lagita@fedora:~$ systemd-analyze
Startup finished in 2.206s (firmware) + 11.249s (loader) + 3.261s (kernel) + 2.290s (initrd) + 23.107s (userspace) = 42.114s
graphical.target reached after 22.993s in userspace.```
im surprised im not the last one, but still horrendous since the last uses HDD. So turns out majority of it is spent on waiting for docker, did not realize it took that long.
lagita@fedora:~$ systemd-analyze critical-chain
The time when unit became active or started is printed after the “@” character.
The time the unit took to start is printed after the “+” character.
graphical.target
└─multi-user.target
└─docker.service
+10.116s
└─network-online.target
└─NetworkManager-wait-online.service
+4.816s
└─NetworkManager.service
+2.317s
└─network-pre.target
└─dkms.service
+2.255s
└─basic.target
└─dbus-broker.service
+49ms
└─dbus.socket
└─sysinit.target
└─systemd-update-utmp.service
+33ms
└─auditd.service
+65ms
└─systemd-tmpfiles-setup.service
+143ms
└─local-fs.target
└─run-snapd-ns-cups.mnt.mount
└─run-snapd-ns.mount
└─swap.target
└─dev-zram0.swap
+134ms
└─systemd-zram-setup@zram0.service
+27ms
└─dev-zram0.device
michael ~ $ systemd-analyze
Startup finished in 13.784s (firmware) + 7.578s (loader) + 2.206s (kernel) + 28.928s (initrd) + 9.007s (userspace) = 1min 1.505s
graphical.target reached after 9.007s in userspace.
I don’t know if it’s fast or not, judging by the others, it’s not exactly great, but it’s enough for me.
michael ~ $ systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.
graphical.target @9.007s
└─multi-user.target @9.007s
└─rsyslog.service @8.913s +92ms
└─network-online.target @8.911s
└─NetworkManager-wait-online.service @3.516s +5.394s
└─NetworkManager.service @3.203s +311ms
└─network-pre.target @3.202s
└─firewalld.service @2.509s +692ms
└─polkit.service @2.362s +142ms
└─basic.target @2.341s
└─dbus-broker.service @2.280s +43ms
└─dbus.socket @2.267s +850us
└─sysinit.target @2.266s
└─systemd-resolved.service @2.199s +66ms
└─systemd-tmpfiles-setup.service @2.050s +145ms
└─local-fs.target @2.045s
└─boot-efi.mount @2.025s +19ms
└─boot.mount @1.997s +25ms
└─systemd-fsck@dev-disk-by\x2duuid-bdb632cf\x2d8731\x2d4ab2\x2d84ea\x2d0c95712ced54.service @1.516s +59ms
└─dev-disk-by\x2duuid-bdb632cf\x2d8731\x2d4ab2\x2d84ea\x2d0c95712ced54.device
By disabling rsyslog.service and NetworkManager-wait-online.service, for some reason the loading became much faster according to the logs, interesting:
michael ~ $ systemd-analyze
Startup finished in 9.381s (firmware) + 4.715s (loader) + 2.040s (kernel) + 14.057s (initrd) + 4.350s (userspace) = 34.546s
graphical.target reached after 4.240s in userspace.
michael ~ $ systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.
graphical.target @4.240s
└─multi-user.target @4.240s
└─tuned-ppd.service @4.083s +156ms
└─tuned.service @3.746s +335ms
└─network.target @3.742s
└─NetworkManager.service @3.460s +280ms
└─network-pre.target @3.459s
└─firewalld.service @2.667s +791ms
└─polkit.service @2.448s +197ms
└─basic.target @2.413s
└─dbus-broker.service @2.354s +41ms
└─dbus.socket @2.342s +833us
└─sysinit.target @2.340s
└─systemd-resolved.service @2.266s +74ms
└─systemd-tmpfiles-setup.service @2.127s +134ms
└─local-fs.target @2.123s
└─boot-efi.mount @2.102s +20ms
└─boot.mount @2.080s +20ms
└─systemd-fsck@dev-disk-by\x2duuid-bdb632cf\x2d8731\x2d4ab2\x2d84ea\x2d0c95712ced54.service @1.585s +72ms
└─dev-disk-by\x2duuid-bdb632cf\x2d8731\x2d4ab2\x2d84ea\x2d0c95712ced54.device
Personally I uninstall rsyslog and clean up the logs in /var/log.
I have little use for log file now that nearly everything in in the system and user journals.
)
Feels like pulling up to a Porsche Cars & Coffee meet with a clapped out Honda.
)