Phantom Reboots ( WAS : is stable stable yet? )
danielc at green-orb.com
Tue Sep 16 10:37:44 PDT 2003
On 12 Sep 2003 at 13:39, Mike Tancsa wrote:
> If you cant get crash dumps then try at least hooking up a serial console
> to it so at least you see the panic saved on your console. If you are not
> seeing a panic, or if the panic is in random places, its probably hardware.
I tried this... unfortunately the serial interface on that particular unit is having its own
(apparently physical) issues... I've never had any good luck with serial ports actually
working (and I am using a high-quality null-modem cable ;-) on any of my boxes. So far (see
below)... its been up decently "okay" (1d+19h) on a very recent CVSup.
Not to be petty or anything, but this unit was rock solid before I CVSupped it at some point,
and handles build/install worlds like nobody's business... I would have a hard time believing
it was hardware all of a sudden, but trust me, I hear you and haven't ruled that out yet. I just
suspect its most likely to be some minor software thing that doesn't/didn't sit well with my
The motherboard is a PCChips 747, which isn't a particularly good one but I wasn't having
any notable problems on it up until these reboot issues. The memory is brand new (well, as
of a month ago) Micron 128MB PC133 DIMMS (running with a normally-clocked PII-450 at a
mere 100Mhz FSB). The NIC is an Intel 10/100 (uses fxp driver, running device polling).
ACPI is NOT compiled into the kernel, and when I tried it, it caused all sorts of hell on my
system like the system clock going backwards and all... so I gave up on that. APM *IS*
enabled though, and doesn't seem to be problematic. Don't know if I need it but it hasn't
seemed to ever be a problem before.
Also, not to jinx it now, but its been up longer now (1 day 19 hours) than it has since the
trouble started... on a CVSup from Sunday the 14th... seriously hoping it stays this way... (*
nervously checks to see if its still up now... *). I don't have any $$ for a new mobo. But if the
problems persist, I'll do things in this order unless advised otherwise:
1. Try to add another HDD with additional swap space for a crashdump (have a 4GB from
another unit on standby -- its a SCSI system and SCSI HDDs don't come cheap for a
personal machine - hence the reason I don't presently have enough swap space for the petty
384MB of RAM in it).
2. When it goes down, get a pciconf for you all (it won't let me right now -- says "access
denied" -- I'm assuming because I'm running securelevels) and of course other relevant
metrics if I can get any.
3. If the crashdump is useful, take appropriate action. If not, I'm going to try going back to
RELENG_4_8 to see if that makes any difference.
Any further suggestions?
Apologies for all the noise on the list if its just my stupid hardware and nothing useful to
> At 12:59 PM 12/09/2003, Daniel wrote:
> >My machine has been rebooting silently (no panic or anything) every 24-36
> >hours since
> >about two weeks ago. It wasn't doing that before then.
More information about the freebsd-stable