Alfresco Search Services 1.4.2 has been released.
Obtaining this release
The ZIP Distribution file can be downloaded from the following URL:
https://download.alfresco.com/cloudfront/release/community/SearchServices/1.4.2/alfresco-search-serv...
If you are using Docker, you can get the new Image by typing:
docker pull alfresco/alfresco-search-services:1.4.2
Alfresco Insight Engine 1.4.2 has been also released.
Insight Engine and Insight Engine Zeppelin artifacts can be downloaded from Quay.io using Enterprise credentials.
New Features
- SOLR REST API improvements described in Alfresco Docs
- Fix JSON Encoding problems for SOLR Repo WebScripts
- Several performance improvements in SOLR Trackers
- YourKit Java Agent available in Docker images by default
Third Party Product Versions
- Solr 6.6.5
- Jetty 9.3.27.v20190418
- Zeppelin 0.8.2
Compatibility
- Search Services 1.4.2 works with ACS 6.0, 6.1, 6.2
- Insight Engine 1.4.2 works with ACS 6.0, 6.1, 6.2 and AGS 3.0.2, 3.2.0
Details
Note on SEARCH-2028
This feature is only available from ACS 6.2.1. If you want to use this feature from ACS 6.0 or upper, you can use this Alfresco Repository module that adds the feature to these releases:
https://github.com/aborroy/alfresco-search-2028-module
Performance improvement metrics
Metrics from our internal testing environment show a 20% performance improvement from 1.3.0 when indexing large transactions. This kind of transactions are common when repositories have been created from a bulk ingestion process or when documents are uploaded in large batches to Alfresco. This happens when the documents are uploaded by applications.
Performance observed when indexing small transactions is equivalent to 1.3.0. This kind of transactions are common when an organisation is using Share or ACA web apps as main source for creating and updating documents. This happens when the documents are uploaded mainly by users.
The metrics have been extracted from a sample repository containing 100,000 documents grouped in folders of 1,000 documents.
SOLR server is running from ZIP Distribution file using following JVM settings:
export GC_TUNE="-XX:NewRatio=3
-XX:SurvivorRatio=4
-XX:TargetSurvivorRatio=90
-XX:MaxTenuringThreshold=8
-XX:+UseConcMarkSweepGC
-XX:ConcGCThreads=4
-XX:ParallelGCThreads=4
-XX:+CMSScavengeBeforeRemark
-XX:PretenureSizeThreshold=64m
-XX:+UseCMSInitiatingOccupancyOnly
-XX:CMSInitiatingOccupancyFraction=50
-XX:CMSMaxAbortablePrecleanTime=6000
-XX:+CMSParallelRemarkEnabled
-XX:+ParallelRefProcEnabled
-XX:-OmitStackTraceInFastThrow"
This server is started using default SOLR command:
./solr/bin/solr start -a "-Dcreate.alfresco.defaults=alfresco,archive" -f
Sharding performance improvement
When using DB_ID_RANGE shard method, every transaction containing DB_ID nodes out of the range is omitted.
Performance improvement in this scenario depends on the number of Shards and the number of Nodes, but it performs quite faster than 1.3.0.
Since 1.4.2 is fully compatible with Alfresco 6.0+, this sharding performance improvement will be only applied when using Alfresco 6.2.0.7 repository.
