSubscriptions to false
* i18n of the exception thrown by ActiveMQMessageHandler if it is not
possible to create a non-shared subscriber
JIRA: https://issues.apache.org/jira/browse/ARTEMIS-1475
With NFSv4 it is now necessary to lock/unlock the byte of the server
lock file where the state information is written so that the
information is then flushed to the other clients looking at the file.
(cherry picked from commit 2ec173bc708daca163c9356cf12440432abc61c4)
There is a leak on replication tokens in the moment when a backup is
shutdowned or killed and the ReplicationManager is stopped. If there
are some tasks (holding replication tokens) in the executor, these
tokens are simply ignored and replicationDone method isn't called on
them. Because of this, some tasks in OperationContextImpl cannot be
finished.
(cherry picked from commit 88a018e17fd49097de1186c65e25cd0af578b6a9)
(cherry picked from commit d6cbc0aa885fa88beb7f1b2450cdfe4da9466947)
This is based on the work @jbertram made at the github pr #1466 and the discussions we had there
(cherry picked from commit ce6942a9aa9375efaa449424fe89de2db3f22e36)
When a large message is replicated to backup, a pendingID is generated
when the large message is finished. This pendingID is generated by a
BatchingIDGenerator at backup.
It is possible that a pendingID generated at backup may be a duplicate
to an ID generated at live server.
This can cause a problem when a large message with a messageID that is
the same as another largemessage's pendingID is replicated and stored
in the backup's journal, and then a deleteRecord for the pendingID
is appended. If backup becomes live and loads the journal, it will
drop the large message add record because there is a deleteRecord of
the same ID (even though it is a pendingID of another message).
As a result the expecting client will never get this large message.
So in summary, the root cause is that the pendingIDs for large
messages are generated at backup while backup is not alive.
The solution to this is that instead of the backup generating
the pendingID, we make them all be generated in advance
at live server and let them replicated to backup whereever needed.
The ID generater at backup only works when backup becomes live
(when it is properly initialized from journal).
(cherry picked from commit d50f577cd50df37634f592db65200861fe3e13d3)
