[3471] in testers
Re: finish-update and other things
daemon@ATHENA.MIT.EDU (Greg Hudson)
Thu May 21 16:41:48 1998
To: Brett David Rosen <bdrosen@MIT.EDU>
Cc: testers@MIT.EDU
In-Reply-To: Your message of "Thu, 21 May 1998 16:19:48 EDT."
<199805212019.QAA01106@cogsworth.mit.edu>
Date: Thu, 21 May 1998 16:41:34 EDT
From: Greg Hudson <ghudson@MIT.EDU>
> /etc/init.d/finish-update (/srvd/etc/init.d/finish-update)
> Has the line:
> /usr/sbin/in.named
Yeah, known bug. I'm surprised your machine hung in config_afs
because of it, though. Certainly didn't happen on my test machines.
> Also, I manually updated this machine (to a newer version of the
> alpha packs) and ran into all sorts of trouble because my / and /usr
> filled up somehow towards the end of the update (while it was
> updating the kernel) and then my machine rebooted, leaving me with
> an unusable machine. I'm not sure how my partitions filled up in
> spite of the space checks in the release, but it seems like it might
> make sense to update the kernel and necessary utilities earlier in
> the process to avoid this sort of problem.
It's not really possible to make sure the machine is always useable
during the track. That's why I'd like to have a miniroot Solaris
update at some point, but there aren't any specific plans to make it
happen.
I'll check the space requirements of the update again. It seems like
every time I turn around we're consuming an extra few megabytes.
> Finally, I seem to have ended up with a /var/kernel tree somehow
> (same as /kernel) and I wanted to make sure that it is something
> that can be safely removed? Is this something that the release
> created? (I may have put it there by accident when I was fixing the
> problem with my unusable kernel)
Never heard of it.
