hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nick Dimiduk (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-12833) [shell] table.rb leaks connections
Date Tue, 13 Jan 2015 23:38:34 GMT

    [ https://issues.apache.org/jira/browse/HBASE-12833?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14276204#comment-14276204

Nick Dimiduk commented on HBASE-12833:

bq. the interface changes and the managed/unmanaged connection issues are separate concerns


I also think that the implementation details of the shell *should not* be considered a part
of our public API contract. Point remains that this is code we've shipped that folks may be
relying on. I asked the question on dev@ yesterday. If no one chimes in, I think we can go
forward with the close and initialize changes for branch-1.0+ (and so long as RM is in agreement).

bq. Do you see any advantages to keeping a mix of managed and unmanaged connections around?

Not long term, no. However, we will continue to support both for 1.x series as a backward
compatibility consideration.

> [shell] table.rb leaks connections
> ----------------------------------
>                 Key: HBASE-12833
>                 URL: https://issues.apache.org/jira/browse/HBASE-12833
>             Project: HBase
>          Issue Type: Bug
>          Components: shell
>    Affects Versions: 1.0.0, 2.0.0, 1.1.0
>            Reporter: Nick Dimiduk
>            Assignee: Solomon Duskis
>             Fix For: 1.0.0, 2.0.0, 1.1.0
>         Attachments: HBASE-12833.patch
> TestShell is erring out (timeout) consistently for me. Culprit is OOM cannot create native
thread. It looks to me like test_table.rb and hbase/table.rb are made for leaking connections.
table calls ConnectionFactory.createConnection() for every table but provides no close() method
to clean it up. test_table creates a new table with every test.

This message was sent by Atlassian JIRA

View raw message