You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Erick Erickson (JIRA)" <ji...@apache.org> on 2017/11/30 17:07:00 UTC

[jira] [Created] (SOLR-11708) Streaming SolrJ clients hang for 50 seconds when closing the stream

Erick Erickson created SOLR-11708:
-------------------------------------

             Summary: Streaming SolrJ clients hang for 50 seconds when closing the stream
                 Key: SOLR-11708
                 URL: https://issues.apache.org/jira/browse/SOLR-11708
             Project: Solr
          Issue Type: Bug
      Security Level: Public (Default Security Level. Issues are Public)
    Affects Versions: 6.6.2, 7.1, 7.0
            Reporter: Erick Erickson


I'll attach my client program in a second but the gist is that closing a CloudSolrStream hangs the client for about 50 seconds no matter how it's closed.

Setup:
- Create a 2-shard, leader-only collection
- Index a million documents to it
- run the attached program.

At the end you'll see the message printed out "We finally stopped", but the program hands for roughly 50 seconds before exiting.

The hang happens with all these situations:
- read to EOF tuple
- stop part way through
- close in a finally block
- close with try-with-resources

In the early-termination case, I do see the following (expected) error in solr's log:

ERROR - 2017-11-30 16:18:20.024; [c:eoe s:shard2 r:core_node4 x:eoe_shard2_replica_n2] org.apache.solr.common.SolrException; null:Early Client Disconnect

I see the same behavior in 7.0

The claim is this worked in earlier 7x versions, 7.0 in particular. I'll test that shortly and report results.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org