https://bz.apache.org/bugzilla/show_bug.cgi?id=67927
--- Comment #7 from Michael Osipov <micha...@apache.org> --- (In reply to Remy Maucherat from comment #6) > Since deallocating the ssl context should be tied to GC when using NIO, I > don't understand how it happens. OTOH, actually doing it is quite risky, so > maybe it could be better to leak it ... > > It would be a good idea to try the FFM code to debug this kind of issue > (shameless ad). It would produce cleaner traces and would be easier to > experiment with, although I doubt the native code itself would need a fix in > that particular case. I will try the FFM code next week. I am still convinced that all requests need to paused and the connector locked during security context reload. -- 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
