On 07/05/2012 17:24, Mladen Turk wrote: > Hi > > Did some testing and removing unlockAccept works equally > fine (and in some circumstances pause/continue and even shutdown is much > faster). > Using a simple socket close on stop/destroy and monitoring > for pause 'after' acceptSocket() by not handling single > connection received after pause eliminates the need for > unlockAccept entirely. > It also fixes various issues with disappearing network > stack, link-local addresses, and various hang-outs on > shutdown/pause.
I'm not entirely clear on what you are proposing. Can you provide a proposed patch for this? > Any objections that I commit that to trunk? At the minute, I'm not sure. > AFAICT this is not covered by any spec and the only > change is that a single connection that could get accepted > during pause phase can receive ECONNRESET instead ECONNREFUSED > in case pause continues beyond client socket timeout. That will be bad for connections from a reverse proxy when taking a Tomcat instance out of a cluster. I'm leaning towards objecting to this patch on that basis alone. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
