On Wed, 2009-03-04 at 14:48 -0800, Todd Cary wrote:
> Bill -
>
> I did remove some of the older files (only a few) and there was then
> enough room to unpack the update. Is there some basic guideline on
> how many to remove?
I suspect you have a minor mess-up in /etc/yum.conf (man yum.conf).
IIRC, installonlypkgs (defaults to various kernel name variations)
interacts with installonly_limit (default 3?). Somewhere, I can't recall
where ATM, when the Xth version, specified by installonly_limit, is
exceeded the oldest package version is automatically removed. However I
seldom play with this stuff and haven't reviewed it in a long time so
I'm unsure about this. I do know that the grub.conf (menu.lst) is
automatically updated to remove the older versions. Maybe I'm confusing
the two?
Let's hope one of the folks that are current and more expert will chime
in.
Now, if you made the mistake I once made, this all breaks. I had a
fall-back boot scheme set up that had a second disk bootable and a copy
of root, /var, etc. on it to allow fast recovery (LVM snapshots were
automatically created at various times to complete the process). I then
set that backup boot disk's /root into my fstab and forgot to remove it
after testing. Yum updates ran on it and never updated my primary root.
Being (apparently) brain-dead I blithely copied stuff over never
realizing what I was doing.
Fortunately, the only real repercussion was when things got pretty full
and I realized what I had done I received numerous bruises from
self-flagellation with a locally available and wielded clue-bat! :-(
As to your current clean-up effort, I think Craig mentioned the rpm
method. That's what I would do. Just be careful that you double-check
the version before you hit enter or you will not have enough clue-bats
available to administer sufficient punitive lessons.
>
> Todd
> <snip>
--
Bill
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
