cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-11082) netty's level hadn't prevent OOM when receiver handle slow.
Date Mon, 08 Feb 2016 10:59:39 GMT


Sylvain Lebresne commented on CASSANDRA-11082:

So you're saying that we don't have any back-pressure mechanism regarding client being too
slow at receiving responses (and hence data to send accumulating server side)? Am I understanding
that correctly?

If so, I suppose that's true or at least I'm not aware of us doing anything special to prevent
that. And it's a good question as to why we configure the high/low watermark given that we
never check {{Channel.isWritable()}}: [~tjake], you introduced those configuration on CASSANDRA-6861?

With that said, there isn't too much evidence so far that OOM due to clients being too slow
to receive their responses is very common so lowering the priority.

> netty's level hadn't prevent OOM when receiver handle slow.
> -----------------------------------------------------------
>                 Key: CASSANDRA-11082
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Streaming and Messaging
>            Reporter: fujian
>            Priority: Critical
> as we know, netty will OOM when received client is slow.
> due to  if receiver can't handle response fast so that cassandra server can't flush data.
it will cause channeloutbuffer with big size.
> we see the cassandra had configure write high/low water level, but I can't found any
code to judge iswritable. so it will have possible OOM
> why cassandra hadn't handle this case?

This message was sent by Atlassian JIRA

View raw message