[
https://issues.apache.org/jira/browse/SOLR-7344?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14494486#comment-14494486
]
Hrishikesh Gadre commented on SOLR-7344:
----------------------------------------
[~andyetitmoves] I like your suggestion and it is not very difficult to
incorporate in my earlier proposal. The main change would be as follows,
- Instead of defining 'externalPort' property, we can define an 'endpoint'
element which would allow specifying details such as
--> port number
--> request type (either a URL suffix of the REST API or a wildcard e.g. '*'
for all requests)
--> Anything else (?)
- We would publish the information about all configured pools as part of the
ZNODE representing the given Solr server (under /live_nodes ZNODE)
- We would change solrj to use the correct endpoint based on the request_type
(may be a URL suffix?). If no endpoint is available, we can fallback on
base_url value (i.e. the current behavior) to maintain backward compatibility.
Just out of curiosity, can you please explain why someone would *not* want to
define separate thread pools for internal and external requests? Also do you
think if there would be use-cases which would require separate endpoints per
request type?
> Use two thread pools, one for internal requests and one for external, to
> avoid distributed deadlock and decrease the number of threads that need to be
> created.
> ---------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-7344
> URL: https://issues.apache.org/jira/browse/SOLR-7344
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Reporter: Mark Miller
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
