|
Problem(Abstract) |
The IBM® WebSphere® Application Server scheduler cancel,
purge, suspend and resume API methods or MBean operations may throw an
incorrect TaskInvalid exception if the task is running or is about to
run. |
|
Cause |
There is a small window in which a task may return a
com.ibm.websphere.scheduler.TaskInvalid Exception incorrectly.
This is caused by some database platforms not locking a row using
READ_COMMITTED transaction isolation unless the data in the row changes.
The iSeries® DB2® database has the following behaviors which increase
the likelihood of this error:
- iSeries may ignore the READ_COMMITTED transaction isolation setting on
the connection and may instead use READ_UNCOMMITTED.
- iSeries prevents row locks when SELECT... FOR UPDATE is used with
READ_COMMITTED transaction isolation where the result set is not updated.
Because the scheduler uses this mechanism to lock the row prior to
updating, This may result in the unexpected TaskInvalid exception.
|
|
Resolving the
problem |
1. Apply iSeries fixes:
- V5R2M0 PTF - MF34388
- V5R3M0 PTF - MF34383
2. Apply an interim fix or service pack that contains scheduler APAR
PK03786. This fix is available in the following service releases:
- Scheduler
cumulative interim fix 1 for WebSphere Enterprise 5.0.2.6
- WebSphere Business Integration Server Foundation version
5.1.1.2
- WebSphere Application Server 6.0.1.2 (All editions)
- WebSphere Application Server 6.0.2 (All editions)
Workaround:
Catch the TaskInvalid exception and submit the cancel, purge, suspend or
resume request again. |
|
|
|
Cross Reference information |
Segment |
Product |
Component |
Platform |
Version |
Edition |
Application Servers |
Runtimes for Java Technology |
Java SDK |
|
|
|
|
|
|
|