[
https://issues.apache.org/jira/browse/DELTASPIKE-1449?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michel Zedler updated DELTASPIKE-1449:
--------------------------------------
Description:
This exemplary job class
{code:java}
package my.package
import org.apache.deltaspike.scheduler.api.Scheduled;
import org.quartz.DisallowConcurrentExecution;
import org.quartz.Job;
@Scheduled(cronExpression = "...", overrideOnStartup = true)
@DisallowConcurrentExecution
public class MyJob implements Job {
...
} {code}
using deltaspike 1.9.0 results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1:
JOB_NAME = MyJob
JOB_CLASS_NAME = my.package.MyJob
IS_NONCONCURRENT = 1 (true){noformat}
using deltaspike 1.9.1 or later
+ a previously empty table results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1:
JOB_NAME = my.package.MyJob
JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
IS_NONCONCURRENT = 0 (false){noformat}
+ a non-empty table results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1: (inserted by a deployment using 1.9.0 or earlier)
JOB_NAME = MyJob
JOB_CLASS_NAME = my.package.MyJob
IS_NONCONCURRENT = 1 (true)
row 2:
JOB_NAME = my.package.MyJob
JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
IS_NONCONCURRENT = 0 (false){noformat}
I tracked this behaviour down to changes by DELTASPIKE-1369 that were released
with 1.9.1
causing:
(1) annotations org.quartz.DisallowConcurrentExecution and
org.quartz.PersistJobDataAfterExecution to be effectively ignored because
Quartz checks for their presence in class
org.apache.deltaspike.scheduler.impl.JobAdapter.
(2) consecutive deployments to produce duplicate job details.
I could not find any release notes, migration guides or open bugs or fixes
concerning this topic.
Expected behaviour:
* deltaspike/quartz should update existing JobDetails instead of inserting
duplicates
(JOB_CLASS_NAME of existing job details can be matched to new JOB_NAME - both
are the fully qualified class name)
* functionality of quartz annotations should be restored
I can provide more context if useful.
Are there any known workarounds?
was:
This exemplary job class
{code:java}
package my.package
import org.apache.deltaspike.scheduler.api.Scheduled;
import org.quartz.DisallowConcurrentExecution;
import org.quartz.Job;
@Scheduled(cronExpression = "...", overrideOnStartup = true)
@DisallowConcurrentExecution
public class MyJob implements Job {
...
} {code}
using deltaspike 1.9.0 results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1:
JOB_NAME = MyJob
JOB_CLASS_NAME = my.package.MyJob
IS_NONCONCURRENT = 1 (true){noformat}
using deltaspike 1.9.1 or later
+ a previously empty table results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1:
JOB_NAME = my.package.MyJob
JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
IS_NONCONCURRENT = 0 (false){noformat}
+ a non-empty table results in:
{noformat}
table: QRTZ_JOB_DETAILS
row 1: (inserted by a deployment using 1.9.0 or earlier)
JOB_NAME = MyJob
JOB_CLASS_NAME = my.package.MyJob
IS_NONCONCURRENT = 1 (true)
row 2:
JOB_NAME = my.package.MyJob
JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
IS_NONCONCURRENT = 0 (false){noformat}
I tracked this behaviour down to changes by DELTASPIKE-1369 that were released
with 1.9.1
likely causing:
(1) annotations org.quartz.DisallowConcurrentExecution and
org.quartz.PersistJobDataAfterExecution to be effectively ignored because
Quartz checks for their presence in class
org.apache.deltaspike.scheduler.impl.JobAdapter.
(2) consecutive deployments to produce duplicate jobs, triggers and concurrent
executions.
I could not find any release notes, migration guides or open bugs concerning
this topic.
Expected behaviour:
* deltaspike/quartz should update existing JobDetails instead of inserting
duplicates
(JOB_CLASS_NAME of existing job details can be matched to new JOB_NAME - both
are the fully qualified class name)
* functionality of quartz annotations should be restored
I can provide more context if useful.
Are there any known workarounds?
> Broken quartz annotations and duplicate job details
> ---------------------------------------------------
>
> Key: DELTASPIKE-1449
> URL: https://issues.apache.org/jira/browse/DELTASPIKE-1449
> Project: DeltaSpike
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: Scheduler
> Affects Versions: 1.9.1
> Reporter: Michel Zedler
> Priority: Major
>
> This exemplary job class
> {code:java}
> package my.package
> import org.apache.deltaspike.scheduler.api.Scheduled;
> import org.quartz.DisallowConcurrentExecution;
> import org.quartz.Job;
> @Scheduled(cronExpression = "...", overrideOnStartup = true)
> @DisallowConcurrentExecution
> public class MyJob implements Job {
> ...
> } {code}
> using deltaspike 1.9.0 results in:
> {noformat}
> table: QRTZ_JOB_DETAILS
> row 1:
> JOB_NAME = MyJob
> JOB_CLASS_NAME = my.package.MyJob
> IS_NONCONCURRENT = 1 (true){noformat}
> using deltaspike 1.9.1 or later
> + a previously empty table results in:
> {noformat}
> table: QRTZ_JOB_DETAILS
> row 1:
> JOB_NAME = my.package.MyJob
> JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
> IS_NONCONCURRENT = 0 (false){noformat}
> + a non-empty table results in:
> {noformat}
> table: QRTZ_JOB_DETAILS
> row 1: (inserted by a deployment using 1.9.0 or earlier)
> JOB_NAME = MyJob
> JOB_CLASS_NAME = my.package.MyJob
> IS_NONCONCURRENT = 1 (true)
> row 2:
> JOB_NAME = my.package.MyJob
> JOB_CLASS_NAME = org.apache.deltaspike.scheduler.impl.JobAdapter
> IS_NONCONCURRENT = 0 (false){noformat}
>
> I tracked this behaviour down to changes by DELTASPIKE-1369 that were
> released with 1.9.1
> causing:
> (1) annotations org.quartz.DisallowConcurrentExecution and
> org.quartz.PersistJobDataAfterExecution to be effectively ignored because
> Quartz checks for their presence in class
> org.apache.deltaspike.scheduler.impl.JobAdapter.
> (2) consecutive deployments to produce duplicate job details.
> I could not find any release notes, migration guides or open bugs or fixes
> concerning this topic.
> Expected behaviour:
> * deltaspike/quartz should update existing JobDetails instead of inserting
> duplicates
> (JOB_CLASS_NAME of existing job details can be matched to new JOB_NAME - both
> are the fully qualified class name)
> * functionality of quartz annotations should be restored
> I can provide more context if useful.
> Are there any known workarounds?
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
