On 2007/11/28 22:27, Daniel Ouellet wrote: > > I guess, that may be the only valid course of action here then. That may be > shown by the difference in the iic1 code in dmesg between not working boot > and stable one where so far, I am up to three time out of may be 50 or 60 by > now in the last few hours that are stable and doing a diff between the dmesg > doesn't show any differences between them. > > Could it be in the iic1 section of the dmesg then?
That's an unsupported device on the iic bus, probably an unsupported sensor, in which case the values will change depending on voltage/ temperature/fan speed/whatever else is measured. The sensors can either be read directly over the iic bus, or via the service processor (and read from there using ipmi). Sensors claimed by ipmi don't attach to iic, my understanding is that on some machines it can result in instability to access them from two places. Hmmm.... how about trying 'disable nviic' > This was fun for some time, but now it sure start not to be so much fun. At least you have the option of using i386 on them for now .. If you're completely fed up with the machine, I notice that an OpenBSD developer in Brisbane is currently asking for an X4100 on http://www.openbsd.org/want.html to work on RAID, not to mention a couple of other requests for amd64 boxes (Europe/N.America) to work on large memory support. ;-) (I'm half-joking here, but I think it doesn't hurt to remind misc@ readers that developers are better able to make devices work well if they use them themselves!)
