Date: Fri, 15 Sep 2000 13:22:54 -0500 From: Dan Nelson <dnelson@emsphone.com> To: Vadim Belman <voland@lflat.org> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Live debugging of a process being hung in a syscall. Message-ID: <20000915132253.A27598@dan.emsphone.com> In-Reply-To: <20000915142543.A3697@lflat.vas.mobilix.dk>; from "Vadim Belman" on Fri Sep 15 14:25:44 GMT 2000 References: <20000915142543.A3697@lflat.vas.mobilix.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
In the last episode (Sep 15), Vadim Belman said: > It seem like I got a NFS-related bug here where a httpd process > hung in a uninterruptable wait (a disk operation, most likely). In order to > locate the problem I need the process' stack trace first. > > gdb doesn't attach to the process for obvious reasons. Making a > crashdump doesn't inspire me at all. > > The question is: is there a way of working with /proc entries? I.e. > is it possible to get all what I need from, say, /proc/<PID>/mem? You can run "gcore <pid>" to generate a coredump from a running program. I do this all the time to programs that I *don't* want to attach directly to with gdb. -- Dan Nelson dnelson@emsphone.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000915132253.A27598>