I'm not sure if that's normal for ufw service to load in 5 seconds with systemd,
is it too long? If not, does alternatives work faster?
An addition to that, automated netctl wireless service takes 10 secs to load. Should I use network-manager or something? I'm sure it takes less than that.
I'm booting up in 21 seconds, so I'm on blank screen just to get a proper internet connection on.
here is my output...
[joyce@linuxg6-joyce tlp]$ systemd-analyze blame 9.609s netctl@wlo1\x2dyou\x20shall\x20not\x20pass.service 4.993s ufw.service 1.731s systemd-logind.service 1.593s systemd-fsck@dev-disk-by\x2duuid-08AD\x2d3F76.service 1.470s systemd-vconsole-setup.service 1.168s systemd-fsck@dev-disk-by\x2duuid-f061c423\x2dcbc1\x2d460b\x2d945c\x2dd85c478c855d.service 839ms systemd-binfmt.service 786ms systemd-fsck@dev-disk-by\x2duuid-e6fc1bb0\x2dc715\x2d4d9d\x2d91ff\x2d561910d1c545.service 785ms sys-kernel-debug.mount 753ms systemd-udev-trigger.service 745ms systemd-tmpfiles-setup-dev.service 725ms dev-mqueue.mount 724ms systemd-remount-fs.service 723ms sys-kernel-config.mount 712ms dev-hugepages.mount 675ms rc-local.service 659ms tmp.mount 627ms alsa-restore.service 588ms proc-sys-fs-binfmt_misc.mount 413ms polkit.service 408ms cpupower.service 304ms home.mount 299ms udisks2.service 298ms systemd-user-sessions.service 293ms systemd-journal-flush.service 288ms archive.mount 278ms systemd-random-seed-load.service 250ms dev-disk-by\x2duuid-195ad48c\x2d39a9\x2d4750\x2d967c\x2dd58a2c84b114.swap 249ms systemd-update-utmp.service 234ms systemd-sysctl.service 224ms systemd-udevd.service 129ms boot-efi.mount 72ms systemd-tmpfiles-clean.service 68ms systemd-tmpfiles-setup.service 56ms upower.service 30ms rtkit-daemon.service 985us sys-fs-fuse-connections.mount
NotaBene: Ignore the rc-local lameness, I was too lazy to change the name in guide while setting switcheroo.
Last edited by Gulver (2013-05-24 07:16:17)
I have always found the ufw takes quite a bit longer to load than iptables does. Why, I have no idea. I can't imagine that it is doing a whole hell of a lot more than plain iptables, since I think it actually uses the rules to make an iptables.rules file. My iptables loaded in 53ms the last boot, but that is probably because i use a SSD.
Netctl@.service (netctl enable <profile>) takes a long time because it does not actually complete until a dhcp address has been resolved. You can make this faster by using a static IP address if that is a possibility. Otherwise, you can set netctl-auto@.service it start instead. Because it is meant to switch between networks, it does not wait for the interface to actually resolve, but rather sends notification of completion after it is able to simply parse and load the profiles you have told it to.
I am curious how you have a service that is reporting in microseconds on the bottom. I thought it was general concensus a couple releases ago to simply round to the ms to make it look nicer. I have several services that start in 1ms, but I have not seen anything reported in us since 200(?) maybe.
As I searched,
it states to enable the temporary kernel datas to be meta-stored. Some trick about loading the kernel, I think. I don't know.
Real automated netctl is the way to got then!
I also shall try the iptables.
Last edited by Gulver (2013-05-24 07:51:34)
Just so you know , iptables isn't easy. There are reasons why there are wrappers around it with "uncomplicated" it its name.