Now solved (we think/hope) read on: On Mon, 12 Sep 2005 09:53:10 -0600, jared r r spiegel wrote:
>On Mon, Sep 12, 2005 at 04:23:01PM +1000, Rod.. Whitworth wrote: >> >> >On Mon, 12 Sep 2005, Rod.. Whitworth wrote: >> > >> >> Any clues/ things to try? I am a bit worried about the build #137 as >> >> that's pretty close to release but I haven't heard anybody else >> >> reporting problems and the archives don't show any relevant posts. >> >> The 4801 has ddb.panic=1 and it isn't going to ddb and I'm fairly sure >> that, on the couple of times I was looking at the terminal when it >> rebooted it was just as though I had pressed reset - no message, but >> even at 9600 you can't be sure. >> >> I hope that if anybody else is getting weird random reboots on recent >> snaps they will put their hands up too so we can get a better handle on >> it. > > what about kern.watchdog.* ? > are you taking advantage of those at all? ( watchdogd, etc ) > > i'm running: > >OpenBSD 3.8 (GENERIC) #137: Thu Sep 1 17:41:20 MDT 2005 > [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC > > on a 4801, but not using any of the watchdog stuff currently. > >4801a $ sudo sysctl -w kern.watchdog >kern.watchdog.period=0 >kern.watchdog.auto=1 > > i could turn it on and see if it happens, as well as see if > there's a newer snap to try. gotta take the puppy to day > care and buy school books for the Mrs. now, though <G> > > jared > >- > >[ openbsd 3.7 GENERIC ( sep 1 ) // i386 ] > > No, I don't have a wtachdog running yet. I've been looking at watchdogd man page but did nothing so far. BUT here is the update: Whilst the 3.7 current boxes in the field still randomly reboot the Soekris 4801 here hasn't since I changed it over from a terminal to tip/OpenBSD (with capture on). I was leaning over a stack of 2 desktop boxes (not towers) with the terminal and the 4801 stacked on top of those. The terminal is just a pizza box that is plugged into a KVM so there's no display bit. Anyway as I pulled the comms cable out of the terminal to plug into the OpenBSD box below it the 4801 rebooted and I wondered what I had disrurbed so I checked the power into the 4801 - nice and firmly in. Thought I'd knocked something with the comms cable but biffing that a little caused no reboot. BUT as I stood up on tip-toe to get a better view of the back panel, vavoom reboot! This time I <knew> what I had disturbed. I had bumped a powerboard on the floor into which the SMPS for the 4801 was attached by a figure-8 plug and cord to a standard Aussie power plug for 240V. The pins on the latter were a slippery fit in the socket and after a slight twist was applied to both pins a solid tap with the size 11 causes no problem. A box in the field running the same build has had zero rebooots in sevearl days so I am reasonably sure that we are ready to roll on 3.8 as far as I can tell. So relax and enjoy another gem from the quality builders. Sorry if this caused any alarm but it is all part of testing. Now I need to build a power monitor to catch any power drops during testing to prevent false alarms.... Thanks to those with suggestions, Rod/ >From the land "down under": Australia. Do we look <umop apisdn> from up over? Do NOT CC me - I am subscribed to the list. Replies to the sender address will fail except from the list-server.
