[SOLVED] Re: Very slow boot in F27 after recent update

[Gianluca Cecchi]

On Fri, May 4, 2018 at 9:32 AM, Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt;
wrote:

>  I'm not subscribed to receive all the e-mails, so I answer to my
> own.
>  Hope not to scramble the thread order...
>
>  Thanks for answering.
>  Here the output requested.
>  Possibly before the update, chronyd and gssproxy (and I see also libvirtd)
>  went into a sort of parallel background not preventing display manager
>  login completion?
>  In fact in about 10-13 seconds (I have an SSD disk on the laptop) I had
>  the gdm login page....
>
>  [g.cecchi@ope46 ~]$ sudo systemd-analyze time
>  Startup finished in 1.379s (kernel) + 1.965s (initrd) + 1min 49.668s
>  (userspace) = 1min 53.012s
>  [g.cecchi@ope46 ~]$
>
>  [g.cecchi@ope46 ~]$ sudo systemd-analyze blame
>      1min 30.134s gssproxy.service
>      1min 15.937s chronyd.service
>           30.043s NetworkManager-wait-online.service
>           17.876s libvirtd.service
>            1.102s dracut-initqueue.service
>            1.015s lvm2-monitor.service
>
>  
It seems latest updates have solved my problems, both if network cable is
connected and without it.

My network card on laptop is this one from "lspci -v":

 05:00.0 Ethernet controller: Qualcomm Atheros AR8151 v2.0 Gigabit Ethernet
(rev c0)
        Subsystem: ASUSTeK Computer Inc. Device 1851
        Flags: bus master, fast devsel, latency 0, IRQ 35
        Memory at dc800000 (64-bit, non-prefetchable) [size=256K]
        I/O ports at 9000 [size=128]
        Capabilities: [40] Power Management version 3
        Capabilities: [48] MSI: Enable+ Count=1/1 Maskable- 64bit+
        Capabilities: [58] Express Endpoint, MSI 00
        Capabilities: [6c] Vital Product Data
        Capabilities: [100] Advanced Error Reporting
        Capabilities: [180] Device Serial Number ff-bd-21-9b-14-da-e9-ff
        Kernel driver in use: atl1c
        Kernel modules: atl1c

It seems combination of latest kernel (4.16.6-202.fc27.x86_64) and latest
packages I have again my old behavior.

With network cable connected to an automatic dhcp connection:

[g.cecchi@ope46 ~]$ sudo systemd-analyze time
Startup finished in 1.386s (kernel) + 1.328s (initrd) + 8.891s (userspace)
= 11.606s
[g.cecchi@ope46 ~]$

[g.cecchi@ope46 ~]$ sudo systemd-analyze blame
          4.568s ntpdate.service
          2.145s NetworkManager-wait-online.service
          1.168s cups.service
          1.158s plymouth-quit-wait.service
          1.019s abrtd.service
           819ms lvm2-monitor.service
           716ms systemd-udev-settle.service
           712ms libvirtd.service
. . .


Without network cable connected (and without wirelsss autoconnect)
[g.cecchi@ope46 ~]$ sudo systemd-analyze time
Startup finished in 1.389s (kernel) + 2.002s (initrd) + 8.702s (userspace)
= 12.094s
[g.cecchi@ope46 ~]$

and

[g.cecchi@ope46 ~]$ sudo systemd-analyze blame
          6.337s NetworkManager-wait-online.service
          1.189s cups.service
          1.181s dracut-initqueue.service
          1.021s abrtd.service
           971ms lvm2-monitor.service
           886ms plymouth-quit-wait.service
           865ms systemd-udev-settle.service
           682ms firewalld.service
           641ms libvirtd.service
. . .


dnf summary for latest updates has been:

Installed:
  kernel.x86_64 4.16.6-202.fc27                       kernel-core.x86_64
4.16.6-202.fc27
  kernel-devel.x86_64 4.16.6-202.fc27                 kernel-modules.x86_64
4.16.6-202.fc27
  kernel-modules-extra.x86_64 4.16.6-202.fc27
 python2-ntlm-auth.noarch 1.1.0-1.fc27

Upgraded:
  ansible.noarch 2.5.2-1.fc27

  cinnamon-themes.noarch 1:1.6.2-1.fc27

  copy-jdk-configs.noarch 3.3-9.fc27

  httpd.x86_64 2.4.33-5.fc27

  httpd-filesystem.noarch 2.4.33-5.fc27

  httpd-tools.x86_64 2.4.33-5.fc27

  hwdata.noarch 0.312-1.fc27

  kdiagram.x86_64 2.6.1-2.fc27

  kernel-headers.x86_64 4.16.6-202.fc27

  libpsl.x86_64 0.18.0-2.fc27

  mate-panel.x86_64 1.20.1-5.fc27

  mate-panel-libs.x86_64 1.20.1-5.fc27

  mint-themes.noarch 1:1.6.2-1.fc27

  mint-themes-gtk3.noarch 1:1.6.2-1.fc27

  mint-x-icons.noarch 1.4.7-1.fc27

  mint-y-icons.noarch 1.1.4-1.fc27

  mint-y-theme.noarch 1:1.6.2-1.fc27

  mod_ssl.x86_64 1:2.4.33-5.fc27

  opencv.x86_64 3.2.0-15.fc27

  opencv-contrib.x86_64 3.2.0-15.fc27

  opencv-core.x86_64 3.2.0-15.fc27

  powerdevil.x86_64 5.12.4-2.fc27
  publicsuffix-list-dafsa.noarch 20180419-1.fc27

  python2-opencv.x86_64 3.2.0-15.fc27

  python2-requests_ntlm.noarch 1.1.0-1.fc27

  python2-winrm.noarch 0.3.0-1.fc27

  selinux-policy.noarch 3.13.1-283.34.fc27

  selinux-policy-devel.noarch 3.13.1-283.34.fc27

  selinux-policy-targeted.noarch 3.13.1-283.34.fc27

  spirv-tools-libs.i686 2018.3.0-0.1.20180407.git26a698c.fc27

  spirv-tools-libs.x86_64 2018.3.0-0.1.20180407.git26a698c.fc27

  vim-common.x86_64 2:8.0.1788-1.fc27

  vim-enhanced.x86_64 2:8.0.1788-1.fc27

  vim-filesystem.noarch 2:8.0.1788-1.fc27

  vim-minimal.x86_64 2:8.0.1788-1.fc27

  vulkan.i686 1.1.73.0-1.fc27

  vulkan.x86_64 1.1.73.0-1.fc27

  vulkan-filesystem.noarch 1.1.73.0-1.fc27


Cheers,
Gianluca

[Gianluca Cecchi]

On Tue, May 8, 2018 at 8:42 AM, Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt;
wrote:

[snip]


>  Possibly before the update, chronyd and gssproxy (and I see also
> libvirtd)
> > went into a sort of parallel background not preventing display manager
> > login completion?
> > In fact in about 10-13 seconds (I have an SSD disk on the laptop) I had
> > the gdm login page....
> >
> > [g.cecchi@ope46 ~]$ sudo systemd-analyze time
> > Startup finished in 1.379s (kernel) + 1.965s (initrd) + 1min 49.668s
> > (userspace) = 1min 53.012s
> > [g.cecchi@ope46 ~]$
> >
> > [g.cecchi@ope46 ~]$ sudo systemd-analyze blame
> >     1min 30.134s gssproxy.service
> >     1min 15.937s chronyd.service
> >          30.043s NetworkManager-wait-online.service
> >          17.876s libvirtd.service
> >           1.102s dracut-initqueue.service
> >           1.015s lvm2-monitor.service
> >
> >
>  It seems latest updates have solved my problems, both if network cable is
>  connected and without it.
>  
I updated the system in the mean time several weeks ago to F28 and with
yesterday update the problem arose again.
Same long elapsed times for gssproxy and chronyd and sendmail service
starting and stopping several times.
Booting with the previous kernel lets the problem disappear, apparently
Good last kernel:  4.16.16-300.fc28.x86_64
Bad new kernel: 4.17.2-200.fc28.x86_64

$ sudo dnf history list | head -6
ID     | Command line             | Date and time    | Action(s)      |
Altered
-------------------------------------------------------------------------------
   353 | update                   | 2018-06-25 09:00 | E, I, U        |
209 EE
   352 | update                   | 2018-06-20 20:52 | E, I, O, U     |
261 EE
   351 | update                   | 2018-06-16 13:32 | E, I, U        |
 89 EE
   350 | update                   | 2018-06-14 09:23 | Update         |
 29
$

Also the shutdown phase yesterday was very slow.
Is there a command to analyse shutdown steps, as you make with boot
and  "systemd-analyze
blame" command?
Anyone else?
Gianluca