Re: Bug in IndexFetcher + Http2SolrClient Interaction

2025-04-05 Thread David Smiley
*Excellent* research Luke! Yes, please file a JIRA and propose a fix. I also have noticed timeouts named one thing and applied in a different way. Very suspicious. Hey Gus; if memory serves, did you discuss that once? On Wed, Mar 19, 2025 at 6:28 PM Luke Kot-Zaniewski (BLOOMBERG/ 919 3RD A) <

Re: Bug in IndexFetcher + Http2SolrClient Interaction

2025-03-23 Thread Luke Kot-Zaniewski (BLOOMBERG/ 919 3RD A)
Yea but was the HttpSolrClient applying the default *idle* timeout (socket timeout) as the total request timeout? These are different things, especially for requests where the response is streamed slowly over long periods of time (so not idle .. just slow). I couldn't find this in the old HttpSo

Re: Bug in IndexFetcher + Http2SolrClient Interaction

2025-03-20 Thread Luke Kot-Zaniewski (BLOOMBERG/ 919 3RD A)
Great! I have added the ticket: https://issues.apache.org/jira/browse/SOLR-17711 Maybe it is best to continue the discussion there? From: dev@solr.apache.org At: 03/19/25 20:24:14 UTC-4:00To: dev@solr.apache.org, gus.h...@gmail.com Subject: Re: Bug in IndexFetcher + Http2SolrClient

Re: Bug in IndexFetcher + Http2SolrClient Interaction

2025-03-20 Thread sanjay dutt
I believe before IndexFetcher was creating HttpSolrClient using default setting for timeouts, I will confirm it once. DEFAULT_CONNECT_TIMEOUT = 6; DEFAULT_SO_TIMEOUT = 60; More Info here :- solr/solrj/src/java/org/apache/solr/client/solrj/impl/HttpClientUtil.java On Thu, Mar 20, 2025 at