Date: Fri, 15 Apr 2005 10:18:16 -0700 (PDT) From: Doug White <dwhite@gumbysoft.com> To: Eirik =?ISO-8859-1?B?2A==?=verby <ltning@anduin.net> Cc: amd64@freebsd.org Subject: Re: Mysterious hangs Message-ID: <20050415100931.O34838@carver.gumbysoft.com> In-Reply-To: <BE829A83.116B6%ltning@anduin.net> References: <BE829A83.116B6%ltning@anduin.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 13 Apr 2005, Eirik [ISO-8859-1] =D8verby wrote: > Sorry, this was meant for the amd64 list. > Got another report for this list. Confused the addresses. > > On 13-04-05 07:28, "Eirik =D8verby" <ltning@anduin.net> wrote: > > > Hi, > > > > For some time I've been seeing infrequent and seemingly random hangs on= my > > dual opteron system. Motherboard is Tyan K8S Pro. Until now I thought i= t was a > > "hard" hang, but I've got serial console hooked up and BREAK_TO_DEBUGGE= R, DDB > > and KDB in the kernel, and when sending a break through telnet I get: > > > > telnet> send brk > > KDB: enter: Line break on console I have a rev-D HDAMA that hangs like this. If you set debug.kdb.stop_cpus=3D0 then ddb will enter, which leads me to believe one of the CPUs is going out to lunch. I had a rev-G HDAMA that upgrading the BIOS fixed this on, so you might try that. I can only assume the BIOS difference includes a tweak to a timing somewhere; I've been wanting to mess with the HyperTransport clocks but haven't had time, but HT problems would likely cause wierdcCPU communication issues. I have an S2881 that does not exhibit the problem. > > And nothing more. Which means the box is sort of alive, but fails to pr= esent > > even the kernel debugger. The machine is currently on 5.4 as of yesterd= ay. > > It's been hanging before (with anything from days to weeks between), un= der > > various variations of 5.x, but I haven't had the serial console before = now. > > > > Has any one seen this before? Any ideas what I could look at? I suspect= ed > > hardware, but it seems that isn't the case as most HW errors I've seen = would > > cause a total freeze. > > > > Here's to hoping. > > > > Thanks, > > /Eirik > > > _______________________________________________ > freebsd-amd64@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-amd64 > To unsubscribe, send any mail to "freebsd-amd64-unsubscribe@freebsd.org" > --=20 Doug White | FreeBSD: The Power to Serve dwhite@gumbysoft.com | www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050415100931.O34838>