On Fri, Jan 10, 2020 at 10:06:41AM +0100, Jan Stary wrote:
> On Jan 09 11:44:25, h...@stare.cz wrote:
> > Installing bsd      100% |**************************|  6248 KB    00:05 ETA
> > Installing bsd.rd   100% |**************************| 11229 KB    00:10 ETA
> > Installing base66.tgz       100% |**************************| 99116 MB  - 
> > 07:12edT-
> > Installing comp66.tgz        83% |*********************     | 45312 KB  - 
> > stalledT-syncing disks... done
> > rebooting...
> 
> On Jan 09 14:48:25, dera...@openbsd.org wrote:
> > > https://marc.info/?l=openbsd-cvs&m=156889553923923&w=2
> > > Looking at install.sub, it seems that the watchdog timer (30 min)
> > > is reset after installing every set. I don't see the watchdog being
> > > reset whenever there is "progress" - did you mean the completion
> > > of a set (as opposed to just not being -stalled-)?
> > > 
> > > Also, the resets only happen with $UU, which is only set with -x
> > > - where should that option be passed? (not in sysupgrade.)
> > > 
> > > Grandpa's analog stopwatch on a chain says
> > > that it indeed reboot around the 30 min mark.
> > 
> > This entire design is intentional.
> > 
> > If a machine gets stuck doing an upgrade, we want it to abandon the
> > upgrade and go back to the /bsd kernel (which hopefully has not been
> > installed yet).
> 
> Well, /bsd is the first thing that gest installed,
> and is relatively small compared to the other sets.
> In this particular machine, the watchdog timeout hits when
> bsd, bsd.rd, base, and half of comp have been installed.
> 
> > Your machine is too slow.
> 
> It seems it's the SD card that is slow (the machine
> is a BeagleBone Black) - will try with a faster one.

Support for using edma with ommmc(4) is lacking, that would help.

> 
> It seems I am missing out on
> http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib/miniroot/install.sub.diff?r1=1.1141&r2=1.1142&f=h
> - I can't figure out how to pass the -x option that sets $UU
> (and thus makes the timer reset before each set is installed).
> 
> Thank you for the insight.
> 
> 

Reply via email to