The default id-cache-size is 20000 and the default
confirmation-window-size is 1MB. It turns out the 1MB
size is too small for id-cache-size.
To fix it we adjust the confirmation-window-size to 10MB. Also
a test is added to guarantee it won't break this rule when this
default value is to be changed to any new value.
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).
It fixes compatibility issues with JMS Core clients using the old address model, allowing the client to query JMS temporary queues too.
you would eventually see this issue when using older clients:
AMQ119019: Queue already exists
This method name would clash with ServiceComponent
As the real meaning here on this method is just to failover
So I've renamed the method to avoid the clash with my next commit
(I've done this on a separate commit as you may need to redo this
commit from scratch again in other branches instead of lots of clashes on cherry-pick)
When a large message is being diverted, a new copy of the original
message is created and replicated (if there is a backup) to the backup.
In LargeServerMessageImpl.copy(long) it reuse a byte array to copy
message body. It is possible that one block of date is read into
the byte array before the previous read has been replicated,
causing the replicated bytes to corrupt.
If we make a copy of the byte array before replication, the corruption
of data will be avoided.
The JDBCSequentialFile blocks on the writeLock when opening. There is
no need to block here, in fact it may cause issues when opening and
syncing concurrently. Instead an AtomicBoolean is enough to prevent the
file from being reloaded.
Before sending of messages to server 0 begins, the test
should wait until consumer is registered at RemoteQueueBindingImpl
on server 0. Otherwise some messages may not be rebalanced
to server 1.
it doesn't really matter the number of files.. as long as the data is valid.
This type of assertion limits the implementation. it's mocking test with too much intrusion
over the implementation. Hence I'm removing these clauses that will fail eventually.
Add extra configuration to address-settings to be able to
control / enable address/queue deletion by pattern,
rather than a global toggle.
Add support in the reload logic to remove address
and/or queues if the address matches an address setting,
where it is enabled.
Armand Roelens
Clebert Suconic
Howard Gao
Dmitrii Tikhomirov
Josh Reagan
Francesco Nigro
Christopher L. Shannon (cshannon)
Justin Bertram
Martyn Taylor
Erich Duda
Guillaume Nodet
Michael Andre Pearce
Jiri Danek
Michal Tóth
Andy Taylor