Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 24 Jul 2000 23:32:17 +0800
From:      Victor Sudakov <sudakov@sibptus.tomsk.ru>
To:        "Andresen,Jason R." <jandrese@mitre.org>
Cc:        questions@FreeBSD.ORG
Subject:   Re: getting out of ddb(4)
Message-ID:  <20000724233217.A18163@sibptus.tomsk.ru>
In-Reply-To: <397C51F2.28A4FDCF@mitre.org>; from jandrese@mitre.org on Mon, Jul 24, 2000 at 10:25:55AM -0400
References:  <20000724221402.A17077@sibptus.tomsk.ru> <397C51F2.28A4FDCF@mitre.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jul 24, 2000 at 10:25:55AM -0400, Andresen,Jason R. wrote:
> > 
> > 
> > After a crash the box got into the kernel debugger. All my attempts
> > to leave the debugger and continue operation were in vain. What I did:
> 
> You don't want to try to keep using the kernel after it has crashed. 

I thought I know better :)

> The last thing you need is to have your disks trashed by trying to run a
> kernel in some inconsistant state.

OK, why is it still considered inconsistent after a cold restart?

>  
> > 1. Tried 'continue'. It just prints the trap message once again.
> 
> The problem is still there, you just ran into it again.
>  
> > 2. Tried 'panic'. It says 'Panic from debugger, Uptime 0s' and hangs.
> > No disk activity, no more messages. I waited about 5 minutes.
> 
> Wasn't the kernel already paniced?

Fine, then what's the command to exit the debugger and run in normal
mode of operation? After a restart, if necessary.

>  
> > 3. Cold restart. After a cold restart the kernel would drop into ddb
> > again and again. There was no way to boot the box.
> 
> That kernel is broken, you need a good kernel to boot the box.

The kernel itself is not broken. The reason of the crash was the failure of a
remote Netware server whose volumes were mounted with mount_nwfs over IPX.

> 
> > The only way out for me was to mount the fixit floopy and
> > 'mv kernel.old kernel'. Only after that I was able to boot. BTW I
> > found out that chflags was not on the floppy, but it is another story.
> 
> Yep, get the good kernel and it comes up.  BTW, if the other kernel was

The kernel WAS good.

> on your root partition, you can load it manually during the boot
> sequence by hitting something other than [enter] at the boot prompt and
> typing "boot kernel.old"

Of course I tried that first. It said 'kernel is already loaded'.

>  
> > Is this behaviour of ddb correct? How do I get out of it next time?
> > And why did cold restart not help?
> 
> Yep, ddb was trying to let you debug your broken kernel.  You can avoid
> this by not running broken kernels. :)  

I dislike the idea of an operating system trying to be smarter than the
admin. And I want to know, how I can work around the ddb trying to be
smarter than myself. Does it mark the kernel as broken somewhere in
its body?

> 
> BTW, did this happen on a newly built kernel or did this just suddenly
> happen to a good kernel?  

It is a perfectly working kernel. I had explained the reason of the
trap earlier in the message. The same trap can be caused by removing a
mounted floppy from the drive.

-- 
Victor Sudakov,  VAS4-RIPE, VAS47-RIPN
2:5005/149@fidonet http://vas.tomsk.ru/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000724233217.A18163>