You are not logged in.
Pages: 1
Hi,
I'm experiencing some issues at boot.
Some services are taking too much time at boot.
Here is the critical-chain
graphical.target @11.271s
└─multi-user.target @11.271s
└─NetworkManager.service @9.085s +2.185s
└─basic.target @9.084s
└─sockets.target @9.084s
└─dbus.socket @9.084s
└─sysinit.target @9.076s
└─systemd-update-utmp.service @8.882s +192ms
└─systemd-tmpfiles-setup.service @8.739s +109ms
└─local-fs.target @8.739s
└─media-home.mount @8.667s +70ms
└─systemd-fsck@dev-sda7.service @7.974s +629ms
└─dev-sda7.device @7.974s
What can I do?
Offline
What filesystem do you have on sda7. Also; where do you mount it, /home? Can you also post output from `systemd-analyze blame`
Last edited by Spider.007 (2015-07-10 11:43:11)
Offline
Is this the whole output?
Offline
this is sda7 on fstab
/dev/sda7 /media/home ext4 defaults 0 2
yes this is the whole output
here is the blame output
5.845s dev-sda4.device
1.870s systemd-journald.service
1.832s NetworkManager.service
1.796s systemd-vconsole-setup.service
1.287s systemd-udevd.service
1.076s systemd-sysctl.service
1.057s polkit.service
898ms systemd-tmpfiles-setup-dev.service
881ms wpa_supplicant.service
845ms systemd-journal-flush.service
533ms systemd-backlight@backlight:acpi_video0.service
503ms systemd-backlight@backlight:intel_backlight.service
345ms systemd-binfmt.service
331ms systemd-rfkill@rfkill0.service
330ms systemd-rfkill@rfkill1.service
312ms systemd-remount-fs.service
284ms systemd-fsck@dev-sda7.service
207ms udisks2.service
163ms systemd-tmpfiles-setup.service
162ms dev-mqueue.mount
157ms systemd-random-seed.service
156ms kmod-static-nodes.service
152ms sys-kernel-config.mount
148ms sys-kernel-debug.mount
147ms systemd-udev-trigger.service
127ms systemd-timesyncd.service
125ms tmp.mount
125ms dev-hugepages.mount
121ms systemd-rfkill@rfkill2.service
117ms user@1000.service
109ms systemd-logind.service
101ms upower.service
95ms systemd-update-utmp.service
60ms media-home.mount
48ms rtkit-daemon.service
46ms systemd-user-sessions.service
21ms alsa-restore.service
20ms proc-sys-fs-binfmt_misc.mount
Last edited by steoguitar (2015-07-10 13:32:50)
Offline
any ideas?
Offline
What's taking too long? Is this a HDD or a SDD? If it's an HDD, these seem like excellent times, 3.3 seconds for userspace is great.
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
It's an HDD.
I think that networkmanager and tmp.service are taking too long...
Last edited by steoguitar (2015-07-13 13:07:37)
Offline
It's an HDD.
I think that networkmanager and tmp.service are taking too long...
2 seconds is pretty quick for network manager - but NM is slow in general if you ask me. You could try an alternative. As for tmp.service, which one? The only one showing above is 95 millseconds - how fast would you want it?
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
Please, take a look at my first output (critical-chains)
It shows that systemd-tmpfiles take very long time and NM too...
Voices with a value within a + character are in red
Last edited by steoguitar (2015-07-13 15:45:24)
Offline
What do you consider a very long time? 192 milliseconds?? I don't mean to be critical - I just can't help but suspect there is a more basic misunderstanding or miscommunication here. Is your boot taking a lot longer than what is represented in that output, or are you misreading the output - because everything that is presented in this thread shows that you actually have a pretty quick boot time.
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
I really appreciate your support.
At startup from the OS selection on GRUB to a usable xfce desktop it takes 45 seconds.
I'm searching the reason why even though systemd takes 15 sec to boot a light DE as xfce takes very long...
I thought that the problem was written on critical-chains.
Sorry but I don't have comparison with other machines
Offline
So what does `systemd-analyze` say? Is it kernel, userspace or is the time nowhere near your 45 seconds?
Offline
I thought that the problem was written on critical-chains.
Where? Your critical-chain output from the first post shows 11.2 seconds. That's been my point: for an HDD system, 11.2 seconds in userspace is a very good time.
But critical chain does not show the time spent in the firmware/bios, the bootloader, or the "kernel-space"; critical chain only shows the userspace processes (I think ... it may also show kernelspace, but definitely not bios or bootloader).
So the comment above is getting at that: which part of the boot process is actually taking up the time. If you don't already have it, adding the systemd hook to mkinitcpio.conf (and rebuilding the initramfs) can get a little more detailed information as that allows systemd-analyze to partition the time more specifically. For example, my systemd-analyze output:
Startup finished in 3.586s (firmware) + 180ms (loader) + 501ms (kernel) + 787ms (initrd) + 5.790s (userspace) = 10.847s
Here I can see almost all the boot time is spent in userspace (though that is misleading as I am in a working desktop long before that 5.7 seconds passes), so if I wanted to streamline my boot time further, I'd look at the critical chain output. But if the userspace time was low and the loader time was where the lag was, the critical-chain output would be useless.
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
Pages: 1