Re: backing up kernel and initramfs for future rescue use with upslug2
On November 26, gordonfarquharson@gmail.com said:
> Yeah, I know, but I was discussing the problem via email, so in this
> case, apart from the reason you mention above, there was no reason
> to file a bug report.
Another use of bug reports (especially ones about bugs that make
machines unbootable, and would therefore be critical) is to keep
packages from filtering down into testing while they're in a dangerous
state.
> Very odd. I have never seen this behaviour. The most similar report
> that I have heard of was due to a hung RTC
> (http://article.gmane.org/gmane.comp.misc.nslu2.linux/16483). Maybe
> the posts in that thread can help (you can read the whole thread
> using Yahoo Groups or Gmane).
Wow, thanks! I think you hit the nail on the head. hwclock --show
drives the CPU to 100% kernel.
When i remove the rtc-dev module, hwclock --show fails, but at least
it returns immediately.
The next time i'm physically near the machine, i'll try reseating the
battery to clear it up, and report back if that clears my problem.
> Can you try removing ntpdate (i.e. dpkg --deinstall ntpdate) and see
> if the behaviour changes. I don't have ntpdate installed on my
> slug. I think that ntpdate is no longer needed if you have the ntp
> package installed because ntp now sets the clock initially
> (http://packages.debian.org/testing/net/ntpdate).
i'll look into this also. i'm used to having ntpdate around, because
it's handy to query external servers, but it'd be nice to not need it
at boot time.
Thanks for the good info, Gordon!
--dkg
Reply to: