Slow Boot Process, sluggish system, rpm-ostree gets stuck

Ask Fedora
22

Hi Marius,

thanks for the hint to this sophisticated journalctl command. Absolutely appreciate your time and effort.

Referring to what Chris noticed in his reply, that attached USB devices might cause the issue, I used your command right after booting once with USB attached and once without

journalctl --boot=0 --output=short-delta | grep --invert-match '< *0.' without attached USB device

[    2.313827                ] fedora kernel: Linux version 6.18.10-200.fc43.x86_64 (mockbuild@bec669e2407a45e7885294d0b7351b99) (gcc (GCC) 15.2.1 20260123 (Red Hat 15.2.1-7), GNU ld version 2.45.1-4.fc43) #1 SMP PREEMPT_DYNAMIC Wed Feb 11 17:20:05 UTC 2026
[   15.136944 <   10.913297 >] fedora systemd-cryptsetup[631]: Set cipher aes, mode xts-plain64, key size 512 bits for device /dev/disk/by-uuid/4a3d4c2d-6b91-4543-be49-ef9c869f3dcd.
[   18.257683 <    3.120739 >] fedora systemd[1]: Found device dev-disk-by\x2duuid-8b3be73f\x2d77bd\x2d4dc0\x2d9efd\x2dd1578d945c55.device - /dev/disk/by-uuid/8b3be73f-77bd-4dc0-9efd-d1578d945c55.
[   19.887208 <    1.159539 >] SilverFed systemd-journald[324]: Received SIGTERM from PID 1 (systemd).
[   29.207870 <    1.259149 >] SilverFed NetworkManager[1520]: <info>  [1771342453.8056] dhcp6 (wlo1): activation: beginning transaction (timeout in 45 seconds)
[   30.683416 <    1.427184 >] SilverFed gdm-password][3321]: gkr-pam: unable to locate daemon control file
[   48.253247 <    3.100120 >] SilverFed rpm-ostree[5672]: Handling GetPackages for caller :1.130

journalctl --boot=0 --output=short-delta | grep --invert-match '< *0.' with USB device

[   10.764495                ] fedora kernel: Linux version 6.18.10-200.fc43.x86_64 (mockbuild@bec669e2407a45e7885294d0b7351b99) (gcc (GCC) 15.2.1 20260123 (Red Hat 15.2.1-7), GNU ld version 2.45.1-4.fc43) #1 SMP PREEMPT_DYNAMIC Wed Feb 11 17:20:05 UTC 2026
[   29.241479 <   13.599145 >] fedora systemd-cryptsetup[638]: Set cipher aes, mode xts-plain64, key size 512 bits for device /dev/disk/by-uuid/4a3d4c2d-6b91-4543-be49-ef9c869f3dcd.
[   52.637966 <   23.396487 >] fedora systemd[1]: Found device dev-disk-by\x2duuid-8b3be73f\x2d77bd\x2d4dc0\x2d9efd\x2dd1578d945c55.device - /dev/disk/by-uuid/8b3be73f-77bd-4dc0-9efd-d1578d945c55.
[   57.572376 <    3.601802 >] SilverFed systemd-journald[325]: Received SIGTERM from PID 1 (systemd).
[   67.104836 <    1.917001 >] SilverFed audit[1445]: NETFILTER_CFG table=firewalld_probe:2 family=1 entries=1 op=nft_register_table pid=1445 subj=system_u:system_r:firewalld_t:s0 comm="firewalld"
[   78.091459 <    1.621623 >] SilverFed chronyd[1297]: Selected source 192.168.178.1 (_gateway)
[   80.337955 <    1.005615 >] SilverFed gnome-shell[1638]: Will monitor session c1
[   85.872636 <    2.578137 >] SilverFed systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully.
[  103.492825 <   17.619991 >] SilverFed gnome-shell[1638]: Error connecting to the screencast service
[  109.108913 <    5.616088 >] SilverFed systemd[1]: systemd-localed.service: Deactivated successfully.
[  112.040343 <    2.625509 >] SilverFed systemd[1578]: dbus-:1.3-org.gnome.Shell.Screencast@0.service: Consumed 5.342s CPU time, 382.5M memory peak.
[  120.074440 <    1.693230 >] SilverFed systemd-logind[1359]: Failed to open '/boot/loader/entries': Remote address changed
[  130.009267 <    1.312458 >] SilverFed systemd[3445]: Started dbus-:1.2-org.gnome.Epiphany.WebAppProvider@0.service.
[  132.079375 <    1.092279 >] SilverFed systemd[3445]: Started dbus-:1.2-org.gnome.Ptyxis@0.service.
[  133.177658 <    1.098283 >] SilverFed fwupd[5951]: 19:33:25.460 FuMain               fwupd 2.0.19 ready for requests (locale de_DE.UTF-8)
[  136.269036 <    1.174733 >] SilverFed rpm-ostree[5952]: In idle state; will auto-exit in 62 seconds
[  138.702304 <    2.042647 >] SilverFed systemd[1]: Stopping user@60578.service - User Manager for UID 60578...
[  141.328491 <    2.448217 >] SilverFed rpm-ostree[5952]: Txn RefreshMd on /org/projectatomic/rpmostree1/fedora successful
[  143.001759 <    1.668001 >] SilverFed realmd[2726]: quitting realmd service after timeout
[  150.291058 <    7.261223 >] SilverFed systemd[3445]: Started dbus-:1.2-org.gnome.Epiphany.WebAppProvider@1.service.
[  151.488894 <    1.076821 >] SilverFed systemd[1]: systemd-hostnamed.service: Deactivated successfully.
[  152.654264 <    1.124428 >] SilverFed systemd[1]: systemd-localed.service: Deactivated successfully.
[  156.253080 <    3.547246 >] SilverFed rpm-ostree[5952]: Importing rpm-md...done
[  162.399836 <    5.834283 >] SilverFed audit: BPF prog-id=93 op=UNLOAD
[  164.065893 <    1.351058 >] SilverFed rpm-ostree[5952]: Handling GetPackages for caller :1.133
[  169.924069 <    5.803428 >] SilverFed rpm-ostree[5952]: Importing rpm-md...done
[  173.598221 <    3.6741 52 >] SilverFed rpm-ostree[5952]: Handling GetPackages for caller :1.133
[  179.413986 <    5.765993 >] SilverFed rpm-ostree[5952]: Importing rpm-md...done
[  180.801587 <    1.385717 >] SilverFed rpm-ostree[5952]: Handling GetPackages for caller :1.133
[  187.151550 <    6.299070 >] SilverFed rpm-ostree[5952]: Importing rpm-md...done
[    2.379436                ] fedora kernel: Linux version 6.18.10-200.fc43.x86_64 (mockbuild@bec669e2407a45e7885294d0b7351b99) (gcc (GCC) 15.2.1 20260123 (Red Hat 15.2.1-7), GNU ld version 2.45.1-4.fc43) #1 SMP PREEMPT_DYNAMIC Wed Feb 11 17:20:05 UTC 2026
[   14.863027 <   10.704077 >] fedora systemd-cryptsetup[634]: Set cipher aes, mode xts-plain64, key size 512 bits for device /dev/disk/by-uuid/4a3d4c2d-6b91-4543-be49-ef9c869f3dcd.
[   17.990307 <    3.127280 >] fedora systemd[1]: Found device dev-disk-by\x2duuid-8b3be73f\x2d77bd\x2d4dc0\x2d9efd\x2dd1578d945c55.device - /dev/disk/by-uuid/8b3be73f-77bd-4dc0-9efd-d1578d945c55.
[   19.692653 <    1.096756 >] SilverFed systemd-journald[326]: Received SIGTERM from PID 1 (systemd).
[   28.865834 <    1.508438 >] SilverFed NetworkManager[1508]: <info>  [1771359309.4180] dhcp6 (wlo1): activation: beginning transaction (timeout in 45 seconds)
[   30.135710 <    1.217296 >] SilverFed gdm-password][3315]: gkr-pam: unable to locate daemon control file
[   47.341412 <    2.906698 >] SilverFed rpm-ostree[5795]: Handling GetPackages for caller :1.128

I lack the Linux competence to evaluate these logs, but find it astounding, that in the case of an attached USB mouse there are so many service showing up in the log like fwupd, gnome.Shell.Screencast@0.service: Consumed 5.342s CPU time, 382.5M memory peak, gnome.Epiphany.WebAppProvider@1.service using up a lot of the boot time. As if an active USB port would slow down or even trigger many services.

23

Glad to hear I could bring you a step further. I am not a USB expert though (and lack a little time atm ^^), and the logs don’t say much explicit other than USB correlates with the issue.

Let’s see if others can help you to get more details. Otherwise, I would test the following

  1. test the LAST update state of your Silverblue with USB
  2. test the LAST update without USB
  3. test the CURRENT update state with USB
  4. test the CURRENT update state without USB

Do not make any changes to your system other than switching between update states and adding/removing USB, in order to ensure the logs are comparable with only the differences we intentionally put in place.

→ compare LAST and CURRENT update state logs both with USB and check out if there are differences that could be linked (which means compare everything up to the time the issue has happened)

→ If the issue/delay occurs only in the CURRENT update state with USB, this might be worth a bug report, containing these logs and an elaboration.

In the old editions, we could change individual components and verify which component makes the difference: e.g., boot with a different kernel but keep all other updates untouched. In Silverblue, I assume you can only switch between massively different update states and no longer easily test if/when an individual component such as the kernel makes the difference. Not sure if there is an alternative in Silverblue? Feels like a major disadvantage in troubleshooting :open_mouth:

24

Hi Pascal, happy to hear you got your boot times figured out. It would br nice to be sble to somehow debug how the USB attached devices cause a stall in the boot and report those as bugs. However as I don’t run the atomic versions of the distro my debug experience kind of stops here :grinning_face_with_smiling_eyes:

Do note that not all the log entries pertain to the boot sequence, they are a collection of all events in the system and the --boot argument only filters the starting/end points (0 current boot, - 1 previous, - 2 etc). So in normal daily usage stuff like the browser (Epiphany) taking up multiple cpu seconds is normal.

25

Hi Chris,

thanks for you reply.

I believe, I found the culprit and solved the issue. (BTW: the “USB issue” came up with any update state)

An incorrect system time was set in the BIOS - you had previously noticed the journalctl logs the time /date jumps

Once I had synched BIOS time with my Linux system time, the boot delay disappeared, now it takes about 1 second longer with a USB device plugged in and total boot time seems OK for me.

1 Like
26

Hi Marius, thanks for you reply. I believe, I found the culprit and solved the issue as written in my reply to Chris.

An incorrect system time was set in the BIOS. Once I had synched BIOS time with my Linux system time, the boot delay disappeared, now it takes about 1 second longer with a USB device plugged in and total boot time seems OK for me.

BTW: I had generated the logs immediately after logging into the Gnome Shell without starting any programs except for the terminal emulator. That’s why I was so surprised to see processes like Epiphany browser showing, I don’t even had it installed as a program