maybe there's an open write lock on /dev/log that keeps the send() hanging?
On Wed, Feb 4, 2009 at 8:39 PM, Jeremy Visser <jeremy.vis...@gmail.com>wrote: > On 02/02/2009, at 11:56 AM, Daniel Pittman wrote: > >> That should let you see which syscalls were delaying the process, and so >> narrow it down to a specific set of options... >> > > Did you read my post 30/01/09 7:14PM? It was basically a running commentary > of it, as I was watching the strace, not just logging it to a file. > > The actual syslogd daemon did not appear to be locked up. It didn't even > hang when I sent a normal HUP to it -- it quits cleanly. > > Just the socket communication between apps and /dev/log seems to be hanging > (though killing syslogd stops the hanging). > > As for the actual apps that hang as a result of the dead communication, > they hang on send(), when writing the log message. > > It's strange that they hang on send(). connect('/dev/log'), which is what > I'd expect them to hang on, returns instantly. > > -- > SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ > Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html > -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html