0

I have run the following via HTTP:

http://solr-uat.cambridgeassessment.org.uk/solr/admin/collections?action=create&name=ocr_education_and_learning_web8&numShards=1&maxShardsPerNode=8&replicationFactor=3&collection.configName=ocr_education_and_learning

and it created the collection but the cores on each server (there are 3 servers) have had the name appended (e.g. ocr_education_and_learning_web8_shard1_replica1). I am integrating with SI4T and it seems to use the core name rather than the collection name so the core names need to be the same across servers but I can't find how to do this.

Can anyone advise how best to do this?

CALJ
  • 53
  • 4

1 Answers1

0

As far as I know you can't do this. Core names must be unique. This naming scheme is internal to SolrCloud and is used to distinguish different indexes ('cores') from each other (which each make up part of the overall collection).

See this nice answer for more information

kellyfj
  • 6,586
  • 12
  • 45
  • 66
  • Thanks for the help, we have now found that you can still use the core name without the shard and replica details which i was unaware of! – CALJ Mar 14 '19 at 15:36
  • Ooohhh cool - Can you share any weblinks or documentation on that? – kellyfj Mar 14 '19 at 17:36
  • This was something that a colleague advised me on so we didn't refer to documentation but it looks like someone had a similar situation to me here https://grokbase.com/t/lucene/solr-user/15224p3g4r/core-property-name-ignored-when-creating-collection-using-api and they got referrred to the documentation here: https://wiki.apache.org/solr/SolrCloud#Distributed_Requests – CALJ Mar 18 '19 at 10:09
  • Also, found this one which details it: https://lucene.apache.org/solr/guide/6_6/distributed-search-with-index-sharding.html – CALJ Mar 20 '19 at 13:33