hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hive QA (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-20040) JDBC: HTTP listen queue is 50 and SYNs are lost
Date Mon, 09 Jul 2018 23:30:00 GMT

    [ https://issues.apache.org/jira/browse/HIVE-20040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16537781#comment-16537781
] 

Hive QA commented on HIVE-20040:
--------------------------------



Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12930868/HIVE-20040.1.patch

{color:red}ERROR:{color} -1 due to no test(s) being added or modified.

{color:green}SUCCESS:{color} +1 due to 14620 tests passed

Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/12496/testReport
Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/12496/console
Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-12496/

Messages:
{noformat}
Executing org.apache.hive.ptest.execution.TestCheckPhase
Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.YetusPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
{noformat}

This message is automatically generated.

ATTACHMENT ID: 12930868 - PreCommit-HIVE-Build

> JDBC: HTTP listen queue is 50 and SYNs are lost
> -----------------------------------------------
>
>                 Key: HIVE-20040
>                 URL: https://issues.apache.org/jira/browse/HIVE-20040
>             Project: Hive
>          Issue Type: Bug
>          Components: JDBC
>    Affects Versions: 3.0.0, 3.1.0
>            Reporter: Gopal V
>            Assignee: Gopal V
>            Priority: Major
>              Labels: Branch3Candidate
>         Attachments: HIVE-20040.1.patch, HIVE-20040.1.patch
>
>
> When testing with 5000 concurrent users, the JDBC HTTP port ends up overflowing on SYNs
when the HS2 gc pauses.
> This is because each getQueryProgress request is an independent HTTP request, so unlike
the BINARY mode, there are more connections being established & closed in HTTP mode.
> {code}
> LISTEN     0      50           *:10004                    *:* 
> {code}
> This turns into connection errors when enabling {{net.ipv4.tcp_abort_on_overflow=1}},
but the better approach is to enqueue the connections until the HS2 is done with its GC pause.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message