New in 18.0.2 Some IBM® Business Automation
Insights capabilities can be disabled only
after you create savepoints on each processing job.
About this task
You can update IBM Business Automation
Insights to disable
such capabilities as HDFS storage or sending data to the Kafka egress topic but you must first
create savepoints of all running processing jobs to be able to restart them correctly after the
update.
Procedure
- Create savepoints for all the running processing jobs by using the script provided in the job
manager pod.
kubectl exec -it <my-release>-bai-flink-jobmanager-<id> -- scripts/create-savepoints.sh -s
- Take note of each savepoint path. In step 3, you replace
savepoint_id with the corresponding identifier values.
- /mnt/pv/savepoints/dba/bai-bpmn/savepoint-<savepoint-id>
- /mnt/pv/savepoints/dba/bai-ingestion/savepoint-<savepoint-id>
- /mnt/pv/savepoints/dba/bai-icm/savepoint-<savepoint-id>
Note: If the
create-savepoints.sh script returns an error such as
Not
able to create savepoint for job 'dba/bai-<job-name>' with ID:
xxxxxxxxxxx., fix the error before you update the
IBM Business Automation
Insights release. For more information, see
Troubleshooting.
- Prepare the parameters for the update of the IBM Business Automation
Insights release.
- To disable HDFS storage
- Remove the HDFS storage path, remove the ingestion job because it is not needed when HDFS is not
enabled, and specify the savepoints for the
jobs.
PARAMS='flink.storageBucketUrl="",ingestion.install=false,bpmn.recoveryPath=/mnt/pv/savepoints/dba/bai-bpmn/savepoint-<savepoint-id>,icm.recoveryPath=/mnt/pv/savepoints/dba/bai-icm/savepoint-<savepoint-id>'
- To disable the capability of sending data to Kafka egress
- Set the data egress option to false and specify the savepoints for the
jobs.
PARAMS='settings.egress=false,bpmn.recoveryPath=/mnt/pv/savepoints/dba/bai-bpmn/savepoint-<savepoint-id>,icm.recoveryPath=/mnt/pv/savepoints/dba/bai-icm/savepoint-<savepoint-id>,ingestion.recoveryPath=/mnt/pv/savepoints/dba/bai-ingestion/savepoint-<savepoint-id>
Kafka topics are documented in Apache Kafka parameters.
- To update your release, follow steps 1 to 3 of Updating your Business Automation Insights deployment.
- Run the Helm upgrade command with the parameters that you defined in 3 above.
helm upgrade <release-name> bai.tar.gz --reuse-values --set $PARAMS --namespace <my-namespace> --tls
Results
After the update, the job manager log returns a message such as the following
one.
2019-01-08 18:11:04,737 INFO org.apache.flink.runtime.checkpoint.Checkpoints
- Skipping savepoint state for operator <operator-id>.
The
message should appear only once for each operator that you disabled. If you see more messages than
expected, contact IBM support.