If you are seeing an error message as below in /loglogic/logs/index.log file on a LogLogic LMI appliance:

com.loglogic.logapp.index.IndexScheduler - indexing error
java.util.concurrent.ExecutionException: java.lang.OutOfMemoryError: Java heap space

then this is evidence that the memory allocations for the indexer's JVM must be increased.

You can increase the indexer memory allocation size in /loglogic/conf/capability.xml to resolve this behavior. Note that if the LogLogic LMI appliance is running LMI 6.2.0 or higher than the procedure for increasing the JVM memory is different. Refer to article 000036251 for that procedure.

1. Search for your platform (e.g. LX4020, ST3010, ST4035, etc.) in the capability.xml file and then look for the following section within the appropriate platform section:

<applianceProperty>
<propertyName>indexer_memory_size</propertyName>
<propertyValue>256</propertyValue>
</applianceProperty>

2. $ mtask stop; mtask start

The default value will differ depending on appliance model and LMI version. The amount to increase the value by will differ also by appliance modem and the current value as well as by how much available physical memory currently exists on the appliance. For those reasons, if you are not sure which value to increase the setting to then contact TIBCO Support so we can help you make that determination.

Note that this change has to be implemented manually on both HA nodes when the affected appliance is in an HA cluster. This is because this setting won't replicate to the other node. When performing the change on both nodes it is recommended to update the slave node first by executing both steps then execute it on the master. When restarting mtask on the master this will cause a failover to the standby node.