Folks,
We were trying to do the Optimize the Tar Persistence manager in our AEM5.6 prod instance. We had modified the repository.xml (both versioning and workspace) and workspace.xml files and added the following entries -
<PersistenceManager class="com.day.crx.persistence.tar.TarPersistenceManager">
<!-- Default value is 8M. To reduce miss in bundlecaching, increasing it to 256M -->
<param name="bundleCacheSize" value="256"/>
<!-- check and fix broken system nodes automatically -->
<param name="consistencyFixSystem" value="true"/>
<param name="consistencyFix" value="true"/>
<param name="consistencyCheck" value="true"/>
<!-- Ensure that heap memory is available to handle index -->
<param name="indexInMemory" value="true" />
</PersistenceManager>
Things went fine in QA and another pre-prod author env because we thing there were not many bundles but in production we had about 1.3 million bundles which had to be processed. When we attempted to do in prod, it was very-2 slow and we figured out that probably it would take complete 2 days to finish the process and we can not keep the system unavailable for 2-3 days :(. System has been running from last 3 years and this is the first time we had attempted to add the above entries.
We wondering if any of you in this forum had similar situation and if yes then how did you fix it or had to do another workaround? Also, if you have any suggestions on how we achieve the TarPM efficiently or what are the options we have to make this process faster?
Thanks,
