https://bz.apache.org/bugzilla/show_bug.cgi?id=66004
--- Comment #7 from Yann Ylavic <ylavic....@gmail.com> --- (In reply to Shun F from comment #6) > What was wrong with 2.4.53? The issue was with the non-accounting of children processes stopped by reloads (ungraceful restarts). 2.4.53 only accounts for processes started and stopped during runtime maintenance (namely perform_idle_server_maintenance() which honors Min/MaxSpareThreads and restarts children exited after MaxRequestsPerChild), while reloads also kill processes outside this mechanism thus outside the accounting (note that graceful restarts don't use this mechanism so are not affected by this issue). At some point, with bad counters, httpd thinks it can't create new children even though none is actually running (besides the main process). The fix consists in updating the counters for started/stopped children in a common function called by both runtime maintenance and reloads. -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
