Troubleshooting

Suggested workarounds in case of errors in purging data

The Data Purge Utility may encounter errors while attempting to purge data from the specified tables. The following workarounds can be used to help overcome these errors:

1. Modification of Cassandra configuration file:
1. Login to Genesys Administrator and stop Workbench, in order to stop the Cassandra process.
2. In an OS command prompt window on the Workbench host, change directory to <WORKBENCH_INSTALLATION>/cassandra/conf>
1. In this directory, copy file cassandra.yaml to cassandra.yaml.sav
3. Open file cassandra.yaml in a text editor and change the following:

cassandra.yaml	Current value	New Value
read_request_timeout_in_ms	5000	30000
request_timeout_in_ms	10000	60000

4. Using Genesys Administrator, start Genesys Workbench, in order to restart the Cassandra process.
5. Re-run the DataPurgeUtility tool.
Important
If there are a lot of records in the database, it is still possible to see the Read timeout errors when deleting records. In that case, it is advisable to change both the timeout numbers to 3600000 instead of 30000 and 60000 (refer the above table).
6. After purging, do the following to roll back the Cassandra configuration changes,
   1. Stop Workbench/Cassandra (see step a.)
   2. Copy cassandra.yaml.sav to cassandra.yaml.
   3. Start Workbench/Cassandra (see step d.).
2. Run the DataPurgeUtility using a parameter -tombstone_seconds when invoking DataPurgeUtility. Specify a small integer value e.g. 300 (for 5 minute). Please note that the GC_GRACE_SECONDS value for a specific table in Cassandra is temporarily set to this specified value. The value is restored to the default value of 8640000 (10 days) after the data is deleted. An example of command line is given below.

java -jar DataPurgeUtility.jar -h 135.17.180.184 -p 9042 -t context_tree_new.events_by_type_name -c ts 
-d 30 -a event_type event_name -pk 0 –prk 1 -tombstone_seconds 300