Long boot time, can't explain why. [SOLVED]

[Anghirrim]

Hi crew,

I am new to this forum, but not to Debian nor the Linux users world. I've been using various distros since 2007, including Debian from time to time.

But since some weeks, I have a crazy long boot time I can't explain, even after googling the problem. I think I am mislead by the ACPI error message, that seems to be a common bug with no impact on performances.

I do have some systemd networking.service problem I could solve. But my kernel time is long and I can't figure out why. Could you please help me figuring out why?

Here is the startup screen:


Systemd-analyze output:

Code: Select all

anghi@asus-k53sd:~$ systemd-analyze time
Startup finished in 32.501s (kernel) + 9.527s (userspace) = 42.029s
anghi@asus-k53sd:~$ systemd-analyze blame
          5.955s NetworkManager-wait-online.service
          1.575s networking.service
          1.461s dev-sda4.device
          1.178s dev-loop1.device
          1.178s dev-loop0.device
           615ms apt-daily.service
           411ms systemd-rfkill.service
           403ms apt-daily-upgrade.service
           377ms ModemManager.service
           337ms mnt-Donnees.mount
           257ms NetworkManager.service
           222ms accounts-daemon.service
           213ms lm-sensors.service
           209ms polkit.service
           160ms keyboard-setup.service
           145ms upower.service
           122ms systemd-timesyncd.service
           100ms systemd-journald.service
            98ms snap-core-4486.mount
            95ms alsa-restore.service
            93ms rsyslog.service
            90ms snap-core-4571.mount
            86ms user@1000.service
lines 1-23...skipping...
          5.955s NetworkManager-wait-online.service
          1.575s networking.service
          1.461s dev-sda4.device
          1.178s dev-loop1.device
          1.178s dev-loop0.device
           615ms apt-daily.service
           411ms systemd-rfkill.service
           403ms apt-daily-upgrade.service
           377ms ModemManager.service
           337ms mnt-Donnees.mount
           257ms NetworkManager.service
           222ms accounts-daemon.service
           213ms lm-sensors.service
           209ms polkit.service
           160ms keyboard-setup.service
           145ms upower.service
           122ms systemd-timesyncd.service
           100ms systemd-journald.service
            98ms snap-core-4486.mount
            95ms alsa-restore.service
            93ms rsyslog.service
            90ms snap-core-4571.mount
            86ms user@1000.service
            83ms systemd-udevd.service
            75ms systemd-tmpfiles-setup-dev.service
            67ms systemd-modules-load.service
            66ms sys-kernel-debug.mount
            64ms dev-mqueue.mount
            63ms systemd-udev-trigger.service
            61ms avahi-daemon.service
            52ms colord.service
            50ms speech-dispatcher.service
            48ms lightdm.service
            47ms udisks2.service
            44ms snapd.autoimport.service
            41ms kmod-static-nodes.service

See below the .svg
https://drive.google.com/open?id=1QHuVg ... eVNFdf4PR0

System config:
Debian 9.4 x86-64
Kernel 4.9.0-6-amd64
Core i5-2450m
Geforce 610m
/ on a ssd, ext4 on sda
Grub on /dev/sdb, a HDD (for my wife to avoid seeing the Grub screen by default everytime she uses the laptop. So I manually boot on sdb to get Grub)
Dualboot with W10


Thanks all

[stevepusser]

Have you tried installing haveged yet?

[Anghirrim]

HI, no.

I've been to the Haveged Wiki page (Archlinux, sorry ), but can't figure out the use for my problem.

[Head_on_a_Stick]

https://lists.debian.org/debian-user/20 ... 00866.html

EDIT: also:

Code: Select all

# journalctl -u networking.service

[Anghirrim]

Thanks @Head_on_a_Stick ,

What mislead me was that damn /dev/sda4 before the suspend/resume message while my swap is on /dev/sdb2.

And I had commented the swap line in my /etc/fstab some time ago, thinking this would remove one parameter to troubleshoot.


So, I followed the link indication and it worked. But to make it work, the weird thing is that I had to create a new UUID for /dev/sdb2. Using the existing UUID, the update-initramfs command failed in finding the swap device... After I changed the UUID, it worked.


While for the networking.service, I already did, it's because it can't find up device eth0 (that I never use). I think I'll disable it later or remove from the list of interfaces.

Thanks for your help, kernel is now up after 2.3 seconds, instead of 33s!!