zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-2809) Unnecessary stack-trace in server when the client disconnect unexpectedly
Date Fri, 01 Sep 2017 21:05:00 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-2809?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16151154#comment-16151154

ASF GitHub Bot commented on ZOOKEEPER-2809:

Github user afine commented on a diff in the pull request:

    --- Diff: src/java/main/org/apache/zookeeper/server/NIOServerCnxn.java ---
    @@ -374,14 +373,12 @@ void doIO(SelectionKey k) throws InterruptedException {
                 // expecting close to log session closure
             } catch (EndOfStreamException e) {
    -            LOG.warn("caught end of stream exception",e); // tell user why
    +            LOG.warn(e.getMessage());
    --- End diff --
    Do we want to log the whole stack trace if debug is enabled here?

> Unnecessary stack-trace in server when the client disconnect unexpectedly
> -------------------------------------------------------------------------
>                 Key: ZOOKEEPER-2809
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2809
>             Project: ZooKeeper
>          Issue Type: Bug
>          Components: server
>    Affects Versions: 3.4.8
>            Reporter: Paul Millar
>            Assignee: Mark Fenes
>            Priority: Minor
>             Fix For: 3.5.0
> In ZK 3.4.x, if the client disconnects unexpectedly then the server logs this with a
stack-trace (see src/java/main/org/apache/zookeeper/server/NIOServerCnxn.java:356).
> This is unfortunate as we are using an embedded ZK server in our project (in a test environment)
and we consider all stack-traces as bugs.
> I noticed that ZK 3.5 and later no longer log a stack-trace.  This change is due to commit
6206b495 (in branch-3.5), which adds ZOOKEEPER-1504 and seems to fix this issue almost as
a side-effect; a similar change in master has the same effect.
> I was wondering if the change in how EndOfStreamException is logged (i.e., logging the
message without a stack-trace) could be back-ported to 3.4 branch, so could be included in
the next 3.4 release.

This message was sent by Atlassian JIRA

View raw message