0

Let me start of by describing my current set up. I am currently having Sitecore 8.2 update 6 running with Solr Cloud 5.5.1 as a search engine. I have 1 CM and 2 CDs. Solr Cloud is running on 5 servers (2 Solr nodes running behind load balance and zookeeper on all 5 servers). CM is pointing to Solr 1 to process indexing and CDs are connected to Solr via load balance.

I am constantly having issue with search results and try to find out if it is set up or config issue.

1- Sometimes I do not see primary Index and secondary Index sync between Solr node 2- Primary index and secondary index do not in sync.

My understanding is Zookeeper should keep both Solr node in sync and Sitecore is taking care the syncing between primary index and secondary during switchindex on rebuild. However I do not see it is happening and wonder if anyone could share opinion if this is potentially the design issue, config issue, etc...? Is it zookeeper or sitecore job to keep primary and secondary in sync during rebuild index? It would be great if you could suggest location to look for to verify the configuration.

Thanks

  • Sitecore describe SolrCloud as experimental in 8.2 kb.sitecore.net/articles/227897 I have also seen issues with it on an 8.2 project, that were fairly strange to say the least. Is your crawling strategy set to Manual on all CD servers ? what is your crawling strategy on the CM server ? Is SwitchOnRebuild enabed ? – TomT Nov 22 '18 at 3:30
0

I think you may be mixing two concepts

  1. Multiple nodes configured in SolrCloud.
  2. Multiple (sets of) indexes configured in Sitecore.

In the first instance, SolrCloud should keep the indexes stored in the individual SolrCores in sync. In SolrCloud, a set of indexes replicated across multiple nodes (SolrCores) is called a Collection 1

In the second instance, Sitecore defines a mechanism called SwitchOnRebuild 2. This allows you to create two sets of indexes (core & core_2, web & web_2, etc.) so that one index can be in use by your website while the other is being rebuilt - I have experienced issues with these being out of sync if the configuration is not just so and even then have had to raise support tickets.

ZooKeeper

As you mention ZooKeeper, here's the definition from the SolrCloud docs: 1

Solr embeds and uses Zookeeper as a repository for cluster configuration and coordination - think of it as a distributed filesystem that contains information about all of the Solr servers

  • Thanks Pirate for the feedback and explanation. I guess to follow up with the question is currently I have CM point to one of Solr node to perform indexing and CDs are connection to load balancer in front of 2 Solr nodes (1 of them is pointed by CM). I did some research and a lot of people mentioned to point Sitecore (means both CM and CD) to Solr load balance (smartsitecore.com/en/setup-solr-sitecore-production). – Ethan P Nov 21 '18 at 16:20
  • Please post another question and I'll answer there - but as a teaser, yes, when using SolrCloud, both CM and CD should use a load balanced address – Pirate Briggs Nov 22 '18 at 18:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.