Le samedi 2 avril 2011 15:05:37, Willy Tarreau a écrit : > Would you like to send a bug report on [email protected], and CC me, > Eric Dumazet (the patch author), and maybe Charles who reported the bug ?
Done (had to resend the mail to fix the list address ;-) ) > Also, could you check if that can also be reproduced with other services ? > The most common one that comes to mind is SSHD. Just login to your machine > via SSH, kill -9 the main daemon and restart it. I suspect it will not be > able to bind anymore until you close your SSH session. I couldn't reproduce the issue with SSHD, killing the daemon allows it to restart and bind the the socket. Note that forcing haproxy to call soft_stop() works, whereas pause_proxy() will break. -- Cyril Bonté

