Hey folks, Anyone else running haproxy? I went directly from 16.4.1 to 17.2.0 for a bunch of native 64 bit zones that I'm running and found the following behavior with a very simple config file:
16.4.1 / haproxy-1.6.9nb1-- works as expected 17.1.0 / haproxy-1.7.3 -- high cpu - eats one whole core 17.2.0 / haproxy-1.7.5 -- high cpu - eats one whole core truss(1) on the process shows a tight loop of "pollsys(0xFFFFFD7FFFDFF970, 3, 0xFFFFFD7FFFDFFA90, 0x00000000) = 1". Starting by hand in debug mode with poll disabled (force select()) -- "-d -dp" restores good behavior, but adding supposedly equivalent flags to the global stanza (noepoll nopoll) doesn't work. Moreover, daemonize with poll disabled - "-D -dp" results in high cpu too. Couldn't find anything suspicious in the changelog at http://www.haproxy.org/download/1.7/src/CHANGELOG Anyone else see similar behavior? Moreover, does anyone have haproxy behaving properly in 17.1 or 17.2? Config file below. -r resolvers dns nameserver goog 8.8.8.8:53 hold valid 10s frontend letsencrypt bind :::80 v4v6 mode tcp use_backend ansible timeout client 15000ms backend ansible mode tcp server ansible ansible.example.com:80 check inter 30000 timeout connect 5000ms timeout server 50000ms ------------------------------------------- smartos-discuss Archives: https://www.listbox.com/member/archive/184463/=now RSS Feed: https://www.listbox.com/member/archive/rss/184463/25769125-55cfbc00 Modify Your Subscription: https://www.listbox.com/member/?member_id=25769125&id_secret=25769125-7688e9fb Powered by Listbox: http://www.listbox.com
