You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@thrift.apache.org by "Jens Geyer (Jira)" <ji...@apache.org> on 2020/04/09 07:24:00 UTC

[jira] [Comment Edited] (THRIFT-5172) NetStd OutOfMemoryException

    [ https://issues.apache.org/jira/browse/THRIFT-5172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17079006#comment-17079006 ] 

Jens Geyer edited comment on THRIFT-5172 at 4/9/20, 7:23 AM:
-------------------------------------------------------------

A transport should not be able to be opened multiple times (unless Open() does really nothing), we even have a nice exception code for this. 

Maybe there's a check missing?

In case of OOMs it could also be some IDisposable class not being disposed properly.

Oh, and we accept pull requests :-)


was (Author: jensg):
A transport should not be able to be opened multiple times (unless Open() does really nothing), we even have a nice exception code for this. 

Maybe there's a check missing?

In case of OOMs it could also be some IDisposable class not being disposed properly.

> NetStd OutOfMemoryException
> ---------------------------
>
>                 Key: THRIFT-5172
>                 URL: https://issues.apache.org/jira/browse/THRIFT-5172
>             Project: Thrift
>          Issue Type: Bug
>          Components: netstd - Library
>    Affects Versions: 0.13.0
>            Reporter: Eric Bléher
>            Priority: Major
>              Labels: patch
>   Original Estimate: 5m
>  Remaining Estimate: 5m
>
> When I have some network issues to connect to my Thrift server, it tries many attempts to connect to it. After a while, I get a "OutOfMemory" Exception.
> I read that it may be related to too many sockets opened.
> Looking quickly at the code in .NET Standard:
> TBaseClient
> public virtual async Task OpenTransportAsync(CancellationToken cancellationToken)
>  {
>  if (!_inputProtocol.Transport.IsOpen)
> { await _inputProtocol.Transport.OpenAsync(cancellationToken); }
> if (!*_inputProtocol*.Transport.IsOpen)
> { await _outputProtocol.Transport.OpenAsync(cancellationToken); }
> }
>  
>  
> This last check seems a bad copy/paste and should probably be
> if (!*_outputProtocol*.Transport.IsOpen)
> Otherwise, we indeed get the transport on the output protocol getting opened many times!
>  
> I just proposed a pull request with the fix.
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)