After about 3 hrs of this, I decided to give my sb-16 something to do
too. Started the soundserver and a forever loop of auplay *.au. About
20 minutes later, I hear 'brrrPOP!' (no have brrrPOP!.au:), and very
shortly thereafter the aic driver began announcing drive timeouts,
restarts and 'trying harder' restarts. The only thing getting to
the console was aic messages, and there were only occasional blinkies.
(was 80 meg deep into swap at the time w. a ldavg of ~45. Worst case
was 160 mb swap at ldavg 145)
I let it choke along for about 15 minutes before intervening with
SysRq. I tried to SysRq-e running processes to no effect, so after
about another 5 minutes hit SysRq-l. This was also nogo, so I hit
SysRq-o (IKD Oops). This worked (boy did it ever work) so well that
it oopsed swapper and locked the machine (safety first).
I rather expected to give fsck a REALLY good workout upon reboot :))
but much to my suprise it found only a non-sunc syslog!! Hardly
believing my eyes, I dropped back to single user and forced another
check of all drives (4) and they came up clean. Naturally, the syslog
was truncated to before the event, so no logdata.
To check to see if there was any hidden damage, I finished the build
of glibc and ran make check.. it works fine as frogs hair! To check
the swap partitions, I ran another make -j for long enough to drive
the machine 80mb deep again.. no damage seemingly.
Any idea why the (very very) active fs wasn't munched? The only thing
I can come up with is..
'The Great Penguin' wards silly testers.
-Mike
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu