Author: alexparvulescu Date: Mon Jun 30 08:10:52 2014 New Revision: 1606639
URL: http://svn.apache.org/r1606639 Log: OAK-1804 TarMK compaction - docs Modified: jackrabbit/oak/trunk/oak-doc/src/site/markdown/FAQ.md jackrabbit/oak/trunk/oak-doc/src/site/markdown/nodestore/segmentmk.md Modified: jackrabbit/oak/trunk/oak-doc/src/site/markdown/FAQ.md URL: http://svn.apache.org/viewvc/jackrabbit/oak/trunk/oak-doc/src/site/markdown/FAQ.md?rev=1606639&r1=1606638&r2=1606639&view=diff ============================================================================== --- jackrabbit/oak/trunk/oak-doc/src/site/markdown/FAQ.md (original) +++ jackrabbit/oak/trunk/oak-doc/src/site/markdown/FAQ.md Mon Jun 30 08:10:52 2014 @@ -27,6 +27,9 @@ You might be missing an index. See [Quer You are accessing a `Session` instance concurrently from multiple threads. Session instances are [not thread safe](dos_and_donts.html#Anti_pattern:_concurrent_session_access). +#### I have a SegmentMK store and the size is growing beyond control + +You need to setup a regular job for [compacting the segments](nodestore/segmentmk.html#Segment_Compaction). #### My question is not listed here Modified: jackrabbit/oak/trunk/oak-doc/src/site/markdown/nodestore/segmentmk.md URL: http://svn.apache.org/viewvc/jackrabbit/oak/trunk/oak-doc/src/site/markdown/nodestore/segmentmk.md?rev=1606639&r1=1606638&r2=1606639&view=diff ============================================================================== --- jackrabbit/oak/trunk/oak-doc/src/site/markdown/nodestore/segmentmk.md (original) +++ jackrabbit/oak/trunk/oak-doc/src/site/markdown/nodestore/segmentmk.md Mon Jun 30 08:10:52 2014 @@ -140,6 +140,21 @@ it possible to quickly find all such ref traversal of all repository content. The list of active references is given to data store garbage collection so it won't collect the referenced binaries. +Segment Compaction +------------------ + +Segment compaction is needed when the repository size grows to an unmanageable +size. The growth can happen when there are a lot of repository write operations +which might create a considerable amount of content revisions that need to get +garbage collected. +The simplest solution is to schedule the revision garbage collection exposed as +a jmx operation via the `RevisionGarbageCollection` MBean. This is a non-blocking +operation, it will set an internal flag for compaction during the next `flush` call. + +As a workaround in the case where the size is already too big, you might want to run +a manual compaction using the [Oak run](https://github.com/apache/jackrabbit-oak/blob/trunk/oak-run/README.md) +tool. + Journals ========
