PDA

View Full Version : [kubuntu] Slow boot after 17.10 -> 18.04 upgrade



piu130
April 30th, 2018, 04:09 PM
I've upgraded my kubuntu 17.10 to 18.04 recently. Now the boot is really slow. From about 8 seconds to 40 seconds. The
dmesg peaks are:


[ 3.772885] [drm] RC6 on
[ 34.502625] EXT4-fs (nvme0n1p5): mounted filesystem with ordered data mode. Opts: (null)

and maybe


[ 36.626981] Bluetooth: hci0: Applying Intel DDC parameters completed
[ 39.306767] Could not find key with description: [448f93ab6910c98f]

Does someone has the same issue? How can I fix it?

Claus7
May 1st, 2018, 02:03 AM
Hello,

I'm searching the issue myself as well for my ubu box. I do not face such an issue as yours, yet I came accross the following commands:

systemd-analyze time
systemd-analyze blame
systemd-analyze critical-chain

which probably will shed more light to your issue.

Some thoughts from the messages you receive: do you have any bluetooth dongle attached to your system? Trying to remove it does it affect your boot time in great extent?

I do not have any messages that are the same as yours appart from:

[ 4.294531] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts: (null)

Regards!

kassanunda
May 1st, 2018, 06:38 AM
Same issue.

[ 2.844760] [drm] RC6 on
[ 63.454571] EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts: (null)


I also tried modifying /etc/initramfs-tools/conf.d/resume (resume=none (instead of UUID)) but no help

piu130
May 1st, 2018, 07:43 AM
systemd-analyze time
Startup finished in 34.554s (kernel) + 5.078s (userspace) = 39.633s
graphical.target reached after 5.074s in userspace

systemd-analyze blame
4.107s NetworkManager-wait-online.service
361ms systemd-logind.service
340ms dev-nvme0n1p6.device
268ms mpd.service

systemd-analyze critical-chain
The time after the unit is active or started is printed after the "@" character.
The time the unit takes to start is printed after the "+" character.

graphical.target @5.074s
└─multi-user.target @5.074s
└─kerneloops.service @5.067s +6ms
└─network-online.target @5.067s
└─NetworkManager-wait-online.service @959ms +4.107s
└─NetworkManager.service @794ms +162ms
└─dbus.service @769ms
└─basic.target @761ms
└─sockets.target @761ms
└─snapd.socket @760ms +1ms
└─sysinit.target @758ms
└─systemd-timesyncd.service @588ms +170ms
└─systemd-tmpfiles-setup.service @574ms +12ms
└─local-fs.target @572ms
└─boot-efi.mount @564ms +7ms
└─systemd-fsck@dev-disk-by\x2duuid-1A02\x2dCCC9.service @506ms +57ms
└─dev-disk-by\x2duuid-1A02\x2dCCC9.device @505ms


Some thoughts from the messages you receive: do you have any bluetooth dongle attached to your system? Trying to remove it does it affect your boot time in great extent?
Build in bluetooth (notebook)

Claus7
May 1st, 2018, 01:25 PM
Hello,



systemd-analyze time
Startup finished in 34.554s (kernel) + 5.078s (userspace) = 39.633s
graphical.target reached after 5.074s in userspace


If you compare it with mine (I use hdd and installing different kernels at the time):

$ systemd-analyze time
Startup finished in 5.483s (kernel) + 2min 32.315s (userspace) = 2min 37.798s
graphical.target reached after 2min 32.308s in userspace

then you will see that you have a huge boot time of your kernel.
Which kerner are you using? Check it with:

uname -a

I would suggest you to try a different one, if possible, and see the results. I cannot see any other component affecting your boot time.

Regards!

piu130
May 2nd, 2018, 07:24 AM
uname -a
gives 4.15 then I tried it with 4.13. Same problem.

I turned off "quiet splash" in grub and the boot stocks at the line


Begin: Running /scripts/local-premount ... [ 3.805190] [drm] RC6 on

piu130
May 2nd, 2018, 08:03 AM
I found the solution here: https://askubuntu.com/questions/1013830/slow-boot-long-kernel-load-time-due-to-wrong-resume-device

In the file /etc/initramfs-tools/conf.d/resume there was an UUID defined that did not match any of my partitions UUID ('sudo blkid' to see the UUIDs).

So I wrote 'RESUME=none' instead of 'RESUME=UUID=xyz' and then executed 'sudo update-initramfs -u'.

Is there a problem with this solution? What UUID should be there to resume instead of none or the wrong one? Should I use cryptswap UUID there? Or the normal swap?

Claus7
May 2nd, 2018, 02:05 PM
Hello,

first of I do not have such a configuration. This means that in my case the numbers are the "correct" ones (the output UUID from blkid of the swap partition is the same as the one in the resume file).

Now in your case:
1) by adding none it seems to me that you are disabling your swap partition. Take a look also here:
https://ubuntuforums.org/showthread.php?t=2291078

2) you could add the UUID of your swap partition instead of none in order not to deactivate it:
so from sudo blkid check the string number that you get for your swap partition

then go to /etc/initramfs-tools/conf.d/resume file and add the string number you just got

finally run sudo update-initramfs -u to apply changes.

You could check it out and see if it works.

Regards!