Etch->Lenny upgrade gone bad
I am hoping for some help with a problem, an "aptitude full-upgrade"
gone bad. The history:
1. Two NSLU2s (hence "slugs", of course), stock (unturboed, unfat)
2. One runs Lenny, installed from the Debian installer
3. The other runs Etch, also installed from the Debian installer back
in the day (I helped diagnose issues with ssh timeouts - see the
wiki); has an 8GB swap partition (more on this later)
4. I decide, "time to upgrade the Etch slug"
5. Update sources.list to point to Lenny [1]
6. aptitude update
7. aptitude full-upgrade
8. watch the oom-killer whack aptitude, many times, in different and
terrifying ways; go back to step 6, or sometimes do a "dpkg
--configure -a", if it tells me it needs that
9. try adding more swap (1GB via dd and mkswap) - problem gets worse
(for in retrospect obvious reasons [2])
10. create 256MB swap file; swapon 256MB file, swapoff 8GB partition
and 1GB file
11. watch aptitude (and a dpkg --configure -a) "smoothly" finish the
upgrade, including:
:
Setting up build-essential (11.4) ...
Processing triggers for initramfs-tools ...
update-initramfs: Generating /boot/initrd.img-2.6.26-2-ixp4xx
Flashing kernel: done.
Flashing initramfs: done.
sh-3.2#
12. install locales it hadn't been there, and I'm not sure if it was
the victim of the oom-killer or wasn't there in etch for some
reason; my logs are full of:
Setting up <some-module> (X.yy) ...
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LANG = "en_US.UTF-8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_CTYPE to default locale: No such file or
directory
locale: Cannot set LC_MESSAGES to default locale: No such file or
directory
locale: Cannot set LC_ALL to default locale: No such file or directory
13. aptitude update; aptitude safe-upgrade; - Yay! - it says I don't
need any more stuff!
14. check my ethernet config; looks OK
15. sync; sync; # deep breath
16. reboot
17. wait; and wait; nothing; no ping response; lights are on green
Disk 1 and Ready/System LEDs; except for a brief time at power up,
no Ethernet LED activity
18. power down hard; plug disk in to workstation; look at var/log/,
and nothing is modified since reboot
19. plug back into slug and try again; after a while, only Disk 1 LED
is lit
That's a lot of steps. I'm guessing that I ended up with a bad flash.
At least I haven't lost the data on the drive (yet).
I mentioned that I have a "good" Lenny slug. I'm hoping there's a
clever way for me to flash its vmlinuz (and anything else from the good
Lenny install) to the former-Etch slug. I'm open to other suggestions;
I haven't tried (yet) plugging my etch->lenny disk to the good lenny
box; aside from MAC address and IP address, is there system-specific
information in the flashed data?
In fact, I happen to have a third slug - still in box - and I'd love to
have a clever way to clone one of the other two without having to go
through the full Debian Installer (or the even slower aptitude
[full|safe]-upgrade process). Seems that if my problem is indeed a bad
flash, I could use that technique and some judicious disk copying to
create new versions. But that's a different topic...
Sorry for a long first post here; I've had worry-free slug operation in
the past, and thought I had the ducks in a row. Thanks for any ideas,
-Del
[1] These steps are somewhat abridged, actually. So there's possibly
something else I did that didn't help.
[2] Turns out I should have read about swap space and Linux more
closely; bigger swap puts *more* pressure on low-memory and hastens the
onset of the oom-killer. *sigh*
Reply to: