Workflow events are not working properly
You need to stop and restart Product Master when workflow events do not seem to be working.
Symptoms
Sometimes reserving an item in a workflow step by clicking Reserve does not succeed. Even clicking Refresh multiple times does not show the item as being reserved. The reservation of items in workflows is processed by the workflow engine. Therefore, the logs for the workflow engine must be checked. The logs are located in the$TOP/logs/workflowengine/ folder.Causes
If there are exceptions in the log files with time stamps close to the time when the Reserve was clicked, then search for Technotes describing such messages, otherwise open a ticket with Support.Diagnosing the problem
When most of the workflow engine log files (especially the ipm.log) have not been updated recently (check with ls -l), then this might be an indication that the workflow engine has stopped running. In this case, you also would see an increasing backlog of new workflow events when triggering new workflow-related actions like checkout items or reserving items. This can be checked by running following query, for example, from within the DB Admin console (click System Administrator > DB Admin):select count(*) from wfe where wfe_event_status ='NEW';
Resolving the problem
If there is indication that workflow engine is no longer responding, follow these steps:- Stop all Product Master services:
$TOP/bin/go/stop_local.sh
- Make sure that there are no other processes running by entering the following command:
The pattern of related Java™ processes has something like the following:
ps –ealf | grep <user>
For example:
AIX
Linuxjonas13 471166 1 0 Jun 15 - 3:59 /opt/WebSphere/AppServer/java/bin/java - DTOP=/opt/wpc/envs/jonas13/wpc_532_IF11_DB2 -DCCD_ETC_DIR=/opt/wpc/envs/jonas13/wpc_532_IF11_DB2/etc ... -Dprofiler_opts=__ com.ibm.ccd.workflow.common.WorkflowEngine force
If you find processes like shown preceding step, you need to end those by using the kill command. The process IDs to use are in the preceding examples. The syntax for the kill command is:0 S stan9 19261 1 0 75 0 - 298022 schedu Jun03 ? 00:43:49 /opt/WebSphere/AppServer6/java/bin/java - DTOP=/wpc/envs/stan9/wpc_5328_DB2 -DCCD_ETC_DIR=/wpc/envs/stan9/wpc_5328_DB2/etc -Dfile.encoding=ISO8859_1 -classpathwpc/envs/stan9/wpc_5328_DB2/jars/ccd_svr.jar:/wpc/envs/stan9/wpc_5 328_DB2/jars/ccd
for example:kill -9 <stale process id> <stale process id> ...
kill -9 471166 19261
- Restart all Product Master services:
$TOP/bin/go/start_local.sh
- Start the user interface and reserve an item in any workflow. This time the item should be reserved successfully, showing the reserve sign. However, if the problem is not solved, contact IBM® Software Support.
Stopping and restarting Product Master
To ensure that Product Master starts correctly, first stop, then restart Product Master by using the following steps.
Procedure
- Attempt to gently stop Product Master by running the
stop_local.sh
script:$TOP/bin/go/stop_local.sh - Wait for approximately one minute, then run the following
command:ps -u $USERNAME
- If there are any active Java processes, a scheduled job
might still be in progress. You can let the job complete or you can stop it manually by using the
abort_local.sh
script:$TOP/bin/go/abort_local.sh - Wait for approximately thirty seconds, then run the following
command:ps -u $USERNAME
- If there continues to be active Java processes, the JVM might have crashed. The Java processes must be manually stopped by running the following command:kill -9 `ps -u $USERNAME | grep java | cut -b1-5
- If any Java processes still exist, restart the system.
- Once all Java processes have been stopped, restart the Product Master using the
start_local.sh
script:$TOP/bin/go/start_local.sh - After the
start_local.sh
script completes, run thermi_status.sh
script to verify that all services return a status and confirm that Product Master has started correctly:$TOP/bin/go/rmi_status.sh - Open a browser and ensure that you can log into Product Master.