Help troubleshooting slow boot

asked 2019-02-20 11:34:59 -0500

keponk gravatar image

updated 2019-02-21 09:44:30 -0500

Hi all, I've noticed lately my boot time is super long. I'm running Fedora 29 on a razer 14 2017 laptop. I'm running on an SSD and it seems a bit crazy this takes so long.

Running systemd-analayze blame shows the follownig:

~ systemd-analyze blame 
     58.232s initrd-switch-root.service
      7.607s dracut-initqueue.service
      6.990s systemd-cryptsetup@luks\x2d58e140f2\x2d55a6\x2d4914\x2d94b3\x2df4b52b7e6faa.service
      5.541s fwupd.service
      5.496s NetworkManager-wait-online.service
      5.078s bolt.service
      4.236s plymouth-quit-wait.service
      2.378s systemd-udev-settle.service
       891ms akmods.service
       601ms firewalld.service
       496ms systemd-journal-flush.service
       391ms lvm2-monitor.service
       297ms udisks2.service
       226ms accounts-daemon.service
       222ms libvirtd.service
       219ms systemd-logind.service
       200ms upower.service
       197ms dnf-makecache.service
       166ms user@42.service
       130ms dracut-pre-pivot.service
       127ms initrd-parse-etc.service
       127ms ModemManager.service
       125ms abrtd.service

So the culpirt seems to be the initrd-switch-root.service . Added to this, I've noticed plymouth gets stuck until I click ESC to see what's going on and as soon as I do that then it seems that it moves along. wtf.

Any clues about what might be happening?

EDIT 21/02/19

dmesg output at :

as requested in the comments here the output of systemd-analyze critical-chain:

~  sudo 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. @7.841s
└─ @7.841s
  └─akmods.service @2.838s +887ms
    └─ @2.837s
      └─ @2.837s
        └─iscsid.socket @2.837s
          └─ @2.833s
            └─sys-fs-fuse-connections.mount @16.214s +4ms
              └─systemd-modules-load.service @652ms +42ms
edit retag flag offensive close merge delete



Hello, Please give me the output of systemd-analyze critical-chain and dmesg from a fresh boot.

ghostv71 gravatar imageghostv71 ( 2019-02-20 15:51:10 -0500 )edit

@ghostv71 updated description with the info

keponk gravatar imagekeponk ( 2019-02-21 09:43:49 -0500 )edit

Big gap between 9 and 39 seconds. I'm guessing some kind of nvme weirdness but I'm not sure how to get more info or work around it. If this system has Windows on it, you might check make/model with the manufacturer if there are any firmware (BIOS) updates, as well as firmware updates for the NVMe drive. I'm not seeing in dmesg the make/model of the drive, you should be able to get this with lspci -vnand plug it into Maybe a bug has and work around are already reported.

cmurf gravatar imagecmurf ( 2019-02-21 18:21:27 -0500 )edit

"initrd-switch-root.service" does do the switch from your initramfs to the real system on disk.

can you please check:

a) what login manager do you use ( lightdm or GDM ) ? b) do you have additional display hw in your device?

because: we had a dell laptop where lightdm showed the login screen on a not connected internal gfx card. if plymouth displays something similar, and you hit esc, you skip this request, which you do not "see". (gdm did not have this problem)

rdtcustomercare gravatar imagerdtcustomercare ( 2019-02-24 11:07:20 -0500 )edit

I have a similar problem but it seems to be related to the Kernel as the long boot time only started after an upgrade to Kernel 5. My boot time is now 2 minutes but 47 seconds if I revert to my previous Kernel 4.

My systemd-analyze blame shows dracut and lvm as the culprits I will continue to work on it

58.386s dracut-initqueue.service 34.834s lvm2-pvscan@8:3.service 22.618s lvm2-monitor.service 12.498s NetworkManager-wait-online.service 10.717s systemd-udev-settle.service

alain gravatar imagealain ( 2019-03-27 07:51:43 -0500 )edit