[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Stretch System Stops Boot Process Immediately After Grub Screen



On 10/23/2016 01:33 PM, Felix Miata wrote:
James P. Wallen composed on 2016-10-23 12:14 (UTC-0400):

On 10/22/2016 18:10 (UTC-0400), Jape Person wrote:

It's confusing to see a response from a different person writing as if he was
responding to himself.

The confusion is caused by my idiotic tendency to confuse which e-mail account I'm using at any given moment. I generally use a separate e-mail account for mailing lists to help with organizational chores. Sorry about that. I did indeed respond to myself using a different e-mail account.

Be that as it may, have either of you tried intercepting Grub and unquieting
the boot process? Remove quiet, and either change splash to splash=0 or
remove splash entirely. Then proceed to boot, and see what if anything shows
up on screen besides a blinking underline cursor in the upper left corner.


Yes, both of us have tried making changes in the boot process.

Heh.

There is simply no change in the experience when removing quiet or using nomodeset. The disk access stops instantly when the grub screen disappears. No keyboard controls are effective.

The fact that just touching the power switch results in instant shut-down makes me think that the kernel has not even started to load. But computers are faster than I am, so I realize that this issue could be happening at the end of grub or the beginning of the kernel load.

Does the same thing happen booting the previous kernel (4.6?)?

Nope. The problem occurred on the first reboot after the upgrade from 4.7.6-1 to 4.7.8-1. The upgrade process didn't leave 4.7.6-1 in place so I could fall back.

I don't remember just when I started seeing that upgrade behavioral change in Debian. I used to always use the new kernel for a week or so, and then I would have to use apt to remove the old one if it was no longer needed.


Have you tried giving it 10 or more minutes before assuming boot won't complete?

I have a Stretch installation last updated about three weeks ago, which
installed a 4.7 kernel of 26 Sept. Booting it just now took >4.5 minutes to
get from the Grub selection to seeing boot messages appear on screen, but
from the point messages started appearing, boot proceeded normally.

I have >20 multiboot PCs with various distros. I've been encountering this
type of boot delay, sometimes as long as more than 13 minutes, with random
kernel/initrd pairs in several different distros, for going on two years.
It's happened only when Dracut builds the initrd, and only with 64 bit
installations. Whenever I've asked anywhere about this I've gotten zero
useful response, if any response at all.


I've left it at the blinking cursor for hours without seeing any change.

I have also been seeing the same odd boot delays (and shutdown delays) on my 64 bit installation for months in Stretch. The delays are not consistent in behavior or duration. I have not been able to find any way to gather useful data. Yeah, kind of annoying. But it's different from what I'm seeing on this i386 installation.

Incidentally, sometimes on the amd64 image on another machine I've switched to TTY1 following logon to the GUI and seen systemd countdown messages for starting of various daemons still going on after the boot has apparently succeeded.

No intention of providing flame bait here, but if anything like daemon startup failures was happening before the switch to systemd, I was happily ignorant of it until I actually needed the daemon, and it didn't slow my boot or shutdown processes. Just saying.

;-)

Now I shall grin, duck, and run.

http://markmail.org/message/yj3l3uphno3cgpgp is probably where I first asked
publicly.


I'll check out your link to see if there's any way I can a) learn something, b) contribute something, c) blame it all on a politician.

Thanks,
JP


Reply to: