Date: Tue, 26 Aug 2003 19:09:27 +0000 (UTC) From: readpunk <readpunk@sdf1.org> To: Matthew Seaman <m.seaman@infracaninophile.co.uk> Cc: freebsd-questions@freebsd.org Subject: Re: Signal 11's all over the place. Message-ID: <Pine.NEB.4.33.0308261906120.7031-100000@otaku.freeshell.org> In-Reply-To: <20030826105345.GC37256@happy-idiot-talk.infracaninophile.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
That is what I believe is happening, which is why I suspect the cpu and not memory. If I do anything that stresses the server (raises the cpu% even when the server is reporting more than enough ram) it will just completely lock and sometimes reboot on it's own. I mention the sig. 11's because I am seeing them in the logs httpd/perl/hlds/top etc. The box has locked up while _only_ running memtest, ha ha. I suppose I need to take out each dimm and really stress the box. If it does well with one and not the other I suppose my answer is found, if not then it has to be the cpu. Thanks for the help all. Nick On Tue, 26 Aug 2003, Matthew Seaman wrote: > Date: Tue, 26 Aug 2003 11:53:45 +0100 > From: Matthew Seaman <m.seaman@infracaninophile.co.uk> > To: readpunk <readpunk@sdf1.org> > Cc: freebsd-questions@freebsd.org > Subject: Re: Signal 11's all over the place. > > On Tue, Aug 26, 2003 at 05:10:20AM +0000, readpunk wrote: > > Signal 11's from different programs and near complete instability usually > > means a fan/heatsink issue, correct? > > > > FreeBSD 4.8-release > > Athlon 2400 XP+ > > 1 gigabyte of RAM > > > > Is anything else really necessary? Occasionally right after a reboot when > > I run top the thing crashes (the machine is remote) when it does get a > > little spat of stability for whatever reason it seems to allocate ram fine > > and run properly. Thanks to anyone who reads this. > > It might be overheating -- but that generally results in the system > simply freezing up when the CPU thermal cutout engages. Instinct > tells me that seeing a lot of Sig 11's like you are could very well be > due to duff memory -- it's not impossible for the memory sticks to > overheat but more likely they've just developed a bad spot. > > If you have console access to the machine, try running a memtest86 > floppy (http://www.memtest86.com/) for a few repetitions (or get the > NOC people at your hosting center to do it for you) That will take > most of a day probably. Note that when memtest86 does find a problem > then it's almost always genuine, but some subtle problems can elude > it, so getting the all clear from it doesn't completely discount > problems with the memory. > > Otherwise, try pulling out each half of the memory sticks in turn and > see if that can isolate the fault. However, that won't help you if > the problem is within the CPU, unless you can lay your hands on a > known good spare to swap in and test with. > > Cheers, > > Matthew > > -- > Dr Matthew J Seaman MA, D.Phil. 26 The Paddocks > Savill Way > PGP: http://www.infracaninophile.co.uk/pgpkey Marlow > Tel: +44 1628 476614 Bucks., SL7 1TH UK > /* Try unix. Then ./revolution */
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.4.33.0308261906120.7031-100000>