hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Oleg Kalnichevski <ol...@apache.org>
Subject Re: High server load due to SessionOutputBufferImpl.flush()
Date Fri, 19 Sep 2014 12:53:00 GMT
On Fri, 2014-09-19 at 14:34 +0200, Tobias Bieniek wrote:
> > Is there a chance you could find out what kind of exception
> > (with the full stack trace) caused the initial 500 status and triggered
> > the whole sequence of events.
> I took another look at the whole logfile, but I couldn't find any
> logged exceptions around this time. We also have a "global" catch
> block that converts any thrown exception into a JSON response and
> returns it with HTTP 200 (legacy design...), so I'm not sure if this
> is really thrown by our application code.

Exception handling with async I/O is hell. The exception is not thrown
by the application code but it is passed to it through a callback event
of the request consumer and response producer. I suspect that exception
may simply get ignored.


> > Could you also consider upgrading to the
> > latest snapshot of the 4.3.x branch?
> >
> > http://svn.apache.org/repos/asf/httpcomponents/httpcore/branches/4.3.x/
> are there any snapshot releases published to a public maven repo
> somewhere that we could use or do we have to compile the jars
> manually?

I published the latest build to the ASF snapshot repo, but building from
source is always preferred 

> > And of course, a reproducer that I could run locally would be awesome.
> we haven't been able yet to reproduce this issue, but as soon as we do
> I'll let you know.

I am looking forward to it.


To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org

View raw message