9:25 a.m.
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public (Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it and
this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
1:19 p.m.
New subject: [JBoss JIRA] Commented: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn commented on JBTM-503:
----------------------------------
The comment 'TODO why?' is probably because the shutdown code has not yet been
tested. It is clear that this routine needs to be called by the RecoveryManagerImple and
not by your code -- the RecoveryManager implementation should clean up its own detritus.
However, after a cursory eyeball of the code I believe the current version of
ExpiredEntryMonitor needs to be supplemented with some synchronization. It should ensure
that the shutdown call does not clear _started and return before the thread held in
_theInstance has exited. If it does and then startup() is called it will be possible for
the old thread to start processing a new scanner list. Synchronization issues aside, it is
probably bad form for RecoveryManagerImple to return from stop when it has not really
stopped a thread which it started.
If you need to get this to work just now then I suggest you either
i) implement and test a version of shutdown which performs the required synchronization,
call it from RecoveryManagerimple.stop(), then send us your patch for inclusion in a
future release (we are very glad to accept contributions from those who know what they are
doing).
or
ii) bodge it as you indicated above and wait for us to release a tested version
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
1:22 p.m.
New subject: [JBoss JIRA] Assigned: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn reassigned JBTM-503:
--------------------------------
Assignee: Andrew Dinn
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
12:24 p.m.
New subject: [JBoss JIRA] Commented: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Mauro Molinari commented on JBTM-503:
-------------------------------------
Hi Andrew,
I was trying to work in this issue to provide you with a patch. I synchronized the
start&stop of ExpiredEntryMonitor but I have a problem understanding the lifecycle of
the RecoveryManagerImple.
It seems that:
- RecoveryManagerImple.start() is called just from
RecoveryManager.startRecoveryManagerThread(), for which I couldn't find any caller
- the ExpiredEntryMonitor is started through ExpiredEntryMonitor.startUp(), which in turn
is called by the constructor of RecoveryManagerImple
So, I can't find the correct place where to call ExpiredEntryMonitor.shutdown().
While things work in this way, I would say that it would be safe to call it from
RecoveryManagerImple.finalize(), however in this way I would have to explicitely call
RecoveryManagerImple.finalize() from my application if I want it to be completely shut
down.
If I add a call to ExpiredEntryMonitor.shutdown() into RecoveryManagerImple.stop(boolean),
then I would rather expect to find the call to ExpiredEntryMonitor.startUp() in
RecoveryManagerImple.start(), so that:
- the RecoveryManagerImple is started and stopped with start() and stop(boolean) by the
RecoveryManager
- the ExpiredEntryMonitor is started and stopped sith startUp() and shutdown() by the
RecoveryManagerImple
However, as it's not clear to me whether RecoveryManagerImple.start() is actually
called and when it is called, I'm not self-confident to move the call to
ExpiredEntryMonitor.startUp() from the constructor of RecoveryManagerImple to
RecoveryManagerImple.start()...
I would be grateful if you could help me to shed light on this.
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
5:19 p.m.
New subject: [JBoss JIRA] Commented: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn commented on JBTM-503:
----------------------------------
Hi Mauro,
RecoveryManager.startRecoveryManagerThread() is called from TransactionManager.start().
There are actually two versions of this class located in the jbossatx.jta and jbossatx.jts
packages. They provide the interface to the Application Server.
I had not noticed that the ExpiredEntryMonitor startup call is made in the constructor.
This is probably not the correct place for it to happen but I htink there is a reason for
it.
I'll note that the periodic recovery instance also gets created in the constructor
(and, if parameter threaded is true, started) whereas it probably ought at most to be
created there and then started and stopped by the recovery manager start and stop calls.
In the periodic recovery case this is done because in some circumstances the instance
needs to be created without being started. Recovery runs as a background thread in the App
Server but it can also be run via direct calls to scan(0 if the recovery manager is
started in a separate recovery process outside the AS. So, this code is dual purpose.
What I have not yet ascertained is why the ExpiredEnrtyMonitor thread is not subject to
the same condition. I think it probably needs to run in either case and since start does
not get called when the recovery manager is in its own process this is a cheap hack to
make sure it always runs.
Anyway, as far as shutting down is concerned I can make one very clear statement. It is
almost never appropriate to use a finalizer to clean up state. The only excuse for using
finalizers is when a library hands out a reference to an instance to client code and
cannot reasonably expect the client code to call a cleanup method to release resources
held by the instance (n.b.in my book any respectable client program ought to guarantee to
release any resources it allocates so this means a finalizer is never justified but
sometimes you cannot pick and choose your clients). The reason it is such a bad thing to
do is that finalizers make GC run like a pig. So it is best to avoid them at all costs.
Ok, that said, I believe it is currently valid to assume that when a recovery manager is
stopped it will not be restarted. Anyway calling start after stop wlll not work because
the start method will try to restart the periodic recovery instance and calling start on a
previously started thread is not valid. Anyway, given this assumption it makes no
difference whether the ExpiredEnrtryMonitor is started in the constructor or in method
start. So, the place to shut down the ExpiredEntryMonitor thread is in
RecoveryManagerImple.stop.
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
10:59 a.m.
New subject: [JBoss JIRA] Commented: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Mauro Molinari commented on JBTM-503:
-------------------------------------
Hi Andrew, I perfectly agree with you regarding the finalizers: I knew it was a very ugly
solution when I wrote that I would have needed to call the finalize() method from my
application. By the way, I saw it done in other places of JBossTS codebase ;-) (TxControl
calls finalize() on the transaction status manager on shutdown).
I hope I have understood your thoughts in the right way, so I'm going to attach the
patch I prepared.
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
10:59 a.m.
New subject: [JBoss JIRA] Updated: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Mauro Molinari updated JBTM-503:
--------------------------------
Attachment: JBossTS.patch
Hi Andrew,
this is the patch I prepared based on what we have discussed. I hope it can be a good
starting point for you.
Please note that I have left a TODO in ExpiredEntryMonitor.shutdown() because I think it
might be desirable to log the catching of an InterruptedException while waiting for the
completion of the monitor thread.
The patch is against 4.5.0GA codebase.
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
12:11 p.m.
New subject: [JBoss JIRA] Commented: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Mauro Molinari commented on JBTM-503:
-------------------------------------
Hi Andrew, any news on this?
Could it be considered for 4.7.0?
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5.0
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
12:15 p.m.
New subject: [JBoss JIRA] Updated: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn updated JBTM-503:
-----------------------------
Fix Version/s: 4.7.0
Might as well try to get this in the next release
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5.0
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Fix For: 4.7.0
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
4:15 p.m.
New subject: [JBoss JIRA] Updated: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn updated JBTM-503:
-----------------------------
Fix Version/s: 4.8.0
(was: 4.7.0)
Fixing this needs to take into account all of the other thread startup and shutdown
operations. Sometimes startup happens explicitly via startup methods, other times it
happens automatically on first call of an API method (or even, in one case at least,
inside a constructor for a related class). We really need to do startup in a more
structured way in order to know exactly what to shutdown and when. Puntiing this to the
next release until this larger problem is addressed
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5.0
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Fix For: 4.8.0
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
5:39 p.m.
New subject: [JBoss JIRA] Closed: (JBTM-503) Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
[
https://jira.jboss.org/jira/browse/JBTM-503?page=com.atlassian.jira.plugi...
]
Andrew Dinn closed JBTM-503.
----------------------------
Resolution: Done
Fixed by JBTM-569
Shutdown the ExpiredEntryMonitor when stopping the RecoveryManager
------------------------------------------------------------------
Key: JBTM-503
URL: https://jira.jboss.org/jira/browse/JBTM-503
Project: JBoss Transaction Manager
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Recovery
Affects Versions: 4.5.0
Environment: Tomcat + Spring + DBCP + JBossJTA
Reporter: Mauro Molinari
Assignee: Andrew Dinn
Fix For: 4.8.0
Attachments: JBossTS.patch
When the recovery manager is started, the implementation RecoveryManagerImple starts the
ExpiredEntryMonitor.
When you stop the recovery manager, however, the ExpiredEntryMonitor is not shut down.
On line 250 of com.arjuna.ats.internal.arjuna.recovery.RecoveryManagerImple there's
the line ExpiredEntryMonitor.shutdown() commented out, with a "TODO why?"
comment immediately before.
I don't know exactly what should be the lifecylce of ExpiredEntryMonitor versus that
of RecoveryManagerImple, however I think that one of the following should happen:
- ExpiredEntryMonitor is started when RecoveryManagerImple is started (rather than
constructed), so that it is shut down when the RecoveryManagerImple is stopped
OTHERWISE
- RecoveryManagerImple provides a way to "dispose" or "disable" it
and this includes the shutdown of ExpiredEntryMonitor
OTHERWISE
- it is documented that stopping the recovery manager needs the ExpiredEntryMonitor to be
stopped too: however, please consider that
com.arjuna.ats.internal.arjuna.recovery.ExpiredEntryMonitor.shutdown() throws a
NullPointerException if called while the ExpiredEntryMonitor has not been started before
and there is no public method to test if it has been started or not; so if I need to write
some code that stops the ExpiredEntryMonitor (without knowing if it has been started or
not) I have to write some ugly code like:
{code:java}
try
{
ExpiredEntryMonitor.shutdown();
}
catch(final NullPointerException e)
{
// ExpiredEntryMonitor had not been started
}
{code}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
5406
days inactive
5524
days old
jbossts-issues@lists.jboss.org
10 comments
2 participants
participants (2)
-
Andrew Dinn (JIRA) -
Mauro Molinari (JIRA)