Thanks Romain for clarification.
Did the quartz handling switched to outside the transaction with
introduction of the "shaded" quartz access?
--
View this message in context:
http://tomee-openejb.979440.n4.nabble.com/TomEE-7-0-2-increasing-quartz-tx-database-connections-tp4681007p4681139.html
Hi Romain!
We configured quartz for using JDBC-JobStoreCMT (see
http://www.quartz-scheduler.org/documentation/quartz-2.2.x/configuration/ConfigJobStoreCMT.html).
It is suggested to be used if "applications are using JTA transactions (such
as via EJB Session Beans) to perform their work."
To my
Hi Romain!
We want to share what we found out so far.
We added some log messages in both versions (7.0.1 & 7.0.2) of
org.apache.openejb.resource.jdbc.managed.local.ManagedConnection.java recording
connection create & close.
What we see is that while transaction completion the established
Hi Romain!
Sure.
We try to reproduce issue without our complete application.
As soon as we identified the root cause we can write an according datasource
test.
Is there a planned release date for upcoming TomEE 7.0.3?
Best regards,
Thomas
--
View this message in context:
We tried latest TomEE 7.0.3 snapshot - same issue (increasing quartz tx
database connections).
What might be worth to know for the log files provided before: We use
several connection pools.
Connections 49 & 50 are related to a non-quartz connection pool.
Connections 28, 29 & 30 (the ones that are
/url?u=https-3A__javaeefactory-2Drmannibucau.rhcloud.com=DwMCAg=ZgVRmm3mf2P1-XDAyDsu4A=tcnVZpsm--xbG4A9tLhh_CVv_TOmzYWmS-P4lauv47I=t0LrDDhqtCTu9Wby2YU3LgQbr2QUrGN8RO1mdlo4CNs=nmP7DqgVP__ao1cpigRcoNgw2MZaFaxDcrm1u4OK2FM=>>
2017-02-13 10:17 GMT+01:00 tschuler <[hidden
email]>:
> Hi Romai
ucau.rhcloud.com=DwMCAg=ZgVRmm3mf2P1-XDAyDsu4A=tcnVZpsm--xbG4A9tLhh_CVv_TOmzYWmS-P4lauv47I=j5PcFhN5wNkiZXRRBmt6aj2_aWlwIXOID5aHgTDLY2c=2MgU1N170sWgWkYfKyl-y7U4UMLxBgQLLNYy9JpKVRU=>>
2017-02-10 15:43 GMT+01:00 tschuler <[hidden
email]>:
> Hi Romain!
>
> We actually cannot reproduce the issue
Hi Romain!
We actually cannot reproduce the issue within a small example.
But what we did was "enhancing" ManagedConnection.java with log messages in
case of connection is created or closed.
Attached text file (startAgent2.log) includes the additional log messages
showing that three
We did some investigations:
The problem (increasing amount of database connections to quartz database)
still occurs if tomcat-jdbc.jar is replaced with same file included in TomEE
7.0.1.
On the other hand we replaced the ManagedConnection*.class files in
openejb-core-7.0.2.jar with the ones
Hi Romain!
I switched the quartz related datasource to dbcp (DataSourceCreator = dbcp).
Now the database connections do not increase any more.
Does this mean for sure that cause is a change in tomcat jdbc
implementation?
Thanks,
Thomas
--
View this message in context:
Hi Romain!
I'm not sure if I get you right: Changes were done but should not have in
impact, right?
Can I somehow revert these changes to get sure if or if not they are cause
for the issue?
Best regards,
Thomas
--
View this message in context:
Hi Romain!
Database configuration is unchanged (same as for TomEE 7.0.1) - is there a
configuration parameter that changed its default value since TomEE 7.0.1?
Thread count is stable, no deadlocks, thread states are same as for TomEE
7.0.1.
Best regards,
Thomas
--
View this message in
Hi!
Our application is using scheduled beans.
The according timers are persisted within quartz tables.
This works as expected using TomEE 7.0.1 (webprofile).
But within TomEE 7.0.2 (also webprofile) we get in trouble because tx
database connections to quartz tables are not released any more.
It
Hi!
We are using TomEE 1.7.2 with message driven beans and ActiveMQ 5.11.1.
In general there are about thirty threads that handles the message driven
bean activities (JmsResourceAdapter-worker- - 1, ...,
JmsResourceAdapter-worker- - 30).
But from time to time we can see that every mdb execution
to something = 0?
Romain Manni-Bucau
@rmannibucau https://twitter.com/rmannibucau | Blog
http://rmannibucau.wordpress.com | Github https://github.com/rmannibucau |
LinkedIn https://www.linkedin.com/in/rmannibucau | Tomitriber
http://www.tomitribe.com
2015-08-05 8:49 GMT+02:00 tschuler [hidden
email
Manni-Bucau
@rmannibucau https://twitter.com/rmannibucau | Blog
http://rmannibucau.wordpress.com | Github https://github.com/rmannibucau |
LinkedIn https://www.linkedin.com/in/rmannibucau | Tomitriber
http://www.tomitribe.com
2015-08-05 9:13 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type
Hi Romain!
We did not observe thread deadlocks (or any other classloading issue) with
TomEE 1.7.2 snapshot from yesterday (21st April).
Best regards,
Thomas
--
View this message in context:
Hi Romain!
The suggested workaround preloading the locking classes in a
ServletContextListener might work, but:
The thread deadlock happens only from time to time (about every 5th day).
Therefore identifying the locking classes is not possible as we observe a
different one every time.
://www.tomitribe.com
2015-04-20 15:00 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4674516i=0:
Hi Romain!
The suggested workaround preloading the locking classes in a
ServletContextListener might work, but:
The thread deadlock happens only from time to time (about every 5th day
Hi Romain!
The example was added as attachment and is available for download.
BTW: Is there a bug tracking system (like JIRA) for TomEE?
Did you already add an according bug so that I can keep an eye on it?
Best regards,
Thomas
--
View this message in context:
-13 11:10 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4674373i=0:
Hi Romain!
The example was added as attachment and is available for download.
BTW: Is there a bug tracking system (like JIRA) for TomEE?
Did you already add an according bug so that I can keep an eye
https://github.com/rmannibucau |
LinkedIn https://www.linkedin.com/in/rmannibucau | Tomitriber
http://www.tomitribe.com
2015-04-13 11:31 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4674375i=0:
Hi Romain!
The example is still available:
http://tomee-openejb.979440.n4
TomEE OpenEJB]
[mailto:ml-node+s979440n4674377...@n4.nabble.com]
Sent: Montag, 13. April 2015 14:46
To: Thomas Schuler
Subject: Re: thread deadlock at URLClassLoaderFirst
2015-04-13 14:42 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4674377i=0:
Hi Romain!
Yes
http://rmannibucau.wordpress.com | Github https://github.com/rmannibucau |
LinkedIn https://www.linkedin.com/in/rmannibucau | Tomitriber
http://www.tomitribe.com
2015-03-30 11:03 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4674223i=0:
Hi Romain!
Can you please provide
@rmannibucau https://twitter.com/rmannibucau | Blog
http://rmannibucau.wordpress.com | Github https://github.com/rmannibucau |
LinkedIn https://www.linkedin.com/in/rmannibucau | Tomitriber
http://www.tomitribe.com
2015-03-27 15:59 GMT+01:00 tschuler [hidden
email]/user/SendEmail.jtp?type
Hi!
We use TomEE 1.7.1 and get in troubles from time to time because of a thread
dead lock resulting in an idle TomEE. It is always the same pattern as shown
below - deadlock at org.apache.openejb.util.classloader.URLClassLoaderFirst.
Extract from JVisualVM thread dump:
Found one Java-level
Hi!
We use TomEE 1.6.0.2.
We have a couple of scheduled beans and provide an own
application.properties (with scheduler MyScheduler) to get the according
quartz jobs persisted.
As we can see (asking the org.quartz.impl.StdSchedulerFactory) that both
MyScheduler and the TomEE default scheduler
: http://rmannibucau.wordpress.com/
LinkedIn: http://fr.linkedin.com/in/rmannibucau
Github: https://github.com/rmannibucau
2014-09-04 15:53 GMT+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4671626i=0:
Hi!
We use TomEE 1.6.0.2.
We have a couple of scheduled beans and provide
+02:00 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4671631i=0:
Hi Romain!
Can you provide a simple example how to add the quartz properties as
properties for the ejb-deployment?
Best regards,
Thomas
Von: Romain Manni-Bucau [via TomEE OpenEJB] [mailto:[hidden
Hi,
we have a TomEE Webprofile 1.6.0 where we have ActiveMQ 5.9 integrated.
We have a simple application which is deployed as EAR file.
In this application we have a JMS queue associated with a MDB (Message
Driven Bean).
We also have a stateless session bean for sending messages to the JMS
: Montag, 24. März 2014 15:40
An: Thomas Schuler
Betreff: Re: Transaction issue for MDB in cluster environment
Is it possible that you are running multiple tomee's on one server/machine,
and using default activemq config?
On Mon, Mar 24, 2014 at 9:54 AM, tschuler [hidden
email]/user
Hi Romain!
We use TomEE 1.6.0 release and still got the problem that the broker shuts
down after a few days.
Is there a way to programmatically restart the broker if it is down using
the configuration in tomee.xml?
Actually we observer the broker using BrokerService retrieved from
No, we do not use JMX.
As described at the very beginning of this topic:
If the ActiveMQ database locker cannot update database lock, the broker
shuts down.
So TomEE is running with stopped ActiveMQ broker.
To avoid this we try to implement a monitoring thread that checks from time
to time if
Hi!
We use TomEE 1.6.0.
We are facing a serious problem in URLClassLoaderFirst as threads get
blocked:
Blocking thread:
---
http-bio-28080-exec-3 - Thread t@795
java.lang.Thread.State: RUNNABLE
at java.io.WinNTFileSystem.getBooleanAttributes(Native Method)
at
Hi!
I'm facing the problem that topic MDBs in a TomEE cluster both get same
(activeMQ) clientid (the class name of the MDB I suppose).
I know of MDB activation config property client id.
If it is set to a specific value for a MDB, the jar containing the MDB
cannot be shared among cluster nodes.
Trying around a bit shows that a LockTimeoutException (that has no effect on
the transaction state) is only thrown if:
- entity manager find (entityManager.find(LockEntity.class, lockId,
LockModeType.PESSIMISTIC_WRITE)) is used.
- and the entity to be locked was not loaded before within the
Hi!
We are using TomEE 1.6.0 snapshot and are actually working on usage of
pessimistic locks for an entity (entityManager.lock(entity,
LockModeType.PESSIMISTIC_WRITE)).
What we see is that if the entity is allready locked the next locking
attempt results in a PessimisticLockException marking the
://fr.linkedin.com/in/rmannibucau
Github: https://github.com/rmannibucau
2013/10/23 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4665711i=0:
Hi!
We are using TomEE 1.6.0 snapshot and are actually working on usage of
pessimistic locks for an entity (entityManager.lock(entity
Hi Romain!
I was able to get the scheduled beans running with oracle database
persisting.
First approach was to deactivate oracle optimization handling disk space for
empty tables (DEFERRED_SEGMENT_CREATION).
This fixes the problem only partially - tomee startup was successful but
adding cron
/10/16 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4665604i=0
Hi!
We are using latest TomEE 1.6.0 snapshot and try to persist scheduled beans
in an oracle database.
The TomEE startup fails because the default scheduler cannot be
initialized.
Attached zip
ringing to anyone.
*Romain Manni-Bucau*
*Twitter: @rmannibucau https://twitter.com/rmannibucau*
*Blog: **http://rmannibucau.wordpress.com/*http://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/10/17 tschuler [hidden
email
tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4665497i=0
Hi!
We are using latest TomEE 1.6.0 snapshot and an oracle database for storing
entity beans.
What we see are warning messages like:
java.lang.IllegalAccessException: Class
org.apache.openjpa.jdbc.sql.OracleDictionary can
: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/10/9 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4665509i=0
Hi Romain!
Do you have an example for synchronizing a scheduled bean with a custom
event?
Best regards,
Thomas
Von
Hi Romain!
Please have a look at:
http://activemq.2283324.n4.nabble.com/Avoid-broker-stop-within-TomEE-tp4671531.html
and
https://issues.apache.org/jira/browse/AMQ-4526
Is there a chance to get ActiveMQ 5.9.0 included in TomEE 1.6.0?
Or is an additional ActiveMQ configuration parameter (in this
Hi!
From time to time (TomEE may run several days without any problems) the
activeMQ broker stops - following log messages can be found in catalina.log:
Sep 9, 2013 1:38:55 PM org.apache.activemq.store.jdbc.DefaultDatabaseLocker
keepAlive
SEVERE: Failed to update database lock:
*
2013/9/12 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4665061i=0
Hi!
From time to time (TomEE may run several days without any problems) the
activeMQ broker stops - following log messages can be found in
catalina.log:
Sep 9, 2013 1:38:55 PM
Hi Romain!
I already have a test suite (scheduled bean basics with and without
persistence) and tried to run it with the actual 1.6.0 snapshot.
What I see is that the used quartz version is now 2.2.0 (before: 2.1.6).
Before I adapt the test suite: Is quartz 2.2.0 only temporary used or will it
Hi Romain!
I don't think that it is related to missing datastore implementation because
it worked with a previous TomEE 1.6.0 snapshot (e.g. 5th august).
Best regards,
Thomas
--
View this message in context:
Hi Romain!
As I am neither familiar with the application composer and the in memory
database:
Can you provide me a running example how to initialize and use the in memory
database for tests?
Best regards,
Thomas
--
View this message in context:
Hi Romain!
If only one webapp is deployed, the exception is not shown. It doesn't
matter which webapp is used.
As soon as at least two webapps are deployed, the exception occurs. It
doesn't matter which webapps are included and start order of the webapps
also has no influence.
Best regards,
in src/test/resources a file import-managed.sql with a statement
by line created from hsql scripts of quartz
Le 3 sept. 2013 09:33, tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664964i=0 a écrit :
Hi Romain!
As I am neither familiar with the application composer and the in memory
Hi Romain!
We observed the messages since using a TomEE 1.6.0 snapshot (I'm not sure if
it happens with TomEE 1.5.2 too), so it has nothing to do with recent
changes.
Providing an example is actually not possible, because it happens within a
very complex application I'm not allowed to share.
I
. Finally add a
@Configuration method returning a Properties with a datasource defined with
properties syntax.
Le 2 sept. 2013 16:37, tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664949i=0 a écrit :
Hi Romain!
Can you reproduce the error I described using persisted quartz jobs?
Can you
/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/8/5 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664531i=0
Hi Romain!
No, it's not the same transaction.
If test client ends (without canceling the calendar timer) and is started
.
*Romain Manni-Bucau*
*Twitter: @rmannibucau https://twitter.com/rmannibucau*
*Blog: **http://rmannibucau.wordpress.com/*http://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/8/2 tschuler [hidden
email]/user
://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/8/5 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664523i=0
Hi Romain!
I have to correct me:
The calendar timer is removed after successful firing only if no quartz
persistence is configured.
It still
*
*Twitter: @rmannibucau https://twitter.com/rmannibucau*
*Blog: **http://rmannibucau.wordpress.com/*http://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/8/5 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664525i=0
/class/project?
*Romain Manni-Bucau*
*Twitter: @rmannibucau https://twitter.com/rmannibucau*
*Blog: **http://rmannibucau.wordpress.com/*http://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/8/5 tschuler [hidden
email
/rmannibucau*
2013/8/5 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664529i=0
Hi Romain!
Here is an example:
- timerTests.zip: A web application including only the scheduled
bean to test - simply put it to %TOMEE_HOME%/webapps.
- timerTests.src.zip: Includes
Hi!
We are doing some tests on TomEE 1.6.0 snapshot.
We use a scheduled bean that gets fired by a calendar timer
(timerService.createCalendarTimer()).
After ejbTimeout happens, the according timer is still available
(timerService.getTimers()).
This is the case only if quartz persistence is
://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/7/26 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664374i=0
Hi Romain!
We define the scheduler using an application.properties file included in a
jar file
Hi!
We use a scheduled bean, the added timers are persisted in a database table.
After server restart the timers still work and the scheduled bean is
triggered as expected.
But we get no access to the timers any more and therefore cannot cancel
them.
The injected TimerService on the scheduled
Hi Romain!
I tried again with actual TomEE 1.6.0 snapshot.
But context retrieved using new InitialContext() does not contain anything.
Only if context is used with initial context property set to
org.apache.openejb.core.LocalInitialContextFactory the ejbs are available as
expected.
Best regards,
with lib and which
is not an EJB
*Romain Manni-Bucau*
*Twitter: @rmannibucau https://twitter.com/rmannibucau*
*Blog: **http://rmannibucau.wordpress.com/*http://rmannibucau.wordpress.com/
*LinkedIn: **http://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/7/29 tschuler
Hi!
We are using the TomEE 1.6.0 snapshot from yesterday.
We have scheduled beans (e.g. JobSchedulerBean) that handle persistent
timers.
While server is starting (and a timeout happened while server was down) the
ejbTimeout happens before according scheduler bean is registered and the
trigger
Hi Romain!
We define the scheduler using an application.properties file included in a
jar file that itself is part of the ear.
Best regards,
Thomas
--
View this message in context:
http://openejb.979440.n4.nabble.com/EjbTimeout-before-scheduler-bean-available-tp4664366p4664373.html
Sent from
Hi!
We have an ear file containing a few war files.
What we see is that wars are started depending on the context root defined
in the application.xml.
Is there a way to get the web applications started in another order? The
initialize-in-order tag defined in application_6.xsd seems not to work.
09:45, tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664151i=0 a écrit :
Hi!
We want to programmatically get the TomEE version.
Is there a class that can be asked for it?
Best regards,
Thomas
--
View this message in context:
http://openejb.979440.n4.nabble.com
://fr.linkedin.com/in/rmannibucau*
*Github: https://github.com/rmannibucau*
2013/7/10 tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664197i=0
Hi Romain!
Thanks.
So there is no way to get the TomEE version itself?
If we want to ensure that, e.g. TomEE 1.6.0 is installed, we have to check
Betreff: Re: Use specific jar version in war
Hi
If not a server jar (these ones can be handled particularly) that's the
case on trunk. You can force it with a system prop but you shouldnt need it
Le 9 juil. 2013 09:36, tschuler [hidden
email]/user/SendEmail.jtp?type=nodenode=4664153i=0 a écrit :
Hi
Hi!
We deploy an ear file containing a specific version of a jar file in the
ear-lib directory (actually using a TomEE 1.6.0 snapshot).
A war included in the ear needs a different version of the same jar file.
Is there a way to separate the war classpath from the ear classpath or is it
possibleto
Hi Romain!
As we are facing the same issue within TomEE 1.5.1:
Will the fix also be available in upcoming TomEE 1.5.2?
Best regards,
Thomas
--
View this message in context:
http://openejb.979440.n4.nabble.com/ActiveMQ-1-5-1-issue-with-an-ActiveMQ-embedded-broker-tp4658482p4661743.html
Sent
Hi Romain!
Putting log4j.jar only into lib folder of the according web applications is
not enough:
The defined log files are just created but all log messages are written into
catalina.log.
Best regards,
Thomas
--
View this message in context:
73 matches
Mail list logo