Oak-run supports all indexing use cases on the command line without having to operate from the JMX level. Advantages of the oak-run approach are:
Below you will find a list of use cases that can be leveraged when performing indexing operations via the oak-run
tool.
For more detailed information regarding this scenario, see Use Case 1 - Index Consistency Check.
oak-run.jar
quickly determines if Lucene oak indexes are corrupt.For more detailed information regarding this scenario, see Use Case 2 - Index Statistics
oak-run.jar
dumps all index definitions, important index stats and index contents for offline analysis.This diagram is a decision tree for when to use the various re-indexing approaches.
For more detailed information regarding this scenario, see Use Case 3 - Reindexing.
Text pre-extraction (a feature that has existed with AEM 6.3) can be used to reduce the time to re-index. Text pre-extraction can be used in conjunction with all re-indexing approaches.
Depending on the oak-run.jar
indexing approach there will be various steps on either side of the Perform Re-index step in the diagram below.
Orange denotes activities where AEM must be in a maintenance window.
For more detailed information regarding this scenario, see Reindex - DocumentNodeStore.
This is the recommended method for re-indexing MongoMK (and RDBMK) AEM installations. No other method should be used.
This process needs to be executed only against a single AEM instance in the cluster.
For more detailed information regarding this scenario, see Reindex - SegmentNodeStore.
Cold Standby considerations (TarMK)
AEM Publish Farms (AE Publish Farms should always be TarMK)
For more detailed information regarding this scenario, see Online Reindex - SegmentNodeStore.
This is the method used before the introduction of the new indexing capabilities of oak-run.jar. It can done by setting the reindex=true
property on the Oak index.
This approach can be used if the time and performance effects to index are acceptable to the customer. This is often the case for small to medium sized AEM installations.
For more detailed information regarding this scenario, see Online Reindex - SegmentNodeStore - The AEM Instance is Running.
Online re-indexing of TarMK using the oak-run.jar is faster than the Online Re-Indexing for TarMK described above. However, it also requires execution during a maintenance window; with the mention that the window will be shorter, and more steps are required to perform the re-indexing.
Orange denotes operations where AEM must be performed in a maintenance period.
For more detailed information regarding this scenario, see Online Reindex - SegmentNodeStore - The AEM Instance is Shut Down.
Offline re-indexing of TarMK is the simplest oak-run.jar
based re-indexing approach for TarMK as it requires a single oak-run.jar
comment. However, it requires the AEM instance to be shutdown.
Red denotes operations where AEM must be shut down.
For more detailed information regarding this scenario, see Out of Band Reindex - SegmentNodeStore.
Out-of-band re-indexing minimizes the impact of re-indexing on in-use AEM instances.
Red denotes operations where AEM may be shut down.
For more detailed information about this scenario, see Use Case 4 - Updating Index Definitions.
ACS Ensure Index is a community supported project, and is not supported by Adobe Support.
This allows shipping index definition via content package which later results in re-indexing via setting the re-index flag to true
. This works for smaller setups where re-indexing does not take long time.
For more info, see the ACS Ensure Index documentation for details.
If the time or performance impact of re-indexing using non oak-run.jar
methods is too high, the following oak-run.jar
based approach can be used to import and re-index Lucene Index definitions in a TarMK based AEM installation.
If the time or performance impact of re-indexing using non oak-run.jar
methods is too high, the following oak-run.jar
based approach can be used to import and re-index Lucene Index definitions in MongoMK based AEM installations.