Unattended reboot was Re: signal 12's everywhere on Current with update this morning.

[Jesper Skriver]

On Thu, Nov 13, 2003 at 07:39:15AM -0800, eculp at encontacto.net wrote:
> Mensaje citado por eculp at encontacto.net:
> 
> | Mensaje citado por Ceri Davies <setantae at submonkey.net>:
> |
> | | On Thu, Nov 13, 2003 at 06:48:35AM -0800, eculp at encontacto.net wrote:
> | | > I installed an older current snap, Oct 26, on a brandnew dell power edge
> | | > with a single Xeon 2.4 GHz cpu and 1G in memory.  It was running great.
> | | > I installed everything except the kitchen sink.  Then I decided it was
> | | > time to update, I've got serveral other machines weathering the storms.
> | | > Bad idea.  It doesn't hang or anything that I can get my teeth into but
> | | > it just give signal 12 core dumps on many if not most applications.  I
> | | > caught this before finishing an install on another box, yesterday
> | morning.
> | | > One of the apps that generates signal 12 is ls so I tried pulling ls
> | | > from the other box and it no longer has a problem.  I am at a lost.
> | | > Any suggestions for where to start would be appreciated.
> | |
> | | Read /usr/src/UPDATING.
> | |
> | Thanks, I missed that.  :(   Sorry for the noise.
> |
> | ed
> 
> I'm building new kernels as I write this.  My next question is:
> One of the machines I'm building on is remote and was last rebuilt
> just before the change.  What would be be better sequence for making
> the change after a fresh cvsup ?
> 
>   1. Build and Install a new kernel
>   2. build a new world
>   3. Run mergemaster
>   4. ReBoot (I'm not sure it will come up multiuser with a new
>              kernel and a 4 day old userland.)
>   5. Installworld and assess possible problems :-)

I upgraded a ancient -current last night, I did

1) cvsup
2) build and install new kernel
3) reboot
4) build and isntall new world
5) mergemaster
6) reboot

/Jesper