VBD (Real-Time Data Grid) nodes remain in Joining state
While restarting all the nodes, the Visual Business Director (VBD) nodes do not join the VBD cluster and are in the JOINING state infinitely.
Steps to Reproduce
Add a VBD node.
Some sort of corruption in the VBD data saved in Cassandra. Enabling debug level logging on "com.pega.decision.vbd" on the node used as RTDG node, we can see lot of log message as below.
VBD data saved in Cassandra was corrupted. When the debug level logging was enabled on 'com.pega.decision.vbd' on the node used as a Real-Time Data Grid (RTDG) node, multiple log messages displayed as below.
(n.request.InitializeVBDRequest) DEBUG - Not all partitions ready. Retrying partition load
Perform the following local-change:
Delete VBD specific data from Cassandra using the below steps,
- Stop the Pega JVM for VBD nodes.
- Decommission all RTDG (VBD) nodes from the RTDG cluster.
- On the Decision Data Store (DDS) node, from the ../cassandra/bin directory run the cqlsh command.
./cqlsh <ip of the dds node> -u dnode_ext -p dnode_ext
- On the cql command prompt, run the Truncate command on the tables below.
cqlsh> truncate vbd.collection_configuration;
cqlsh> truncate vbd.v2_collection_Actuals;
cqlsh> truncate vbd.v2_partition_summary;
- Start Pega JVM for the VBD nodes.
- Add the node to RTDG cluster.
0% found this useful