Dear Apache Ignite Support Team,
I hope this email finds you in good health. I am writing to seek assistance
regarding an issue I am currently encountering with Apache Ignite while
persisting data in memory. I have successfully set up Apache Ignite within
a Docker container, allocating 120GB of memory to the virtual machine.
Specifically, I have configured an off-heap memory of 80GB for Apache
Ignite in the configuration file.
The relevant part of the configuration file pertaining to data storage is
as follows:
<property name="dataStorageConfiguration">
<bean class="org.apache.ignite.configuration.DataStorageConfiguration">
<property name="defaultDataRegionConfiguration">
<bean
class="org.apache.ignite.configuration.DataRegionConfiguration">
<property name="name" value="Default_Region"/>
<property name="evictionThreshold" value="0.5"/>
<property name="initialSize" value="#{256 * 1024 * 1024}"/>
<property name="maxSize" value="#{80L * 1024 * 1024 *
1024}"/>
<property name="pageEvictionMode" value="RANDOM_LRU"/>
</bean>
</property>
</bean>
</property>
Despite configuring the eviction policy settings, I have observed that when
the Ignite memory usage approaches nearly 100%, the eviction process does
not initiate as expected. I have also attempted caching the data with a
TTL, but upon reviewing the Ignite logs, I noticed that there is no
increase in the freeRam percentage for the specified data region when the
data is destroyed. When the freeRam percentage approaches 0%, I am forced
to restart the container.
I have tried implementing the suggested actions from the advisory message,
including increasing the maximum off-heap memory size and enabling Ignite
persistence and eviction or expiration policies. However, these actions did
not resolve the issue.
I would greatly appreciate your assistance in enabling the eviction policy
to function effectively and initiating eviction when the Ignite memory
usage approaches the specified threshold. If there are any additional
solutions or insights from the documentation regarding this matter, please
do share them.
Thank you for your time and support. I look forward to your guidance.
Best regards,
Furkan Sahin
