On Sun, Mar 14, 2010 at 12:26:01AM -0800, patrick keshishian wrote:
> if I control-C out of the install/upgrade script and run the command I
> *think* the upgrade script is running, base46.tgz installs just fine.
>
> This is the command I *think* is being run by the upgrade script but
> i'm not 100% positive:
>
> # ftp -V -o - -m file://mnt2/4.6/i386/base46.tgz | cat | tar zxphf - -C /mnt
>
> I'm hoping someone with much more familiarity with the install/upgrade
> scripts can give me a clue.
we had very few reports of this before, but this remains a mystery,
since rerunning it from the command line always seems work. It could
be related to /mnt being mounted async by the install script, but
actually that's just a wild guess and it does not explain why it works
from the command line. Anyway, rare and very hard to reproduce, so
even harder to diagnose and fix.
-Otto
>
> --patrick
>
>
> On Sat, Mar 13, 2010 at 11:29 PM, patrick keshishian <pkesh...@gmail.com>
> wrote:
> > I've gone through the steps outlined in release(8) to create a release
> > set which then i used mkisofs and cdrecord from cdrtools (ports) to
> > burn a CD image that I use to upgrade my other machines with. The last
> > two times I had to do this, after an errata update, I notice that
> > during the set install time base46.tgz stalls towards the end (this
> > time at 77%):
> >
> > base46.tgz 77% |***************** | 36648 KB - stalled -
> > (above is hand copied obviously)
> >
> > The sets are being installed off of the same PC's CD drive. So I'm not
> > sure what is this indicating?
> >
> > At first I thought maybe bad CD-Rs, but copying the .tgz sets to local
> > disk and using this path during the upgrade also stalls at the same
> > point (77% 36648 KB).
> >
> > I don't recall what I did to get past this last time.
> >
> > Any ideas?
> >
> > --patrick
> >
> > This is a GENERIC i386 4.6 system, following errata and using cvs up
> > -rOPENBSD_4_6
