By stopping the JMeter test. It just violently kills all threads in test.
2015-02-09 15:45 GMT+01:00 Jon V. <[email protected]>: > How are you interrupting during SSL handshake? > On Feb 9, 2015 4:04 AM, "Ing. Jan Kaláb" <[email protected]> wrote: > >> The version is 2.0.9. >> >> I can't provide you the whole code, only what's in the question on >> StackOverflow, but the JMeter plan is attached. Maybe if you ask more >> precisly about part of the code you are interested in, I can give you >> that. >> >> I am pretty sure the session is not closed. The file descriptors for >> sockets are still there, and the server stops responding after a while >> (too many open files). But I found out it happens only when the >> connection is interrupted during the SSL handshake phase. If I >> interrupt the connection later (or sooner), the session is closed >> properly. If I fail the handshake for other reasons (bad cipher), it's >> also closed properly. >> >> So maybe I should file a bug for this? >> >> 2015-02-07 1:53 GMT+01:00 Emmanuel Lécharny <[email protected]>: >> > Le 06/02/15 14:05, Ing. Jan Kaláb a écrit : >> >> Hi, >> >> can someone please help me with this problem: >> >> https://stackoverflow.com/q/28366323 >> >> >> >> Thanks! >> > First of all, which MINA version are you using ? >> > >> > Also can you provide the code that you use, and the Jmeter plan ? >> > >> > FTR, calling session.close(true) does not close the session immediatly, >> > it just close the session regardless the waiting pending writes. What >> > happens is that the session is scheduled to be closed, and the >> > sessionClosed event is fired. As it's going to be done in two different >> > thread, it's perfectly possible that the thread handling the >> > sessionClosed() event might not be associated with any listener, as the >> > addListener() call might be executed after the session has been fully >> > closed. >> > >> > Thanks ! >> > >>
