In SOLR-12610, I noticed that requests being retried by CloudSolrClient when there is no communication exception and no stale state. It seems that CloudSolrCient will retry when there is a RouteException however RouteException is not a communication error but only a wrapper exception for all underlying RemoteExceptions. In other words, CloudSolrClient will retry request on all HTTP 500 errors coming from the server including on admin requests such as split shards.
- blocks
-
SOLR-12610 Inject failures during synchronous update requests during shard splits
-
- Open
-