David Smiley created SOLR-18247:
-----------------------------------
Summary: Refactor: SolrClientCache should only cache
CloudSolrClients
Key: SOLR-18247
URL: https://issues.apache.org/jira/browse/SOLR-18247
Project: Solr
Issue Type: Task
Reporter: David Smiley
SolrClientCache caches both CloudSolrClient and HttpSolrClient instances. It's
only used for solrj-streaming. Nowadays, a single HttpSolrClient can
communicate with any URL via requestWithBaseUrl, so we don't need a per-URL
cache of them. And we can get an HttpSolrClient for a CloudSolrClient (*).
Each caller in solrj-streaming is actually working with a CloudSolrClient. We
could even consider adding a convenience method to StreamContext, which must be
referenced all over solrj-streaming, to have a convenience method or two in
order to reduce type-exposure of SolrClientCache needlessly.
Desired Outcomes:
* SolrClientCache will become simpler in Solr 11, just a cache of
CloudSolrClient
* SolrClientCache is referred to in much fewer places (less important)
Note: as of this writing, this isn't quite ready as
CloudSolrClient.getHttpClient doesn't yet exist. That signature is blocked by
CloudLegacySolrClient to be removed very soon.
Constraint: the timeout tuning of SolrClientCache must continue to be honored.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
