https://issues.apache.org/bugzilla/show_bug.cgi?id=56725
--- Comment #9 from Christopher Schultz <ch...@christopherschultz.net> ---
(In reply to mardkrause from comment #8)
> The first set of logs are the access logs (via configuring the
> AccessLogValve) on the "server server" aka "receiving server". These were
> requested by Konstantin Kolinko.
> The second set of logs are the normal console logs from the "client server".
I was referring to the discontinuity observed between these two lines of the
"client server" logs:
Retrying request to {}->http://localhost:8700
13:50:58.538 INFO http-nio-8710-exec-3 Logger.info - After request
[method=PUT;server=http://localhost:8710;uri=/application/api/file/conflictApp/files/a57f597a-bb7f-4221-8097-e4d9222c61cd?null;client=127.0.0.1]
Jul 16, 2014 1:50:58 PM org.apache.catalina.core.StandardWrapperValve invoke
SEVERE: Servlet.service() for servlet [ourServlet] in context with path
[/application] threw exception [Request processing failed; nested exception is
org.springframework.web.client.ResourceAccessException: I/O error on PUT
request for
"http://localhost:8700/management/api/storage/article/aApp/c47a43b3-88a8-474b-a532-99efc3baa68a":
Connection reset by peer: socket write error; nested exception is
java.net.SocketException: Connection reset by peer: socket write error] with
root cause
> > I also notice that every request gets a 401 (Unauthorized) response. Are you
> > checking for that and stopping?
>
> We are aware of the double request. The RestOperations is handling is. It
> seems to first send a request and when 401 is returned it then adds the
> authorization portion to the request and tries again.
Sorry, I didn't notice the 201s in that sea of 401s.
Weird. Do you know if "Expect" is sent along with either or both requests?
> > Read Konstantin's comment #3 about aborted uploads: if you get a 401
> > response, you are probably going to hit that limit with a "large" file and
> > find your connection forcibly closed.
>
> Is this something new with 8.0.9? Because again, that version is the only
> difference between a successful test run and one that fails.
Yes, it's new. See the last item under the "coyote" section here:
http://tomcat.apache.org/tomcat-8.0-doc/changelog.html
But note that it's not a limit on legitimate file uploads: instead, its a limit
on how much data Tomcat will "swallow" if it's going to reject the request.
Tomcat used to swallow essentially unlimited data but now it will fail after
2MiB by default by forcibly closing the connection.
--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org
