Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-26 Thread Mark Thomas
On 26/09/2013 02:47, Rainer Jung wrote: > The TestCoyoteAdapter now neither fails with tcnative 1.1.28 nor with > the additional APR_NOTFOUND change. I did 50 runs for each of the two > versions and also for bio/nio/apr and none failed. Good! +1 > I'm still a bit nervous about coupling the conne

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-26 Thread Rainer Jung
Hi Mark, On 25.09.2013 01:50, Mark Thomas wrote: > On 24/09/2013 16:19, Mark Thomas wrote: >> On 24/09/2013 08:16, Mark Thomas wrote: >>> On 23/09/2013 00:54, Rainer Jung wrote: >>> I agree that there's probably another problem further up the stack. >>> >>> I'm looking at this now. >> >> I've

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-25 Thread Rainer Jung
On 25.09.2013 01:50, Mark Thomas wrote: > On 24/09/2013 16:19, Mark Thomas wrote: >> On 24/09/2013 08:16, Mark Thomas wrote: >>> On 23/09/2013 00:54, Rainer Jung wrote: >>> I agree that there's probably another problem further up the stack. >>> >>> I'm looking at this now. >> >> I've found a c

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-24 Thread Mark Thomas
On 24/09/2013 16:19, Mark Thomas wrote: > On 24/09/2013 08:16, Mark Thomas wrote: >> On 23/09/2013 00:54, Rainer Jung wrote: >> >>> I agree that there's probably another problem further up the stack. >> >> I'm looking at this now. > > I've found a couple of issues. Not sure if either of these are

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-24 Thread Mark Thomas
On 24/09/2013 08:16, Mark Thomas wrote: > On 23/09/2013 00:54, Rainer Jung wrote: > >> I agree that there's probably another problem further up the stack. > > I'm looking at this now. I've found a couple of issues. Not sure if either of these are the root cause of the remaining issues we see but

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-24 Thread Mark Thomas
On 23/09/2013 00:54, Rainer Jung wrote: > I agree that there's probably another problem further up the stack. I'm looking at this now. > Since the apr endpoint explicitely uses the return value of remove to > decide whether something was removed, APT_NOTFOUND seems better. But as > you said: why

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-23 Thread Mladen Turk
On 09/23/2013 09:54 AM, Rainer Jung wrote: On 23.09.2013 07:27, Mladen Turk wrote: The patch seems fine. I mean any return value should do in theory. The main question is why is particular socket removed twice from the Poller. I agree that there's probably another problem further up the sta

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-23 Thread Rainer Jung
On 23.09.2013 07:27, Mladen Turk wrote: > On 09/22/2013 03:39 PM, Rainer Jung wrote: >> On 22.09.2013 13:17, Rainer Jung wrote: >>> I debugged around my occasional failures for TestCoyoteAdapter when >>> using APR. >>> >>> Error is: >>> >>> SEVERE [http-apr-127.0.0.1-auto-13-Poller] >>> org.apache.

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-22 Thread Mladen Turk
On 09/22/2013 03:39 PM, Rainer Jung wrote: On 22.09.2013 13:17, Rainer Jung wrote: I debugged around my occasional failures for TestCoyoteAdapter when using APR. Error is: SEVERE [http-apr-127.0.0.1-auto-13-Poller] org.apache.tomcat.util.net.AprEndpoint$Poller.run Poller failed with error [81]

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-22 Thread Rainer Jung
On 22.09.2013 15:39, Rainer Jung wrote: > On 22.09.2013 13:17, Rainer Jung wrote: >> I debugged around my occasional failures for TestCoyoteAdapter when >> using APR. >> >> Error is: >> >> SEVERE [http-apr-127.0.0.1-auto-13-Poller] >> org.apache.tomcat.util.net.AprEndpoint$Poller.run Poller failed

Re: Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-22 Thread Rainer Jung
On 22.09.2013 13:17, Rainer Jung wrote: > I debugged around my occasional failures for TestCoyoteAdapter when > using APR. > > Error is: > > SEVERE [http-apr-127.0.0.1-auto-13-Poller] > org.apache.tomcat.util.net.AprEndpoint$Poller.run Poller failed with > error [81] : [File descriptor in bad sta

Issue in AprEndpoint detected by TestCoyoteAdapter

2013-09-22 Thread Rainer Jung
I debugged around my occasional failures for TestCoyoteAdapter when using APR. Error is: SEVERE [http-apr-127.0.0.1-auto-13-Poller] org.apache.tomcat.util.net.AprEndpoint$Poller.run Poller failed with error [81] : [File descriptor in bad state] In the native world it is: 19372/128: port_a