In Adobe Experience Manager (AEM), binary data can be stored independently from the content nodes. The binary data is stored in a data store, whereas content nodes are stored in a node store.
Both data stores and node stores can be configured using OSGi configuration. Each OSGi configuration is referenced using a persistent identifier (PID).
To configure both the node store and the data store, perform these steps:
Copy the AEM quickstart JAR file to its installation directory.
Create a folder
crx-quickstart/install in the installation directory.
First, configure the node store by creating a configuration file with the name of the node store option you want to use in the
For example, the Document node store (which is the basis for AEM’s MongoMK implementation) uses the file
Edit the file, and set your configuration options.
Create a configuration file with the PID of the data store you want to use. Edit the file to set the configuration options.
See Node Store Configurations and Data Store Configurations for configuration options.
Newer versions of Oak employ a new naming scheme and format for OSGi configuration files. The new naming scheme requires that the configuration file be named .config and the new format requires values to be typed and is documented here.
If you upgrade from an older version of Oak, ensure that you make a backup of the
crx-quickstart/installfolder first. After the upgrade, restore the contents of the folder to the upgraded installation and modify the extension of the configuration files from .cfg to .config.
In case you are reading this article in preparation for an upgrade from an AEM 5.x installation, ensure that you consult the upgrade documentation first.
The segment node store is the basis of Adobe’s TarMK implementation in AEM6. It uses the
org.apache.jackrabbit.oak.segment.SegmentNodeStoreService PID for configuration.
The PID for the Segment node store has changed from
org.apache.jackrabbit.oak.plugins.segment.SegmentNodeStoreService in previous versions of AEM 6 to
org.apache.jackrabbit.oak.segment.SegmentNodeStoreService in AEM 6.3. Make sure you make the necessary configuration adjustments to reflect this change.
You can configure the following options:
repository.home: Path to repository home under which repository-related data is stored. By default, segment files are stored under the
tarmk.size: Maximum size of a segment in MB. The default maximum is 256MB.
customBlobStore: Boolean value indicating that a custom data store is used. The default value is true for AEM 6.3 and later versions. Prior to AEM 6.3 the default was false.
The following is a sample
#Path to repo repository.home="crx-quickstart/repository" #Max segment size tarmk.size=I"256" #Custom data store customBlobStore=B"true"
The document node store is the basis of AEM’s MongoMK implementation. It uses the
org.apache.jackrabbit.oak.plugins.document.DocumentNodeStoreService* *PID. The following configuration options are available:
mongouri: The MongoURI required to connect to Mongo Database. The default is
db: Name of the Mongo database. The default is Oak
. However, new AEM 6 installations use **aem-author** as the default database name.
cache: The cache size in MB. This is distributed among various caches used in DocumentNodeStore. The default is
changesSize: Size in MB of capped collection used in Mongo for caching the diff output. The default is
customBlobStore: Boolean value indicating that a custom data store will be used. The default is
The following is a sample
#Mongo server details mongouri="mongodb://localhost:27017" #Name of Mongo database to use db="aem-author" #Store binaries in custom BlobStore customBlobStore=B"false"
When dealing with large number of binaries, it is recommended that an external data store be used instead of the default node stores to maximize performance.
For example, if your project requires many media assets, storing them under the File or S3 Data Store makes accessing them faster than storing them directly inside a MongoDB.
The File Data Store provides better performance than MongoDB, and Mongo backup and restore operations are also slower with large number of assets.
Details on the different data stores and configurations are described below.
To enable custom Data Stores, you must make sure that
customBlobStore is set to
true in the respective Node Store configuration file (segment node store or document node store).
This is the implementation of FileDataStore present in Jackrabbit 2. It provides a way to store the binary data as normal files on the file system. It uses the
These configuration options are available:
repository.home: Path to repository home under which various repository related data is stored. By default, binary files would be stored under
path: Path to the directory under which the files would be stored. If specified then it takes precedence over
minRecordLength: The minimum size in bytes of a file stored in the data store. Binary content less than this value would be inlined.
When using a NAS to store shared file data stores, make sure you use only high performing devices to avoid performance issues.
AEM can be configured to store data in Amazon’s Simple Storage Service (S3). It uses the
org.apache.jackrabbit.oak.plugins.blob.datastore.S3DataStore.config PID for configuration.
To enable the S3 data store functionality, a feature pack containing the S3 Datastore Connector must be downloaded and installed. Go to the Adobe Repository and download the latest version from the 1.10.x versions of the feature pack (for example, com.adobe.granite.oak.s3connector-1.10.0.zip). Also, you must download and install the latest AEM service pack as listed on the AEM 6.5 Release Notes page.
When using AEM with TarMK, binaries will be stored by default in the
FileDataStore. To use TarMK with the S3 Datastore, you must start AEM using the
crx3tar-nofds runmode, for example:
java -jar <aem-jar-file>.jar -r crx3tar-nofds
Once downloaded, you can install and configure the S3 Connector as follows:
Extract the contents of the feature pack zip file to a temporary folder.
Go to the temporary folder and navigate to the following location:
Copy all the contents from the above location to
If AEM is already configured to work with the Tar or MongoDB storage, remove any existing configuration files from the <aem-install>/crx-quickstart/install folder before proceeding. The files that must be removed are:
For MongoMK: org.apache.jackrabbit.oak.plugins.document.DocumentNodeStoreService.config
For TarMK: org.apache.jackrabbit.oak.segment.SegmentNodeStoreService.config
Return to the temporary location where the feature pack has been extracted, and copy the contents of the following folder:
Make sure you only copy the configuration files needed by your current configuration. For both a dedicated data store and a shared data store setup copy the
In a cluster setup, perform above steps on all nodes of cluster one by one. Also, make sure to use same S3 settings for all nodes.
Edit the file and add the configuration options required by your setup.
To upgrade to a new version of the 1.10.x S3 connector (for example, from 1.10.0 to 1.10.4) follow these steps:
Stop the AEM instance.
<aem-install>/crx-quickstart/install/15 in the AEM installation folder and make a backup of its contents.
After the backup, delete the old version of the S3 Connector and its dependencies by deleting all the jar files in the
<aem-install>/crx-quickstart/install/15 folder, for example:
The file names presented above are used for illustration purposes only.
Download the latest version of the 1.10.x feature pack from the Adobe Repository.
Unzip the contents to a separate folder, then navigate to
Copy the jar files to <aem-install>/crx-quickstart/install/15 in the AEM installation folder.
Start AEM and check the connector functionality.
You can use the configuration file with the options detailed below.
The S3 connector supports both IAM user authentication and IAM role authentication. To use IAM role authentication, omit the
secretKey values from your configuration file. The S3 connector will then default to the IAM role assigned to the instance.
|accessKey||Access Key ID for the IAM user with access to the bucket.||Yes, when not using IAM roles.|
|secretKey||Secret access key for the IAM user with access to the bucket.||Yes, when not using IAM roles.|
|cacheSize||The size (in bytes) of the local cache.||64GB||No.|
|connectionTimeout||Set the amount of time to wait (in milliseconds) before timing out when initially establishing a connection.||10000||No.|
|maxCachedBinarySize||Binaries with size less than or equal to this value (in bytes) are stored in the memory cache.||17408 (17 KB)||No.|
|maxConnections||Set the maximum number of allowed open HTTP connections.||50||No.|
|maxErrorRetry||Set the maximum number of retry attempts for failed (retriable) requests.||3||No.|
|minRecordLength||The minimum size of an object (in bytes) that should be stored in the data store.||16384||No.|
|path||The local path of the AEM datastore.||
|proxyHost||Set the optional proxy host the client connects through.||No.|
|proxyPort||Set the optional proxy port the client connects through.||No.|
|s3Bucket||Name of the S3 bucket.||Yes|
|s3EndPoint||S3 REST API endpoint.||No.|
|s3Region||Region where the bucket resides. See this page for more details.||Region where AWS instance is running.||No.|
|socketTimeout||Set the amount of time to wait (in milliseconds) for data to be transferred over an established, open connection before the connection times out and is closed.||50000||No.|
|stagingPurgeInterval||The interval (in seconds) for purging finished uploads from the staging cache.||300||No.|
|stagingRetryInterval||The interval (in seconds) to retry failed uploads.||600||No.|
|stagingSplitPercentage||The percentage of
|uploadThreads||The number of upload threads used for asynchronous uploads.||10||No.|
|writeThreads||The number of concurrent threads used for writing via S3 Transfer Manager.||10||No.|
The DataStore implementations of
AzureDataStore support local file system caching. The
CachingFileDataStore implementation is useful when the DataStore is on NFS (Network File System).
When upgrading from an older cache implementation (pre Oak 1.6) there is a difference in the structure of the local file system cache directory. In the old cache structure, both the downloaded and the uploaded files were put directly under the cache path. The new structure segregates the downloads and uploads and stores them in two directories named
download under cache path. The upgrade process should be seamless and any pending uploads should be scheduled for upload and any previously downloaded files in the cache are put in the cache on initialization.
You can also upgrade the cache offline by using the
datastorecacheupgrade command of oak-run. For details on how to execute the command, check the readme for the oak-run module.
The cache has a size limit and it can be configured by using the cacheSize parameter.
The local cache is checked for the record of the requested file/blob before accessing it from the DataStore. When the cache exceeds the configured limit (see the
cacheSize parameter) while adding a file into the cache, then some of the files are evicted to reclaim space.
The cache supports asynchronous uploads to the DataStore. The files are staged locally, in the cache (on the file system), and an asynchronous job starts to upload the file. The number of asynchronous uploads is limited by the size of the staging cache. The size of the staging cache is configured by using the
stagingSplitPercentage parameter. This parameter defines the percentage of cache size to be used for the staging cache. Also, the percentage of cache available for downloads is calculated as (100 -
The asynchronous uploads are multi-threaded and the number of threads is configured by using the
The files are moved to the main download cache after the uploads are complete. When the staging cache size exceeds its limit, the files are uploaded synchronously to the DataStore until the previous asynchronous uploads are complete and space is again available in the staging cache. The uploaded files are removed from the staging area by a periodic job whose interval is configured by the
Failed uploads (for example, because of a network disruption) are put on a retry queue and retried periodically. The retry interval is configured by using the
To configure binaryless replication with S3, the following steps are required:
Install the author and publish instances and make sure they are started properly.
Go to the replication agent settings, by opening a page to https://localhost:4502/etc/replication/agents.author/publish.html.
Press the Edit button in the Settings section.
Change the Serialization type option to Binary less.
Add the parameter "
true" in the transport uri. After the change, the uri should look similar to the following:
Restart all author and publish instances to let the changes take effect.
Unpack CQ quickstart using the following command:
java -jar cq-quickstart.jar -unpack
After AEM has been unpacked, create a folder inside the installation directory crx-quickstart/install.
Create these two files inside the
After the files have been created, add the configuration options as needed.
Install the two bundles required for the S3 data store as explained above.
Make sure MongoDB is installed and an instance of
mongod is running.
Start AEM with the following command:
java -Xmx1024m -jar cq-quickstart.jar -r crx3,crx3mongo
Repeat steps 1 through 4 for the second AEM instance.
Start the second AEM instance.
First, create the data store configuration file on each instance that is required to share the data store:
If you are using a
FileDataStore, create a file named
org.apache.jackrabbit.oak.plugins.blob.datastore.FileDataStore.config and place it in the
If using S3 as the data store, create a file named o
rg.apache.jackrabbit.oak.plugins.blob.datastore.S3DataStore.config in the
<aem-install>/crx-quickstart/install folder as above.
Modify the data store configuration files on each instance so they point to the same data store. For more information, see this article.
If the instance has been cloned from an existing server, you must remove the
clusterId of the new instance by using the latest oak-run tool while the repository is offline. The command you must run is:
java -jar oak-run.jar resetclusterid < repository path | Mongo URI >
If a Segment node store is configured, then the repository path must be specified. By default, the path is
<aem-install-folder>/crx-quickstart/repository/segmentstore. If a Document node store is configured you can use a Mongo Connection String URI.
The Oak-run tool can be downloaded from this location:
Different versions of the tool must be used depending on the Oak version you use with your AEM installation. Check the version requirements list below before using the tool:
* For Oak versions **1.2.x** use the Oak-run **1.2.12 or newer** * For Oak versions **newer than the above**, use the version of Oak-run that matches the Oak core of your AEM installation.
Lastly, validate the configuration. To validate, look for a unique file added to the data store by each repository that is sharing it. The format of the files is
repository-[UUID], where the UUID is a unique identifier of each individual repository.
Therefore, a proper configuration should have as many unique files as there are repositories sharing the data store.
The files are stored differently, depending on the data store:
FileDataStorethe files are created under the root path of the data store folder.
S3DataStorethe files are created in the configured S3 bucket under the
AEM can be configured to store data in Microsoft®’s Azure storage service. It uses the
org.apache.jackrabbit.oak.plugins.blob.datastore.AzureDataStore.config PID for configuration.
To enable the Azure data store functionality, a feature pack containing the Azure Connector must be downloaded and installed. Go to the Adobe Repository and download the latest version from the 1.6.x versions of the feature pack (for example, com.adobe.granite.oak.azureblobconnector-1.6.3.zip).
When using AEM with TarMK, binaries are stored by default in the FileDataStore. To use TarMK with the Azure DataStore, you must start AEM using the
crx3tar-nofds runmode, for example:
java -jar <aem-jar-file>.jar -r crx3tar-nofds
Once downloaded, you can install and configure the Azure connector as follows:
Extract the contents of the feature pack zip file to a temporary folder.
Go to the temporary folder and copy the contents of
jcr_root/libs/system/install to the
If AEM is already configured to work with the Tar or MongoDB storage, remove any existing configuration files from the
/crx-quickstart/install folder before proceeding. The files that must be removed are:
Return to the temporary location where the feature pack has been extracted and copy the contents of
jcr_root/libs/system/config to the
Edit the configuration file and add the configuration options required by your setup.
You can use the configuration file with the following options:
azureSas=“”: In version 1.6.3 of the connector, Azure Shared Access Signature (SAS) support was added. If both SAS and storage credentials exists in the configuration file, SAS has priority. For more information about SAS see the official documentation. Ensure that the ‘=’ character is escaped like ‘=’.
azureBlobEndpoint=“”: The Azure Blob Endpoint. For example, https://<storage-account>.blob.core.windows.net.
accessKey=“”: The storage account name. For more details about the Microsoft® Azure authentication credentials, see the official documentation.
secretKey=“”: The storage access key. Ensure that the ‘=’ character is escaped like ‘=’.
container=“”: The Microsoft® Azure blob storage container name. The container is a grouping of a set of blobs. For additional details, read the official documentation.
maxConnections=“”: The concurrent number of simultaneous requests per operation. The default value is 1.
maxErrorRetry=“”: Number of retries per request. The default value is 3.
socketTimeout=“”: The timeout interval, in milliseconds, used for the request. The default value is 5 minutes.
Besides the settings above, the following settings can also be configured:
All settings should be put between quotes, for example:
The data store garbage collection process is used to remove any unused files in the data store, thus freeing up valuable disk space in the process.
You can run data store garbage collection by:
Going to the JMX console at https://<serveraddress:port>/system/console/jmx
Searching for RepositoryManagement. Once you find the Repository Manager MBean, click it to bring up the available options.
Scroll to the end of the page, and click the startDataStoreGC(boolean markOnly) link.
In the following dialogue, enter
false for the
markOnly parameter, then click Invoke:
markOnly parameter signifies whether the sweep phase of garbage collection runs or not.
When performing garbage collection in a clustered or shared data store, setup (with Mongo or Segment Tar) the log might display warnings about the inability to delete certain blob IDs. Blob IDs deleted in a previous garbage collection are incorrectly referenced again by other cluster or shared nodes which do not have information about the ID deletions. As a result, when garbage collection is performed it logs a warning when it tries to delete an ID that has already been deleted in the last run. This behavior does not affect performance or functionality.
If you are using a shared datastore setup and datastore garbage collection is disabled, running the Lucene Binary cleanup task can suddenly increase the disk space used. Consider disabling BlobTracker on all author and publish instances by doing the following:
blobTrackSnapshotIntervalInSecs=L"0"parameter in the
crx-quickstart/install/org.apache.jackrabbit.oak.segment.SegmentNodeStoreService.configfile. This parameter requires Oak 1.12.0, 1.10.2 or later.
With newer versions of AEM, data store garbage collection can also be run on data stores shared by more than one repository. To be able to run data store garbage collection on a shared data store, take the following steps:
Make sure that any maintenance tasks configured for the data store garbage collection are disabled on all repository instances sharing the data store.
Run the steps mentioned in Binary Garbage Collection individually on all repository instances sharing the data store. However, make sure to enter
true for the
markOnly parameter before clicking the Invoke button:
After completing the above procedure on all instances, run the data store garbage collect again from any of the instances:
All the files found are collated using the mark phase used before and delete the rest that are unused from the data store.