https://bz.apache.org/bugzilla/show_bug.cgi?id=65975
--- Comment #19 from Martin Stangl <martin.sta...@t-base.pro> --- (In reply to Christopher Schultz from comment #18) > I think you misunderstood what I was suggesting: using Wireshark to monitor > *outbound* traffic during your testing. This is exactly what I captured. Every packet leaving the machine during the authentication process. > making an outgoing connection and timing out. Wireshark should definitely be > able to tell you *something* is happening. DNS lookup, or connection-attempt > to some remote service which fails after 60 seconds when you are connected > and fails *immediately* when you are not connected. Wireshark does not get anything if no interface is connected. And this makes sense to me, because to which of the not connected interface should the traffic go? It does not make it so far, therefore Wireshark will not get anything to capture. And as I mentioned before: Wait time is still 60 seconds, even with immediate network communication failures due to having no network connection. So the delay might still be related to communication, but is most likely not originating from the network code. The relevant request should be in the first 2 to 3 seconds of each capture, but there is nothing in there which is common to the three captures. The only SYNs without ACKs I see are related to some devices on the WLAN (a Huawai Honor 7, I could not identify the other device). Also they start a little too late and the retries end way before 60 seconds. So as far as I can tell, nothing goes out which relates to the issue. -- 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
