You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@geode.apache.org by "Galen O'Sullivan (JIRA)" <ji...@apache.org> on 2017/07/12 19:07:00 UTC
[jira] [Comment Edited] (GEODE-3170) Geode keeps logging
org.apache.geode.protocol.exception.InvalidProtocolMessageException
[ https://issues.apache.org/jira/browse/GEODE-3170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16084514#comment-16084514 ]
Galen O'Sullivan edited comment on GEODE-3170 at 7/12/17 7:06 PM:
------------------------------------------------------------------
This issue is caused by a socket that's closed on the client side, which causes a SocketException to be thrown in ProtobufProtocolSerializer.deserialize . This throws a InvalidProtocolMessageException wrapping the SocketException (as an IOException), which is then rethrown as an IOException (whew!).
We need to implement shutdown properly to have the shutdown method to call when the socket is no longer readable (hence the dependency on GEODE-3079).
I think it's reasonable to close a socket if we ever get an IOException from it. Are there other cases of IOException I'm missing?
was (Author: gosullivan):
This issue is caused by a socket that's closed on the client side, which causes a SocketException to be thrown in ProtobufProtocolSerializer.deserialize . This throws a InvalidProtocolMessageException wrapping the SocketException (as an IOException), which is then rethrown as an IOException (whew!).
We need to implement shutdown properly to have the shutdown method to call when the socket is no longer readable.
I think it's reasonable to close a socket if we ever get an IOException from it. Are there other cases of IOException I'm missing?
> Geode keeps logging org.apache.geode.protocol.exception.InvalidProtocolMessageException
> ---------------------------------------------------------------------------------------
>
> Key: GEODE-3170
> URL: https://issues.apache.org/jira/browse/GEODE-3170
> Project: Geode
> Issue Type: Bug
> Components: client/server
> Reporter: Alexander Murmann
> Assignee: Galen O'Sullivan
> Fix For: 1.3.0
>
> Attachments: Proto-parse-error (1).txt
>
>
> We've been seeing the logs getting filled up with exceptions if the new binary protocol is enabled. This quickly fills up the hard drive.
> See attached file for example log
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)