[iPAQ] RE: [Familiar] ipkg update; ipkg upgrade process

Hello all,

I have a small ipkg feature request related to doing a mass upgrade.

When ipkg finds some configuration files in /etc when upgrading that
have been edited by the user, it warns and asks whether to install the
ones from the new package or keep the existing ones. If you're not sure,
however, then you need to quit ipkg (Ctrl+c) and then manually back it
up. This is fine for one file, but a pain for upgrading the system. Can
we please have an extra option that backs up the existing file - e.g.
say ipkg.conf needed backing up to ipkg.conf.08jul2002 ?

Any chance of having this implemented? I just lost my stowaway setup
because I thought the upgrade process would be intelligent enough to
realise I needed the relevant modules/config information...It wasn't :/

Cheers,

Adam.

-----Original Message-----
From: familiar-admin@handhelds.org [mailto:familiar-admin@handhelds.org]
On Behalf Of Brian F. Hensch
Sent: 10 July 2002 16:22
To: familiar@handhelds.org
Subject: [Familiar] ipkg update; ipkg upgrade process

Hi,

I had minor problems with the ipkg update; ipkg upgrade process in an
attempt to upgrade from 0.5.2 to 0.5.3. Following the directions I made
the changes to /etc/ipkg.conf and then cleaned up some space by removing
some packages I would add back later if I still had room. I had 957k
free
when I issued the update command.

Everything was going very well until the kernel package was upgraded.
It
made it past the download just fine. I waited 15+ min. for it to
complete
the upgrade before I decided it was stuck. I may have jumped the gun.
There was a defunct ipkg process running with a couple sleeping ones
when
I hit ^C.

I tried upgrading the package by itself and did a ipkg upgrade a couple
more times. The last time seemed to work and only took a ~minute. Now
most of the packages are doing the same thing, taking a long time to
complete but there aren't any long running defunct ipkg processes. I'm
watching the disk usage and it floats around 95% but jumps briefly to
99%.
Does anyone else think this is a space issue? Why is it randomly
working
and then not working?

I did find this message once in the /proc/kmsg.
<4>jffs2_read_inode(): No data nodes found for ino #4936
<5>Eep. read_inode() failed for ino #4936

The real mistake I know I made was not to reboot with a serial cable
attached and kernel console param. set so I could report error messages.
It seems a little risky to do so now.

FYI: After the process has completed I have 636k available on root but I
want to try to verify I have everything before I reboot. Any
suggestions?

-- 
Brian F. Hensch
Farmington, MN
_______________________________________________
The Familiar Linux Distribution
Familiar mailing list
Familiar@handhelds.org
http://handhelds.org/mailman/listinfo/familiar
irc://irc.openprojects.net #familiar