Bug 15230 - long boot time - mandriva-everytime.service
Summary: long boot time - mandriva-everytime.service
Status: NEW
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: x86_64 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Colin Guthrie
QA Contact:
URL: http://www.mageialinux-online.org/for...
Whiteboard: MGA5TOO MGA6TOO
Keywords:
Depends on: 25075
Blocks:
  Show dependency treegraph
 
Reported: 2015-02-07 22:03 CET by Stéphane Couturier
Modified: 2019-07-09 12:10 CEST (History)
10 users (show)

See Also:
Source RPM:
CVE:
Status comment:


Attachments
boot screenshot (185.02 KB, image/jpeg)
2015-04-12 22:23 CEST, Stéphane Couturier
Details
quick fix to avoid the splash-hide in plymouth (456 bytes, patch)
2016-09-29 15:37 CEST, Giuseppe Ghibò
Details | Diff

Description Stéphane Couturier 2015-02-07 22:03:45 CET
Description of problem:
Hi, 
As per this forum post, our boot time to start Mageia 5 is abnormal.

Booting Mageia 4 use to take about 20s with an SSD.
Now using Mageia 5, the boot time exceeds 1min30s.

Looking at the log generated with the command systemd-analyze blame, we see that the mandriva-everytime.service takes more than a minute to start.

    1min 39.156s mandriva-everytime.service
          3.068s plymouth-quit-wait.service
          2.129s shorewall6.service
          1.250s mysqld.service
          1.244s shorewall.service
          1.181s network.service
           883ms systemd-udev-settle.service
           565ms systemd-journald.service
           286ms httpd.service
           ...
Note: in the MCC / System / manage services: the tool tip of this services indicates: "Reconfigure the system on administrator request"

Thank you,

Reproducible: 

Steps to Reproduce:
David Walser 2015-02-08 02:23:37 CET

Assignee: bugsquad => mageia

Sylvain Goumas 2015-02-08 07:34:36 CET

CC: (none) => timidou0409

Comment 1 magnux77 2015-03-04 14:30:18 CET
I have the same problem : Boot is far more longer in M5B3 than in M4 and the guilty is 'mandriva-everytime'.

bash-4.3$ systemd-analyze blame
         58.306s mandriva-everytime.service
          4.741s spamd.service
          4.621s network-up.service
          2.431s shorewall.service
          2.027s systemd-udev-settle.service
          1.310s shorewall6.service
           700ms systemd-fsck@dev-disk-by\x2duuid-a1c27c40\x2d1c58\x2d4694\x2d8c
           670ms systemd-fsck@dev-disk-by\x2duuid-6f6d9d6d\x2dd3c8\x2d47a3\x2d8b
           643ms systemd-fsck@dev-disk-by\x2duuid-9811a0f0\x2d2a90\x2d424e\x2d90
           616ms SaveHomes.mount
           440ms SaveSlash.mount

CC: (none) => magnux77

Comment 2 Ingvar Klimenkov 2015-03-16 06:50:43 CET
I revealed a small bug in Mageia 5 betta3. Incorrectly works when loading file:///usr/lib/systemd/system/mandriva-everytime .service system.
If to add the new device (I added the Wi-Fi network interface card) to a computer, when loading mandriva-everytime.service system gives out the black screen, but then the ravine are restored and loading is restored. At repeated loading everything is normal.
As this service, at start in the terminal works, assumed that he clashes with other service during loading of system.
  I changed in p. 6 of file:///usr/lib/systemd/system/mandriva-everytime .service:

After=mandriva-save-dmesg.service

Everything improved. The black screen isn't present. The truth and a conclusion of dens, during loading, the same isn't present.

CC: (none) => ingvar45

Comment 3 Ingvar Klimenkov 2015-03-16 09:18:58 CET
I note that boot time decreased

bash-4.3$ systemd-analyze blame
    2min 11.678s sendmail.service
    1min 10.703s sm-client.service
         53.542s mandriva-everytime.service
         15.577s systemd-remount-fs.service
         15.326s systemd-udev-settle.service
         10.094s magos-postinit.service
          8.169s webmin.service
          5.699s NetworkManager-wait-online.service
          5.283s atieventsd.service
          4.911s httpd.service
          4.521s network.service
          3.703s sensord.service
          3.691s NetworkManager.service
          3.626s adb.service
          2.674s iptables.service
          2.663s cups.service
          2.539s ModemManager.service
          2.137s drakxtools_http.service
          1.955s squid.service
          1.882s resolvconf.service
          1.847s plymouth-quit-wait.service
          1.707s partmon.service
          1.656s autofs.service
          1.384s lm_sensors.service
          1.381s upower.service
          1.356s polkit.service
          1.353s smb.service
          1.256s rsyslog.service
          1.204s nfs-server.service
          1.113s saslauthd.service
          1.042s preload.service
           954ms systemd-logind.service
           909ms proftpd.service
           904ms systemd-resolved.service
           852ms proc-fs-nfsd.mount
           841ms nfs-lock.service
           817ms nslcd.service
           806ms openl2tp.service
lines 1-38
Comment 4 magnux77 2015-04-06 12:23:02 CEST
I read somewhere that a possible cause was dkms on multiple old kernels. Since i am on Mageia5 i had a lot of unused kernels. I have deleted them an yet the duration of the boot, and specially mandriva-everytime is 1 minute longer.
I also noticed that desactivation of mandriva-everytime.service in MCC (perhaps stupid) has no effect.
bash-4.3$ systemd-analyze blame
    1min 49.099s mandriva-everytime.service
          4.621s network-up.service
          4.314s spamd.service
          2.251s shorewall6.service
          1.928s systemd-udev-settle.service
          1.235s shorewall.service
           817ms systemd-fsck@dev-disk-by\x2duuid-9811a0f0\x2d2a90\x2d424e\x2d90
           593ms SaveHomes.mount
           561ms systemd-fsck@dev-disk-by\x2duuid-54a08537\x2da117\x2d471b\x2db7
           417ms systemd-fsck@dev-disk-by\x2duuid-a1c27c40\x2d1c58\x2d4694\x2d8c
           377ms SaveSlash.mount
           334ms network.service
           250ms systemd-journal-flush.service
           226ms systemd-fsck@dev-disk-by\x2duuid-6f6d9d6d\x2dd3c8\x2d47a3\x2d8b
           207ms cups.service
           185ms VirtualBox\x20VMs.mount
           143ms freshclam.service
           139ms home.mount
           112ms resolvconf.service
           109ms acpid.service
            86ms systemd-logind.service
Comment 5 magnux77 2015-04-09 16:51:11 CEST
So what ?
Comment 6 Rémi Verschelde 2015-04-10 19:17:25 CEST
Any idea about how to debug this further Colin?

CC: (none) => rverschelde

Comment 7 Manuel Hiebel 2015-04-12 11:51:56 CEST
I disable it, and didn't found any issue yet,

if you check the file /etc/init.d/mandrake_everytime (which is launched from the service) you will see some stuff, not needed sometimes (in my case, every times)
Comment 8 Sander Lepik 2015-04-12 12:03:10 CEST
I've seen similar delays and for that I added this line:

HARDDRAKE_ONBOOT=no

into

/etc/sysconfig/system

CC: (none) => mageia

Comment 9 magnux77 2015-04-12 17:01:24 CEST
@ Manuel Hiebel : 1° As I wrote before, desactivating the mandriva-everytime.service has no effect, it does run at next boot anyway.
2° The stuff in /etc/init.d/mandrake_everytime is not clear for me (TOTAL BLACKNESS !) to do any moficiation.

@ Sander Lepik : HARDDRAKE_ONBOOT=no in /etc/sysconfig/system has no effect at reboot
cat /etc/sysconfig/system
CLASS=beginner
LIBSAFE=no
META_CLASS=download
SECURITY=3
HARDDRAKE_ONBOOT=no

bash-4.3$ systemd-analyze blame
    1min 46.613s mandriva-everytime.service
          4.620s network-up.service
          4.453s spamd.service
          2.681s shorewall.service
          2.051s systemd-udev-settle.service
          1.238s shorewall6.service
          1.030s systemd-fsck@dev-disk-by\x2duuid-54a08537\x2da117\x2d471b\x2db7
          1.030s systemd-fsck@dev-disk-by\x2duuid-6f6d9d6d\x2dd3c8\x2d47a3\x2d8b
           717ms systemd-fsck@dev-disk-by\x2duuid-a1c27c40\x2d1c58\x2d4694\x2d8c

But in /etc/init.d/mandrake_everytime, I saw a test on HARDDRAKE_ONBOOT=no of /etc/sysconfig/system proposed by Sander. I saw also a test on DKMS_ONBOOT=no of that same file. So, I added DKMS_ONBOOT=no into /etc/sysconfig/system and it looks a little bit better. mandriva-everytime.service decreases from 1min46 to 3s. !
[root@AMD664 ~]# systemd-analyze blame
          4.622s network-up.service
          4.036s spamd.service
          3.182s mandriva-everytime.service
          2.480s shorewall.service
          1.922s systemd-udev-settle.service
          1.208s shorewall6.service
           928ms systemd-fsck@dev-disk-by\x2duuid-54a08537\x2da117\x2d471b\x2db7
The problem looks solved but some residual questions :
 - what are the consequences of dkms_onboot=no ?
 - am i the only with that problem ? 
 - should the bug be considerd as resolved ?
Comment 10 Stéphane Couturier 2015-04-12 22:22:25 CEST
Hi, thanks for the workaround! 
I have now a normal boot time after adding the 2 lines:

HARDDRAKE_ONBOOT=no   
DKMS_ONBOOT=no

in

/etc/sysconfig/system

I am going to attach a screen shot of the boot text as seen on the screen.
It looks like the system is trying to recompile everytime broadcom and nvidia drivers that are "Already installed on this kernel".
I understand better why the DKMS option helps avoiding this. 
But it is a normal behaviour?
Comment 11 Stéphane Couturier 2015-04-12 22:23:57 CEST
Created attachment 6249 [details]
boot screenshot

by the way this, was also causing some overheating...
Comment 12 magnux77 2015-04-13 11:21:23 CEST
Because of no effects, I removed the "HARDDRAKE_ONBOOT=no.

I only have DKMS_ONBOOT=no.

And i would like to have a response for those questions :
 - what are the consequences of dkms_onboot=no ?
 - am i the only with that problem ? 
 - should the bug be considerd as resolved ?
Comment 13 Rémi Verschelde 2015-04-13 11:25:58 CEST
(In reply to epilip from comment #12)
> 
> And i would like to have a response for those questions :
>  - what are the consequences of dkms_onboot=no ?
I guess DKMS_ONBOOT is meant to make sure that your kernel modules (like dkms-nvidia-current for example) are properly built for the kernel you are running. I'm not 100% sure, but I guess the reason for this feature is to ensure that you can boot _any_ kernel without having issue, since the kernel modules will be rebuilt on boot if they are not already built.

But the tests should be really quick, on my system I have two dkms modules, and it does not delay the boot too much. I wonder if you're not having a dkms module that fails rebuild at each boot, and that's why your boot takes so long. Could you check?

>  - am i the only with that problem ? 
>  - should the bug be considerd as resolved ?

The second point depends on the first. If you are the only one with the problem, the bug can probably be closed, but if it's more general it calls for a better fix.
Comment 14 magnux77 2015-04-13 11:45:41 CEST
Rémi Verschelle "Could you check?"
Yes, for sure. How ? (I'm a user not a developper...)
Comment 15 magnux77 2015-04-13 12:13:32 CEST
As i don't know what to check, i tried to search where to check. But too many locations. Perhaps the problem is the last file ?
root@AMD664 log]# find / -name "*dkms*"
/var/cache/urpmi/rpms/dkms-nvidia-current-346.35-5.mga5.nonfree.x86_64.rpm
/var/lib/dkms-binary
/var/lib/dkms-binary/virtualbox/4.3.20-2.mga5/3.18.3-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-2.mga5/3.19.0-desktop-0.rc7.3.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-2.mga5/3.19.0-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-2.mga5/3.19.0-desktop-0.rc7.2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.26-1.mga5/3.19.2-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.26-1.mga5/3.19.2-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.26-1.mga5/3.19.3-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.1-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.3-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.1-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.0-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.1-desktop-4.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.0-desktop-0.rc7.1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.1-desktop-3.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.0-desktop-0.rc7.2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.17.4-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.0-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.20-1.mga5/3.18.0-desktop-0.rc6.1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.24-1.mga5/3.19.0-desktop-6.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.24-1.mga5/3.19.1-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.24-1.mga5/3.19.0-desktop-5.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.24-1.mga5/3.19.1-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.22-1.mga5/3.19.0-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.22-2.mga5/3.19.0-desktop-3.mga5/dkms.conf
/var/lib/dkms-binary/virtualbox/4.3.22-2.mga5/3.19.0-desktop-4.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.35-5.mga5.nonfree/3.19.0-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.35-5.mga5.nonfree/3.19.0-desktop-0.rc7.3.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.35-5.mga5.nonfree/3.19.0-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.35-5.mga5.nonfree/3.19.0-desktop-3.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.35-5.mga5.nonfree/3.19.0-desktop-0.rc7.2.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.59-1.mga5.nonfree/3.19.3-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.2-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.0-desktop-6.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.2-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.1-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.0-desktop-5.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.0-desktop-4.mga5/dkms.conf
/var/lib/dkms-binary/nvidia-current/346.47-1.mga5.nonfree/3.19.1-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-2.mga5/3.18.3-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-2.mga5/3.19.0-desktop-0.rc7.3.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-2.mga5/3.19.0-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-2.mga5/3.19.0-desktop-0.rc7.2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.26-1.mga5/3.19.2-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.26-1.mga5/3.19.2-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.26-1.mga5/3.19.3-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.1-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.3-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.1-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.0-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.1-desktop-4.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.0-desktop-0.rc7.1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.1-desktop-3.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.0-desktop-0.rc7.2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.17.4-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.0-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.20-1.mga5/3.18.0-desktop-0.rc6.1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.24-1.mga5/3.19.0-desktop-6.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.24-1.mga5/3.19.1-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.24-1.mga5/3.19.0-desktop-5.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.24-1.mga5/3.19.1-desktop-1.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.22-1.mga5/3.19.0-desktop-2.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.22-2.mga5/3.19.0-desktop-3.mga5/dkms.conf
/var/lib/dkms-binary/vboxadditions/4.3.22-2.mga5/3.19.0-desktop-4.mga5/dkms.conf
/var/lib/dkms
/var/lib/dkms/dkms_dbversion
/var/lib/dkms/virtualbox/4.3.26-1.mga5/build/dkms.conf
/var/lib/dkms/virtualbox/4.3.26-1.mga5/build/do_dkms
/var/lib/dkms/nvidia-current/346.59-1.mga5.nonfree/build/dkms.conf
/var/lib/dkms/vboxadditions/4.3.26-1.mga5/build/dkms.conf
/var/lib/dkms/vboxadditions/4.3.26-1.mga5/build/vboxadditions-dkms-mkrpm.spec
/etc/kernel/postinst.d/dkms
/etc/kernel/prerm.d/dkms
/etc/bash_completion.d/dkms
/etc/depmod.d/dkms.conf
/etc/dkms
/etc/dkms/template-dkms-mkrpm.spec
/usr/share/ldetect-lst/dkms-modules.description
/usr/share/ldetect-lst/dkms-modules.alias
/usr/share/man/man8/dkms.8.xz
/usr/share/doc/dkms-nvidia-current
/usr/share/doc/dkms
/usr/share/doc/dkms/template-dkms-mkrpm.spec
/usr/sbin/dkms_autoinstaller
/usr/sbin/dkms_autoload
/usr/sbin/dkms
/usr/sbin/dkms_mkkerneldoth
/usr/src/nvidia-current-346.59-1.mga5.nonfree/dkms.conf
/usr/src/virtualbox-4.3.26-1.mga5/dkms.conf
/usr/src/virtualbox-4.3.26-1.mga5/do_dkms
/usr/src/vboxadditions-4.3.26-1.mga5/dkms.conf
/usr/src/vboxadditions-4.3.26-1.mga5/vboxadditions-dkms-mkrpm.spec
/usr/lib/modules/3.18.1-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.18.1-desktop-1.mga5/dkms
/usr/lib/modules/3.19.2-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.18.3-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.18.3-desktop-2.mga5/dkms
/usr/lib/modules/3.19.0-desktop-6.mga5/dkms-binary
/usr/lib/modules/3.18.3-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.18.1-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.18.0-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.18.0-desktop-2.mga5/dkms
/usr/lib/modules/3.19.2-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.19.2-desktop-2.mga5/dkms
/usr/lib/modules/3.19.1-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.19.1-desktop-2.mga5/dkms
/usr/lib/modules/3.19.0-desktop-2.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-5.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-5.mga5/dkms
/usr/lib/modules/3.19.3-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.19.3-desktop-1.mga5/dkms
/usr/lib/modules/3.19.0-desktop-0.rc7.3.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-0.rc7.3.mga5/dkms
/usr/lib/modules/3.19.0-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-1.mga5/dkms
/usr/lib/modules/3.18.1-desktop-4.mga5/dkms-binary
/usr/lib/modules/3.18.1-desktop-4.mga5/dkms
/usr/lib/modules/3.18.0-desktop-0.rc7.1.mga5/dkms-binary
/usr/lib/modules/3.18.0-desktop-0.rc7.1.mga5/dkms
/usr/lib/modules/3.19.0-desktop-3.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-3.mga5/dkms
/usr/lib/modules/3.18.1-desktop-3.mga5/dkms-binary
/usr/lib/modules/3.18.1-desktop-3.mga5/dkms
/usr/lib/modules/3.18.0-desktop-0.rc7.2.mga5/dkms-binary
/usr/lib/modules/3.18.0-desktop-0.rc7.2.mga5/dkms
/usr/lib/modules/3.19.0-desktop-4.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-4.mga5/dkms
/usr/lib/modules/3.17.4-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.17.4-desktop-1.mga5/dkms
/usr/lib/modules/3.19.1-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.19.1-desktop-1.mga5/dkms
/usr/lib/modules/3.18.0-desktop-1.mga5/dkms-binary
/usr/lib/modules/3.18.0-desktop-1.mga5/dkms
/usr/lib/modules/3.19.0-desktop-0.rc7.2.mga5/dkms-binary
/usr/lib/modules/3.19.0-desktop-0.rc7.2.mga5/dkms
/usr/lib/modules/3.18.0-desktop-0.rc6.1.mga5/dkms-binary
/tmp/dkms_rpm_safe_upgrade_lock.2905.NzM8F7
[root@AMD664 log]#
Comment 16 Rémi Verschelde 2015-04-13 12:22:01 CEST
(In reply to epilip from comment #14)
> Rémi Verschelle "Could you check?"
> Yes, for sure. How ? (I'm a user not a developper...)

By removing the DKMS_ONBOOT=no workaround, and then watch the boot logs on boot to see if it says: "blablabla: Already installed on this kernel." or something like:
"blablabla: ...............
...............................
...........................
Module compilation failed"
Comment 17 magnux77 2015-04-13 18:14:01 CEST
(In reply to Rémi Verschelde from comment #16)
> > Rémi Verschelle "Could you check?"
> > Yes, for sure. How ? (I'm a user not a developper...)
> 
> By removing the DKMS_ONBOOT=no workaround, and then watch the boot logs on
> boot to see if it says: "blablabla: Already installed on this kernel." or
> something like:
> "blablabla: ...............
> ...............................
> ...........................
> Module compilation failed"
1° No need to test. While i was suffering those too long boots, I have so often seen those messages "Already installed.." for both nvidia and virtual box but none "Module compilation failed".

2° I made the test with workaround "DKMS_ONBOOT=no" removed and rebbot.
The messages "Already intalled.." appear and no "Module compilation failed".
But the boot is really faster cause i have deleted the most obsolete of /var/lib/dkms-binary conf files in the list in comment15.
Comment 18 magnux77 2015-04-14 09:22:49 CEST
As said by TuxMad here (http://www.mageialinux-online.org/forum/topic-20039+bug-15230-mga5b2-long-boot-time-mandriva-everytime-service.php), the fresh install of Mageia5B3 doesn't encounter the problem, but the migration from 4 to 5 will cause this long boot due to lots of useless dkms. 

Few users are trying this migration today, so almost nobody complain.
But, when Mageia 5 will be released official, there will be lots of users who will experiment the problem when migrating from Mageia4 . 

So I think all these useless dkms (and other stuff I don't know) have to be deleted or another solution has to be found for the migration from 4 to 5 to be successfull.
Comment 19 magnux77 2015-04-14 09:50:03 CEST
Precision : All these useless and obsolete files still stay in the system while all obsolete kernels where previously deleted by CCM.
Comment 20 Rémi Verschelde 2015-04-14 10:01:28 CEST
Thomas, what is your opinion on comments 18 and 19? What is the current best way for an end user to get rid of old built modules for already-removed kernels? And why do those delay the boot time so much?

CC: (none) => tmb

Comment 21 Thomas Backlund 2015-04-14 10:15:28 CEST
Hm, the prebuilt modules should be removed when you remove the kernel as they have strict versioned deps.
Comment 22 Thomas Backlund 2015-04-14 10:18:16 CEST
and dkms on boot should only build the ones matching the booting kernel
Comment 23 Rémi Verschelde 2015-04-14 10:22:39 CEST
epilip: Are you sure that you removed old kernels, or did you just drop them from the grub bootloader?
Comment 24 Rémi Verschelde 2015-04-14 10:26:31 CEST
I confirm that uninstalling a kernel also uninstalls the corresponding prebuilt modules, I just removed kernel-3.19.0-desktop-6.mga5 and that removed nvidia-current-kernel-3.19.0-desktop-6.mga5.
Comment 25 Thomas Backlund 2015-04-14 10:32:58 CEST
I even have a secondary cleanup in kernel %postun that should catch anything that dkms leaves behind:


rm -rf /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel} >/dev/null
if [ -d /var/lib/dkms ]; then
    rm -f /var/lib/dkms/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
    rm -rf /var/lib/dkms/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
    rm -f /var/lib/dkms-binary/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
    rm -rf /var/lib/dkms-binary/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
fi

So I dont see how the files could be left behind (unless you still have all those kernels installed)
Comment 26 magnux77 2015-04-14 11:03:14 CEST
(In reply to Rémi Verschelde from comment #23)
> epilip: Are you sure that you removed old kernels, or did you just drop them
> from the grub bootloader?

I remove obsolete kernels by CCM / Démarrage : COnfigure. One by one, then finish. 

Strange but consistent with Comment 24 & 25. After reboot or not, obsoletes kernels are still remaining.
[root@AMD664 ~]# ll /boot
total 434964
-rw-r--r-- 1 root root   167785 nov.   4 01:32 config-3.17.2-desktop-3.mga5
-rw-r--r-- 1 root root   167764 nov.  21 21:17 config-3.17.4-desktop-1.mga5
-rw-r--r-- 1 root root   170060 nov.  30 00:56 config-3.18.0-desktop-0.rc6.1.mga5
-rw-r--r-- 1 root root   170060 déc.   1 10:43 config-3.18.0-desktop-0.rc7.1.mga5
-rw-r--r-- 1 root root   170060 déc.   5 20:43 config-3.18.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root   170054 déc.   8 09:19 config-3.18.0-desktop-1.mga5
-rw-r--r-- 1 root root   170054 déc.  14 13:58 config-3.18.0-desktop-2.mga5
-rw-r--r-- 1 root root   170054 déc.  15 22:46 config-3.18.1-desktop-1.mga5
-rw-r--r-- 1 root root   170054 déc.  21 21:37 config-3.18.1-desktop-2.mga5
-rw-r--r-- 1 root root   170114 déc.  30 23:11 config-3.18.1-desktop-3.mga5
-rw-r--r-- 1 root root   170122 janv.  7 16:39 config-3.18.1-desktop-4.mga5
-rw-r--r-- 1 root root   170122 janv. 16 20:14 config-3.18.3-desktop-1.mga5
-rw-r--r-- 1 root root   170122 janv. 21 16:42 config-3.18.3-desktop-2.mga5
-rw-r--r-- 1 root root   171240 févr.  3 00:24 config-3.19.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root   171240 févr.  4 02:31 config-3.19.0-desktop-0.rc7.3.mga5
-rw-r--r-- 1 root root   171234 févr.  9 07:48 config-3.19.0-desktop-1.mga5
-rw-r--r-- 1 root root   171234 févr. 23 00:20 config-3.19.0-desktop-2.mga5
-rw-r--r-- 1 root root   171245 févr. 24 21:51 config-3.19.0-desktop-3.mga5
-rw-r--r-- 1 root root   171245 févr. 26 21:31 config-3.19.0-desktop-4.mga5
-rw-r--r-- 1 root root   171245 mars   2 17:09 config-3.19.0-desktop-5.mga5
-rw-r--r-- 1 root root   171245 mars   7 12:10 config-3.19.0-desktop-6.mga5
-rw-r--r-- 1 root root   171245 mars  10 23:42 config-3.19.1-desktop-1.mga5
-rw-r--r-- 1 root root   171245 mars  14 23:29 config-3.19.1-desktop-2.mga5
-rw-r--r-- 1 root root   171245 mars  18 22:02 config-3.19.2-desktop-1.mga5
-rw-r--r-- 1 root root   171245 mars  24 19:18 config-3.19.2-desktop-2.mga5
-rw-r--r-- 1 root root   171245 avril  3 02:28 config-3.19.3-desktop-1.mga5
drwxr-xr-x 2 root root     4096 avril  6 22:30 dracut/
-rwxr-xr-x 1 root root   562176 mars   5 17:53 gfxmenu*
drwxr-xr-x 2 root root     4096 nov.  28 09:05 grub/
drwxr-xr-x 6 root root     4096 avril 14 10:57 grub2/
-rw------- 1 root root 11427597 nov.  28 09:06 initrd-3.17.2-desktop-3.mga5.img
-rw------- 1 root root  8644935 nov.  28 08:49 initrd-3.17.4-desktop-1.mga5.img
-rw------- 1 root root  8652148 nov.  30 08:31 initrd-3.18.0-desktop-0.rc6.1.mga5.img
-rw------- 1 root root  8695267 déc.   1 14:54 initrd-3.18.0-desktop-0.rc7.1.mga5.img
-rw------- 1 root root  8654908 déc.   6 10:10 initrd-3.18.0-desktop-0.rc7.2.mga5.img
-rw------- 1 root root  8695977 déc.   8 13:39 initrd-3.18.0-desktop-1.mga5.img
-rw------- 1 root root  8654576 déc.  15 09:18 initrd-3.18.0-desktop-2.mga5.img
-rw------- 1 root root  8695752 déc.  16 08:44 initrd-3.18.1-desktop-1.mga5.img
-rw------- 1 root root  9237697 déc.  22 10:10 initrd-3.18.1-desktop-2.mga5.img
-rw------- 1 root root  9238007 janv.  5 17:44 initrd-3.18.1-desktop-3.mga5.img
-rw------- 1 root root  9238957 janv.  8 10:23 initrd-3.18.1-desktop-4.mga5.img
-rw------- 1 root root  9200814 janv. 19 14:45 initrd-3.18.3-desktop-1.mga5.img
-rw------- 1 root root  9200077 janv. 22 09:17 initrd-3.18.3-desktop-2.mga5.img
-rw------- 1 root root  9257988 févr.  3 14:50 initrd-3.19.0-desktop-0.rc7.2.mga5.img
-rw------- 1 root root  9272857 févr.  4 16:00 initrd-3.19.0-desktop-0.rc7.3.mga5.img
-rw------- 1 root root  9232142 févr.  9 15:49 initrd-3.19.0-desktop-1.mga5.img
-rw------- 1 root root  9237159 févr. 23 11:13 initrd-3.19.0-desktop-2.mga5.img
-rw------- 1 root root  9277688 févr. 25 09:31 initrd-3.19.0-desktop-3.mga5.img
-rw------- 1 root root  9237718 févr. 28 17:14 initrd-3.19.0-desktop-4.mga5.img
-rw------- 1 root root  9237053 mars   3 10:05 initrd-3.19.0-desktop-5.mga5.img
-rw------- 1 root root  9237084 mars   8 09:38 initrd-3.19.0-desktop-6.mga5.img
-rw------- 1 root root  9236523 mars  11 08:45 initrd-3.19.1-desktop-1.mga5.img
-rw------- 1 root root  9236979 mars  15 16:26 initrd-3.19.1-desktop-2.mga5.img
-rw------- 1 root root  9235803 mars  19 08:49 initrd-3.19.2-desktop-1.mga5.img
-rw------- 1 root root  9237119 mars  26 14:16 initrd-3.19.2-desktop-2.mga5.img
-rw------- 1 root root  9235908 avril  4 09:56 initrd-3.19.3-desktop-1.mga5.img
lrwxrwxrwx 1 root root       32 avril  4 09:56 initrd-desktop.img -> initrd-3.19.3-desktop-1.mga5.img
lrwxrwxrwx 1 root root       32 avril  4 09:56 initrd.img -> initrd-3.19.3-desktop-1.mga5.img
-rw-r--r-- 1 root root   241192 nov.   4 01:32 symvers-3.17.2-desktop-3.mga5.xz
-rw-r--r-- 1 root root   241436 nov.  21 21:17 symvers-3.17.4-desktop-1.mga5.xz
-rw-r--r-- 1 root root   242384 nov.  30 00:56 symvers-3.18.0-desktop-0.rc6.1.mga5.xz
-rw-r--r-- 1 root root   242384 déc.   1 10:43 symvers-3.18.0-desktop-0.rc7.1.mga5.xz
-rw-r--r-- 1 root root   242400 déc.   5 20:43 symvers-3.18.0-desktop-0.rc7.2.mga5.xz
-rw-r--r-- 1 root root   242400 déc.   8 09:19 symvers-3.18.0-desktop-1.mga5.xz
-rw-r--r-- 1 root root   242400 déc.  14 13:58 symvers-3.18.0-desktop-2.mga5.xz
-rw-r--r-- 1 root root   242344 déc.  15 22:47 symvers-3.18.1-desktop-1.mga5.xz
-rw-r--r-- 1 root root   242344 déc.  21 21:37 symvers-3.18.1-desktop-2.mga5.xz
-rw-r--r-- 1 root root   242344 déc.  30 23:11 symvers-3.18.1-desktop-3.mga5.xz
-rw-r--r-- 1 root root   242068 janv.  7 16:39 symvers-3.18.1-desktop-4.mga5.xz
-rw-r--r-- 1 root root   242080 janv. 16 20:14 symvers-3.18.3-desktop-1.mga5.xz
-rw-r--r-- 1 root root   242112 janv. 21 16:42 symvers-3.18.3-desktop-2.mga5.xz
-rw-r--r-- 1 root root   245372 févr.  3 00:24 symvers-3.19.0-desktop-0.rc7.2.mga5.xz
-rw-r--r-- 1 root root   245372 févr.  4 02:31 symvers-3.19.0-desktop-0.rc7.3.mga5.xz
-rw-r--r-- 1 root root   245356 févr.  9 07:48 symvers-3.19.0-desktop-1.mga5.xz
-rw-r--r-- 1 root root   245356 févr. 23 00:20 symvers-3.19.0-desktop-2.mga5.xz
-rw-r--r-- 1 root root   245356 févr. 24 21:51 symvers-3.19.0-desktop-3.mga5.xz
-rw-r--r-- 1 root root   245356 févr. 26 21:31 symvers-3.19.0-desktop-4.mga5.xz
-rw-r--r-- 1 root root   245376 mars   2 17:09 symvers-3.19.0-desktop-5.mga5.xz
-rw-r--r-- 1 root root   245656 mars   7 12:10 symvers-3.19.0-desktop-6.mga5.xz
-rw-r--r-- 1 root root   245720 mars  10 23:42 symvers-3.19.1-desktop-1.mga5.xz
-rw-r--r-- 1 root root   245512 mars  14 23:29 symvers-3.19.1-desktop-2.mga5.xz
-rw-r--r-- 1 root root   245512 mars  18 22:02 symvers-3.19.2-desktop-1.mga5.xz
-rw-r--r-- 1 root root   245512 mars  24 19:18 symvers-3.19.2-desktop-2.mga5.xz
-rw-r--r-- 1 root root   245420 avril  3 02:28 symvers-3.19.3-desktop-1.mga5.xz
-rw-r--r-- 1 root root  3165755 nov.   4 01:32 System.map-3.17.2-desktop-3.mga5
-rw-r--r-- 1 root root  3167215 nov.  21 21:17 System.map-3.17.4-desktop-1.mga5
-rw-r--r-- 1 root root  3191008 nov.  30 00:56 System.map-3.18.0-desktop-0.rc6.1.mga5
-rw-r--r-- 1 root root  3191050 déc.   1 10:43 System.map-3.18.0-desktop-0.rc7.1.mga5
-rw-r--r-- 1 root root  3191201 déc.   5 20:43 System.map-3.18.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root  3191201 déc.   8 09:19 System.map-3.18.0-desktop-1.mga5
-rw-r--r-- 1 root root  3191226 déc.  14 13:58 System.map-3.18.0-desktop-2.mga5
-rw-r--r-- 1 root root  3191226 déc.  15 22:46 System.map-3.18.1-desktop-1.mga5
-rw-r--r-- 1 root root  3191297 déc.  21 21:37 System.map-3.18.1-desktop-2.mga5
-rw-r--r-- 1 root root  3191537 déc.  30 23:11 System.map-3.18.1-desktop-3.mga5
-rw-r--r-- 1 root root  3191857 janv.  7 16:39 System.map-3.18.1-desktop-4.mga5
-rw-r--r-- 1 root root  3191894 janv. 16 20:14 System.map-3.18.3-desktop-1.mga5
-rw-r--r-- 1 root root  3191673 janv. 21 16:42 System.map-3.18.3-desktop-2.mga5
-rw-r--r-- 1 root root  3222671 févr.  3 00:24 System.map-3.19.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root  3222628 févr.  4 02:31 System.map-3.19.0-desktop-0.rc7.3.mga5
-rw-r--r-- 1 root root  3222798 févr.  9 07:48 System.map-3.19.0-desktop-1.mga5
-rw-r--r-- 1 root root  3222798 févr. 23 00:20 System.map-3.19.0-desktop-2.mga5
-rw-r--r-- 1 root root  3222831 févr. 24 21:51 System.map-3.19.0-desktop-3.mga5
-rw-r--r-- 1 root root  3222831 févr. 26 21:31 System.map-3.19.0-desktop-4.mga5
-rw-r--r-- 1 root root  3222929 mars   2 17:09 System.map-3.19.0-desktop-5.mga5
-rw-r--r-- 1 root root  3222929 mars   7 12:10 System.map-3.19.0-desktop-6.mga5
-rw-r--r-- 1 root root  3222929 mars  10 23:42 System.map-3.19.1-desktop-1.mga5
-rw-r--r-- 1 root root  3222968 mars  14 23:29 System.map-3.19.1-desktop-2.mga5
-rw-r--r-- 1 root root  3222998 mars  18 22:02 System.map-3.19.2-desktop-1.mga5
-rw-r--r-- 1 root root  3223103 mars  24 19:18 System.map-3.19.2-desktop-2.mga5
-rw-r--r-- 1 root root  3223368 avril  3 02:28 System.map-3.19.3-desktop-1.mga5
lrwxrwxrwx 1 root root       29 avril  4 09:56 vmlinuz -> vmlinuz-3.19.3-desktop-1.mga5
-rw-r--r-- 1 root root  4254000 nov.   4 01:32 vmlinuz-3.17.2-desktop-3.mga5
-rw-r--r-- 1 root root  4257840 nov.  21 21:17 vmlinuz-3.17.4-desktop-1.mga5
-rw-r--r-- 1 root root  4274224 nov.  30 00:56 vmlinuz-3.18.0-desktop-0.rc6.1.mga5
-rw-r--r-- 1 root root  4287728 déc.   1 10:43 vmlinuz-3.18.0-desktop-0.rc7.1.mga5
-rw-r--r-- 1 root root  4278192 déc.   5 20:43 vmlinuz-3.18.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root  4283504 déc.   8 09:19 vmlinuz-3.18.0-desktop-1.mga5
-rw-r--r-- 1 root root  4284400 déc.  14 13:58 vmlinuz-3.18.0-desktop-2.mga5
-rw-r--r-- 1 root root  4288816 déc.  15 22:47 vmlinuz-3.18.1-desktop-1.mga5
-rw-r--r-- 1 root root  4284464 déc.  21 21:37 vmlinuz-3.18.1-desktop-2.mga5
-rw-r--r-- 1 root root  4285552 déc.  30 23:11 vmlinuz-3.18.1-desktop-3.mga5
-rw-r--r-- 1 root root  4284080 janv.  7 16:39 vmlinuz-3.18.1-desktop-4.mga5
-rw-r--r-- 1 root root  4286320 janv. 16 20:14 vmlinuz-3.18.3-desktop-1.mga5
-rw-r--r-- 1 root root  4289008 janv. 21 16:42 vmlinuz-3.18.3-desktop-2.mga5
-rw-r--r-- 1 root root  4343168 févr.  3 00:24 vmlinuz-3.19.0-desktop-0.rc7.2.mga5
-rw-r--r-- 1 root root  4343488 févr.  4 02:31 vmlinuz-3.19.0-desktop-0.rc7.3.mga5
-rw-r--r-- 1 root root  4341120 févr.  9 07:48 vmlinuz-3.19.0-desktop-1.mga5
-rw-r--r-- 1 root root  4341568 févr. 23 00:20 vmlinuz-3.19.0-desktop-2.mga5
-rw-r--r-- 1 root root  4339392 févr. 24 21:51 vmlinuz-3.19.0-desktop-3.mga5
-rw-r--r-- 1 root root  4342976 févr. 26 21:31 vmlinuz-3.19.0-desktop-4.mga5
-rw-r--r-- 1 root root  4340096 mars   2 17:09 vmlinuz-3.19.0-desktop-5.mga5
-rw-r--r-- 1 root root  4339200 mars   7 12:10 vmlinuz-3.19.0-desktop-6.mga5
-rw-r--r-- 1 root root  4340800 mars  10 23:42 vmlinuz-3.19.1-desktop-1.mga5
-rw-r--r-- 1 root root  4340928 mars  14 23:29 vmlinuz-3.19.1-desktop-2.mga5
-rw-r--r-- 1 root root  4340928 mars  18 22:02 vmlinuz-3.19.2-desktop-1.mga5
-rw-r--r-- 1 root root  4341376 mars  24 19:18 vmlinuz-3.19.2-desktop-2.mga5
-rw-r--r-- 1 root root  4342144 avril  3 02:28 vmlinuz-3.19.3-desktop-1.mga5
lrwxrwxrwx 1 root root       29 avril  4 09:56 vmlinuz-desktop -> vmlinuz-3.19.3-desktop-1.mga5
[root@AMD664 ~]#
Comment 27 Stéphane Couturier 2015-04-15 20:37:42 CEST
Hi thank you, I made good progress thanks to all this feedback, 
- note: I use Mga5 since beta 1.
- went to the MCC / Software Management (Install and Remove Software)  (rpmdrake)
- searched "kernel" packages
- noticed the long list of kernel-desktop-3.* , about 25 of them, starting from 3.12.18-1
- Removed them one by one. When unselecting one, it also proposed to remove the dependent packages, typically nvidia, broadcom, & virtualbox. So I accepted every time.
- Removing just one kernel was quite long at the beginning, but it got faster and faster while less and less were remaining.
- stopped to keep the last 3 versions.
- removed the temporary workaround line DKMS_ONBOOT=no from /etc/sysconfig/system
- rebooted and the boot time is now correct


There is still a long list of kernel-desktop-devel packages, but it seems they were not causing problems. I might remove them just for space.


Then regarding Mageia5 release:
- when upgrading from 4 to 5, are old kernels kept? if not, users should not have such issue right away
- but in time, while new kernels are released, could this problem appear again? so it could be worth keeping it opened.
Comment 28 Stéphane Couturier 2015-04-15 20:41:46 CEST
Note: the best way to search for the kernel packages is to type:
kernel-desktop-3
Comment 29 magnux77 2015-04-16 10:43:26 CEST
Conclusions
Without comments of any Mageia one, for those who will experiment long time boot and read this post, i resume here MY conclusions before leaving this bug :

- I misunderstood CCM Démarrage : it just update grub, i was believing it deletes kernels
- I tried it today, uninstalling kernel-desktop-3.19.3 also delete modules as said in comment 24 comment 25
- So, if you experiment long time boot perhaps the reason why is multiple useless dkms. The correct workaround consists of deleting obsolete kernels by CCM / Install-Uninstall.
- I don't know the origin of the bug, does anybody :  bug 10771 ?
Comment 30 Rémi Verschelde 2015-04-16 10:59:42 CEST
So the real bug seems to be that having many dkms modules installed delays the DKMS_ONBOOT function too much. That should be investigated IMO, because on the other hand forcing users to remove older kernels is not a good idea.
Samuel Verschelde 2015-05-31 22:47:44 CEST

Whiteboard: (none) => MGA5TOO FOR_ERRATA

Comment 31 Florian Hubold 2015-10-07 20:54:38 CEST
It seems it can be partly reproduced by simply running "time dkms status" - takes 12 seconds here with SSD on a very fast system, and I'm only using 4 dkms modules: dkms-broadcom-wl, dkms-nvidia-current, dkms-sysdig and dkms-virtualbox. But overall there are 27 versions of those modules present, including the binary dkms modules that get installed by default but which I don't use.

Most time seems to be consumed scanning for the "binary" dkms modules for nvidia-current and virtualbox. The actual dkms modules that I use, where no binary (read precompiled modules that we install by default, e.g. for nvidia and virtualbox, called e.g. nvidia-current-kernel and virtualbox-kernel ) dkms modules are present it is much quicker.

----

Steps to reproduce:

time dkms status 1>before.txt
urpme $(rpm -qa|grep kernel|grep -E "nvidia|virtualbox") 
# warning, the above removes all precompiled dkms kernel modules - make sure you don't need them and have the normal dkms modules built for the kernel you use
time dkms status 1>after.txt

----

[doktor5000@Mageia5]â[20:44:34]â[~] time sudo dkms status 1>before.txt

real    0m12.544s
user    0m0.882s
sys     0m1.257s


[doktor5000@Mageia5]â[20:45:49]â[~] sudo urpme $(rpm -qa|grep kernel|grep -E "nvidia|virtualbox")
nvidia-current-kernel-3.19.3-desktop-1.mga5-346.47-8.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.4-desktop-1.mga5-346.59-2.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.4-desktop-2.mga5-346.59-3.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.6-desktop-1.mga5-346.59-4.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.6-desktop-2.mga5-346.59-5.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.7-desktop-1.mga5-346.59-6.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.8-desktop-1.mga5-346.59-10.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.8-desktop-2.mga5-346.72-1.mga5.nonfree.x86_64 nvidia-current-kernel-3.19.8-desktop-3.mga5-346.82-1.mga5.nonfree.x86_64 nvidia-current-kernel-4.1.8-desktop-1.mga5-346.82-3.mga5.nonfree.x86_64 nvidia-current-kernel-desktop-latest-346.82-3.mga5.nonfree.x86_64 virtualbox-kernel-3.19.3-desktop-1.mga5-4.3.26-4.mga5.x86_64 virtualbox-kernel-3.19.4-desktop-1.mga5-4.3.26-5.mga5.x86_64 virtualbox-kernel-3.19.4-desktop-2.mga5-4.3.26-6.mga5.x86_64 virtualbox-kernel-3.19.6-desktop-1.mga5-4.3.26-7.mga5.x86_64 virtualbox-kernel-3.19.6-desktop-2.mga5-4.3.26-8.mga5.x86_64 virtualbox-kernel-3.19.7-desktop-1.mga5-4.3.26-9.mga5.x86_64 virtualbox-kernel-3.19.8-desktop-1.mga5-4.3.28-1.mga5.x86_64 virtualbox-kernel-3.19.8-desktop-2.mga5-4.3.28-2.mga5.x86_64 virtualbox-kernel-3.19.8-desktop-3.mga5-5.0.2-2.mga5.x86_64 virtualbox-kernel-4.1.8-desktop-1.mga5-5.0.4-2.mga5.x86_64 virtualbox-kernel-desktop-latest-5.0.4-2.mga5.x86_64 will be removed



[doktor5000@Mageia5]â[20:48:17]â[~] time sudo dkms status 1>after.txt

real    0m0.213s
user    0m0.054s
sys     0m0.018s
[doktor5000@Mageia5]â[20:48:31]â[~]

CC: (none) => doktor5000

Comment 32 Giuseppe Ghibò 2016-09-29 15:35:50 CEST
Apparently the problem seems in the way plymouth interacts with service_harddrake; there is a block in the code, which calls plymouth with hide-splash option and ha sexactly 2 minutes timeout. That's exactly the timeout "systemd-analyze blame" shows; the code is:

sub dialog_with_timeout {
    my ($type, $title, $msg, $timeout, $plymouth) = @_;
    my ($pid, $res, $timed_out);

    $SIG{ALRM} = sub { $timed_out = 1; kill 15, $pid };
    unless ($pid = fork()) {
        $plymouth and system('plymouth', 'hide-splash');
        exec("/usr/share/harddrake/confirm", $type, $title, $msg);

what the code seems to suggest is that in case of some new hardware detected a dialog box should appear (getting rid of the splash screen in the meanwhile) asking for an interactive confirm, but apparently this interactive dialog box doesn't appear, probably because plymouth --hide-splash is not able to get rid of the splash screen and switch to some interactive console or vt where one can answer about the questions, I suppose.

A quick fix would be to just get rid the hide-splash call option and call plymouth without it, or even replace with the debug option. In this case there is no more the timeout (and there isn't also this apparently missing interactive dialog box too).

CC: (none) => ghibomgx

Comment 33 Giuseppe Ghibò 2016-09-29 15:37:29 CEST
Created attachment 8467 [details]
quick fix to avoid the splash-hide in plymouth

I attach a quick fix to get rid of the hide-splash option with plymouth.
Samuel Verschelde 2016-10-18 13:10:41 CEST

Whiteboard: MGA5TOO FOR_ERRATA => MGA5TOO

Comment 34 Jeff Robins 2017-02-06 08:12:34 CET
I am also having very long boot times with MGA5.  After looking at "systemd-analyze blame", I see that the worst offenders are sendmail.service with 2min-2s, sm-client.service with 1min and mandriva-everytime.service with 59s.

I see the solution to mandriva-everytime.service, but I haven't seen one for sendmail and sm-client yet.  To fix the sendmail issue just alter /etc/hosts so that there is one line with localhost.localdomain.  I'm not sure if this is a valid FQDN, but it seems to work.  "systemctl restart sendmail" now takes 0.17sec, where it was taking 2min.

"127.0.0.1 <local system name> localhost.localdomain localhost"

I still need to uninstall my old kernels and make sure that fixes the problem.

Does anyone have a fix for sm-client?

--Jeff

CC: (none) => jeffrobinsSAE

Comment 35 Jeff Robins 2017-02-06 08:14:50 CET
Nevermind about sm-client.  It seems to be fixed with the same fix for sendmail.  I'm guess the "sm" stands for SendMail?
Comment 36 Florian Hubold 2017-07-22 08:55:27 CEST
(In reply to Jeff Robins from comment #35)
> Nevermind about sm-client.  It seems to be fixed with the same fix for
> sendmail.  I'm guess the "sm" stands for SendMail?

That's not actually on topic in this bugreport, this is about mandriva-everytime, not about any particular service that may take longer to start. And you can see for yourself via e.g. "systemctl show sm-client --property Description"


(In reply to Giuseppe Ghibò from comment #33)
> 
> I attach a quick fix to get rid of the hide-splash option with plymouth.

As plymouth has no maintainer, do you mind adding this for cauldron, if you can reproduce the boot delay and if that removes this delay? FWIW, this report is still valid for mga6, sadly

Summary: mga5b2 long boot time - mandriva-everytime.service => long boot time - mandriva-everytime.service
Whiteboard: MGA5TOO => MGA5TOO MGA6TOO

Comment 37 Sara Taylor 2019-01-14 07:55:08 CET
(In reply to magnux77 from comment #18)
> As said by TuxMad here
> (https://whatstatus.co/happy-birthday-meme), the fresh install of Mageia5B3
> doesn't encounter the problem, but the migration from 4 to 5 will cause this
> long boot due to lots of useless dkms. 
> 
> Few users are trying this migration today, so almost nobody complain.
> But, when Mageia 5 will be released official, there will be lots of users
> who will experiment the problem when migrating from Mageia4 . 
> 
> So I think all these useless dkms (and other stuff I don't know) have to be
> deleted or another solution has to be found for the migration from 4 to 5 to
> be successfull.

I guess DKMS_ONBOOT is meant to make sure that your kernel modules (like dkms-nvidia-current for example) are properly built for the kernel you are running. I'm not 100% sure, but I guess the reason for this feature is to ensure that you can boot _any_ kernel without having issue, since the kernel modules will be rebuilt on boot if they are not already built.

CC: (none) => adamgrayling

Christian Lohmaier 2019-07-09 12:10:23 CEST

Depends on: (none) => 25075


Note You need to log in before you can comment on or make changes to this bug.