Description
There is evidence in some recent jenkins failures that we may have some manor of bug in our http2 client/server code that can cause intra-node query requests to stall / timeout non-reproducibly.
In at least one known case, forcing the jetty & SolrClients used in the test to use http1.1, seems to prevent these test failures.
Attachments
Attachments
Issue Links
- is related to
-
SOLR-13338 HdfsAutoAddReplicasIntegrationTest failures
- Resolved
- relates to
-
SOLR-13541 Upgrade Jetty to 9.4.19.v20190610
- Closed