Date: Tue, 5 Feb 2002 07:20:02 -0800 (PST) From: Laurent Wacrenier <lwa@teaser.fr> To: freebsd-bugs@FreeBSD.org Subject: Re: bin/32295 Message-ID: <200202051520.g15FK2Z12710@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR bin/32295; it has been noted by GNATS. From: Laurent Wacrenier <lwa@teaser.fr> To: Alan Judge <Alan.Judge@eircom.net> Cc: freebsd-gnats-submit@freebsd.org, lwa@teaser.fr Subject: Re: bin/32295 Date: Tue, 5 Feb 2002 16:19:24 +0100 Alan Judge wrote: > Laurent, have you had any more info since you submitted this report? Not much. The problem hasn't been repeated on my servers since the report. I've looked in MySQL source code, the SIGALRM signal sound to be used to check time to time is the server or the connection hasn't been killed. MySQL provide a CFLAGS/CXXFLAGS (described in the INSTALL-SOURCE) named DONT_USE_THR_ALARM to avoid this, but some feature are disabled when you use it. It's not easy to debug random bugs on a multi-thread process with heavy load :( but if you want to, you should seek why the signals are queued but never flushed. Maybe the signal is blocked somewhere in MySQL or maybe it's a bug in libc_r that does not advertise the application. I guess it's the second option (you should compile libc_r with debugging symbols to check here). To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200202051520.g15FK2Z12710>