Hi Chris, On Mon, Mar 18, 2013 at 07:14:08PM +0000, Chris Allen wrote: > Hi, > > This weekend our haproxy died with a "SIGABRT". As luck would have it, > apport was running > and it kept a core, details of which I've posted below. Is this a known > bug? Or did we do something > wrong? Any insight would be appreciated.
No this crash is new and totally unexpected. Could you please send me the core and the executable (in private) ? That would help a lot, I'll disassemble the code to find the fd causing the issue, because I don't see why this one would not be present in 1.4.22. Is there any reason you're using select() instead of poll() or epoll() ? Thanks, Willy
