Search Services 1.3.0.1 sharding

[ivanovpavel1983]

Hello.

We have 2 application servers installed in production environment : alfresco 6.1 + search services 1.3.0.1
Transactions number is about 40 millions, i would like to create sharded index.
Could you tell me how to cluster solr-index properly?
I tried to install near one more solr 1.3.0.1, installed in solr.in.sh and shared.properties port 8984,
created 8 shards for workspacestore and 4 shards for archivestore by

http://alfapp01:8984/solr/admin/cores?action=newCore&coreName=alfresco&storeRef=workspace://SpacesStore&numShards=8&numNodes=1&nodeInstance=1&template=rerank&property.data.dir.root=/index/solrhome
http://alfapp01.isb:8984/solr/admin/cores?action=newCore&coreName=archive&storeRef=archive://SpacesStore&numShards=4&numNodes=1&nodeInstance=1&template=rerank&property.data.dir.root=/index/solrhome
but there are constantly errors in logs

2019-06-01 12:25:07.392 ERROR (SolrTrackingPool-alfresco-7-MetadataTracker-2) [   ] o.a.s.SolrInformationServer SolrInformationServer problem 
org.alfresco.solr.TrackerStateException: 05010427 The trackers work was rolled back by another tracker error 
        at org.alfresco.solr.SolrInformationServer.canUpdate(SolrInformationServer.java:4029) 
        at org.alfresco.solr.SolrInformationServer.indexNonShardCascade(SolrInformationServer.java:3654) 
        at org.alfresco.solr.SolrInformationServer.indexNodes(SolrInformationServer.java:2590) 
        at org.alfresco.solr.tracker.MetadataTracker$NodeIndexWorkerRunnable.doWork(MetadataTracker.java:882) 
        at org.alfresco.solr.tracker.AbstractWorkerRunnable.run(AbstractWorkerRunnable.java:45) 
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) 
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) 
        at java.base/java.lang.Thread.run(Thread.java:834) 
2019-06-01 12:25:11.242 ERROR (SolrTrackingPool-alfresco-7-MetadataTracker-4) [   ] o.a.s.SolrInformationServer SolrInformationServer problem 
org.alfresco.solr.TrackerStateException: 05010428 The trackers work was rolled back by another tracker error 
        at org.alfresco.solr.SolrInformationServer.canUpdate(SolrInformationServer.java:4029) 
        at org.alfresco.solr.SolrInformationServer.indexNonShardCascade(SolrInformationServer.java:3654) 
        at org.alfresco.solr.SolrInformationServer.indexNodes(SolrInformationServer.java:2590) 
        at org.alfresco.solr.tracker.MetadataTracker$NodeIndexWorkerRunnable.doWork(MetadataTracker.java:882) 
        at org.alfresco.solr.tracker.AbstractWorkerRunnable.run(AbstractWorkerRunnable.java:45) 
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) 
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) 
        at java.base/java.lang.Thread.run(Thread.java:834) 
2019-06-01 12:25:12.701 ERROR (SolrTrackingPool-alfresco-7-MetadataTracker-5) [   ] o.a.s.SolrInformationServer SolrInformationServer problem 
org.alfresco.solr.TrackerStateException: 05010429 The trackers work was rolled back by another tracker error 
        at org.alfresco.solr.SolrInformationServer.canUpdate(SolrInformationServer.java:4029) 
        at org.alfresco.solr.SolrInformationServer.indexNonShardCascade(SolrInformationServer.java:3654) 
        at org.alfresco.solr.SolrInformationServer.indexNodes(SolrInformationServer.java:2590) 
        at org.alfresco.solr.tracker.MetadataTracker$NodeIndexWorkerRunnable.doWork(MetadataTracker.java:882) 
        at org.alfresco.solr.tracker.AbstractWorkerRunnable.run(AbstractWorkerRunnable.java:45) 
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) 
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) 
        at java.base/java.lang.Thread.run(Thread.java:834) 

and index doesn't grow IN SIZE

at the same time the core cpu loading is 100% alfresco application.

On 15:10 index building started:

We use applications servers with 6 cores cpu intel xeon gold.

Tell me please the best practies how to create sharded index in online to not stop all cluster.

Thank you

[cesarista]

Hi:

Maybe this helps:

https://docs.alfresco.com/5.2/references/sharding-best-practices.html

https://docs.alfresco.com/5.2/concepts/solr6-shard-approaches.html

https://www.youtube.com/watch?v=_JkO2yjznh8

https://es.slideshare.net/JosePortillo26/jose-portillo-dev-con-presentation-1138

Kind regards.

--C.

P.S: If you have an Enterprise license, try checking with Alfresco support too. 

[ivanovpavel1983]

Yes, this is theoretical knowledge. I read it and created shards as documented.

But in solr.log every second an errors occurs. And size was began to grow in only 3 shards. Index building speed is very small. I think this is caused by blocking in java.