On Mon, Mar 09, 2020 at 06:47:10PM +0100, Sebastien Marie wrote:
| On Mon, Mar 09, 2020 at 04:51:00PM +0000, Anthony Campbell wrote:
| > On 09 Mar 2020, Otto Moerbeek wrote:
| > > On Mon, Mar 09, 2020 at 03:56:53PM +0000, Anthony Campbell wrote:
| > > 
| > > > This discussion is very interesting. The same thing happened to me
| > > > on 6 March, when after completing the upgrade my Dell Optiplex 3020
| > > > refused to boot. I assumed it was a hardware failure and spent the
| > > > next three days bringing up an older Acer n460 which the Dell had
| > > > replaced.
| 
| yes, it looks like a hardware failure.

Indeed it did :)  My machine would not POST anymore (Dell Optiplex
9020; dmesg at the end)

| in my case, 4 hosts with the same motherboard model failed at the same time (I
| ran sysupgrade via ansible), so hardware failure was a bit excluded.

I only have this one machine that showed the behaviour.  Several VMs,
my gateway and my laptop worked fine so I didn't really tie it to the
bootloader changes (especially since the machine didn't POST).  I
couldn't boot from any other medium as long as the boot disk (an SSD)
was connected; my conclusion was that a failed SSD prevented the
system from POSTing (something I've seen in the past with failed
HDDs).

| > > > I don't have the facility at present to put the disk in another
| > > > machine so it looks like I'm stuck. 
| 
| I agree it could be difficult. If the disk is plugged, bios stuck. If the disk
| is unplugged, bios is fine, but you can't modify the disk data.
| 
| As sthen@ said, you could try to change bios setting to make the bios to not
| look at the disk. I dunno if it would work or not.

I played around with that a little bit, but didn't get to a working
machine.

| Alternatively, if you disk support hotplugging (sata disk should), try to
| connect the disk after the bios started could help. If so, I would try to plug
| it as soon as possible after bios init.

That was a bit of a scary option for me :)

| Depending your configuration, you could also try to use USB/SATA or USB/IDE
| adapter (depending your disk), in order to plug the disk after bios init. For
| me, I had problem with this method too: when my sata disk is plugged in sata
| connector it is showed with 512 bytes/sector, whereas with USB/SATA connector 
it
| showed with 4096 bytes/sector and so disklabel is incoherent.

In the end, after reading Otto's mail about reverting his changes, I
connected the SSD from my not-booting machine to my laptop and
upgraded the snapshot on it.  That allowed my desktop machine to boot
properly again.

I've seen Otto's commit message from earlier today, so I will test out
the next snap on my machine tomorrow.  At least now I know not to jump
to conclusions about failing hardware :)

Thanks to Otto for his work on this area; looking forward to running
my machine on all-ffs2.

Cheers,

Paul

-- 
>++++++++[<++++++++++>-]<+++++++.>+++[<------>-]<.>+++[<+
+++++++++++>-]<.>++[<------------>-]<+.--------------.[-]
                 http://www.weirdnet.nl/                 

Reply via email to