Hardware - Support > Start up and Shutdown
Long startup due to hardcoded if-up apt-get update
rew:
--- Quote from: Jerry on January 05, 2023, 04:04:41 AM ---People post 'fixes' here. But that doesn't automatically translate to one size fits all. I have the /etc/network/if-up.d/update file and my boot times are not adversely affected by it. The other contributors here have also not reported a problem. As an OS provider, one must tread extremely carefully and usually only 'fix' things that affect the majority of people, and only then, with no foreseeable adverse affects. I can predict with more certainty how a necessary bug fix will affect everyone, but with this, I don't like to play 'trial and error' with 1000's of computers. There's a balance to consider. Interesting thread nonetheless. Thank you.
--- End quote ---
Thanks for your explanation. To me it's important that the maintainers of linuxliteos are aware of this problem. How you react to it, remains indeed your responsibility.
I have now tested that bug also on my somewhat newer and stronger machine.
--- Code: ---neofetch
,xXc rew@studlin460s
.l0MMMMMO ---------------
.kNMMMMMWMMMN, OS: Linux Lite 6.2 x86_64
KMMMMMMKMMMMMMo Host: 20FAS54D00 ThinkPad T460s
'MMMMMMNKMMMMMM: Kernel: 5.15.0-56-generic
kMMMMMMOMMMMMMO Uptime: 11 mins
.MMMMMMX0MMMMMW. Packages: 2401 (dpkg)
oMMMMMMxWMMMMM: Shell: bash 5.1.16
WMMMMMNkMMMMMO Resolution: 2560x1440
:MMMMMMOXMMMMW DE: Xfce
.0MMMMMxMMMMM; WM: Xfwm4
:;cKMMWxMMMMO WM Theme: Materia
'MMWMMXOMMMMl Theme: Materia [GTK2/3]
kMMMMKOMMMMMX: Icons: Papirus-Adapta [GTK2], Adwaita [GTK3]
.WMMMMKOWMMM0c Terminal: xfce4-terminal
lMMMMMWO0MNd:' Terminal Font: Droid Sans Mono 12
oollXMKXoxl;. CPU: Intel i7-6600U (4) @ 3.400GHz
':. .: .' GPU: Intel Skylake GT2 [HD Graphics 520]
.. Memory: 1615MiB / 19739MiB
--- End code ---
Turns out that I face severe boot issues also there ... around 10 secs.
Which doesn't really astonish me, it was clear that it is a network issue ... provoked by the 'apt update' enforcement.
I share my data and let you make your wise choices.
The attachments and measurements:
blame =
--- Code: ---systemd-analyze blame
--- End code ---
networking-services =
--- Code: ---journalctl -u networking.service -b
--- End code ---
plot =
--- Code: ---system-analyze plot
--- End code ---
- before the fix: Startup finished in 14.112s (firmware) + 20.537s (loader) + 4.202s (kernel) + 14.398s (userspace) = 53.250s graphical.target reached after 14.374s in userspace
- after the fix: Startup finished in 8.840s (firmware) + 19.691s (loader) + 4.185s (kernel) + 9.536s (userspace) = 42.254s graphical.target reached after 9.529s in userspace
... we clearly see that I have problem on boot to reach the update sites. See journalctl output. This is not issue at real runtime. I don't mind however ... chmod -x update and I have a blazing fast liteos.
users voice
To share my view, I was experimenting with other distros despite my preference for linuxliteos, as I was frustrated with the latest boot performance. And I almost left you guys, wouldn't I have found this cause for the severe boot problem ... so I now stay with the distro. If other users make the same choices, the price for this 'apt update' may be much bigger than you think.
I saw other threads that report slow boot ... around networking.service ... I assume they suffer from the same effect.
Someone delete my post to help these users, thats ok, your business. But with this we never get more data whether more users are affected.
Jerry:
Sure, here's my output. Is there anything suspicious in it? - Define suspicious?
People post 'fixes' here. But that doesn't automatically translate to one size fits all. I have the /etc/network/if-up.d/update file and my boot times are not adversely affected by it. The other contributors here have also not reported a problem. As an OS provider, one must tread extremely carefully and usually only 'fix' things that affect the majority of people, and only then, with no foreseeable adverse affects. I can predict with more certainty how a necessary bug fix will affect everyone, but with this, I don't like to play 'trial and error' with 1000's of computers. There's a balance to consider. Interesting thread nonetheless. Thank you.
rew:
--- Quote from: Jerry on January 03, 2023, 09:24:40 PM ---What is the output of:
--- Code: ---systemctl list-dependencies --reverse network-online.target
--- End code ---
--- End quote ---
Sure, here's my output. Is there anything suspicious in it?
--- Code: ---systemctl list-dependencies --reverse network-online.target network-online.target
● ├─cups-browsed.service
● ├─hddtemp.service
● ├─nmbd.service
● ├─packagekit.service
● └─smbd.service
--- End code ---
--- Quote from: stevef on January 04, 2023, 12:07:55 AM ---
--- Quote ---With third-party I am talking about the packages I tend to install ... I don't think it makes sense to post a full list of all my deps.
--- End quote ---
Do you notice the problem with the boot times of the machines prior to installing your third party packages ?
If I suspected third party package may be responsible I'd be doing analysis of boot ups before and after the third party stuff installation.
--- End quote ---
The boot time is evidently lost by an `apt update` call on if-up. So the third-parties that cause that delay can be best reported by the sources I'm going to look up on during this. In my case the firefox ppa is indeed a third-party. But any other repo that users may use, will cause further delays on everydays boot routines.
--- Code: ---sudo apt update
[sudo] password for rew:
Hit:1 http://ubuntu.ethz.ch/ubuntu jammy InRelease
Hit:2 http://ubuntu.ethz.ch/ubuntu jammy-updates InRelease
Hit:3 http://ubuntu.ethz.ch/ubuntu jammy-backports InRelease
Hit:4 http://ubuntu.ethz.ch/ubuntu jammy-security InRelease
Hit:5 https://ppa.launchpadcontent.net/mozillateam/ppa/ubuntu jammy InRelease
Hit:6 http://archive.canonical.com/ubuntu jammy InRelease
Hit:7 http://repo.linuxliteos.com/linuxlite fluorite InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
9 packages can be upgraded. Run 'apt list --upgradable' to see them.
--- End code ---
--- Quote from: Jerry on January 03, 2023, 09:34:55 PM ---If this is your PC - http://www.hp.com/hpinfo/newsroom/press_kits/2008/connecting/ds_bn_6730b.pdf I wouldn't expect miracle boot times once the networking issue is resolved. Windows Vista :) Throw in an SSD (if possible) and tweak the boot, static IP helps a lot, and you should be in business.
--- End quote ---
Yes, this is my PC ... it's pretty old. The full boot time, after the fix, is around 45 secs (including bios).
See also my plots on the actual boot time (with the fix applied of course):
--- Code: ---Startup finished in 6.053s (kernel) + 8.989s (userspace) = 15.043s graphical.target reached after 8.962s in userspace
--- End code ---
So I'm happy with that. And yes it has an SSD. But actually I wasn't looking for advice how to make it faster: I've measured and isolated a clear problem, and want to share with other users how to get rid of it, or even better get it fixed in the linuxliteos distribution.
At any rate, we all know that a call of `apt update` doesn't come for free regardless of the disk. The more repos we're contacting the more this command delays the actual boot performance. We're making a real-world network operation.
My latest boot time measures are attached: today I only waste 10secs with it ... looks as if it highly depends on the activity on my network. What about if I'm connected to a slow network, having DNS problems, or a slow link, or using a slow mirror? The boot time will decrease almost beyond limits ... without any value for me as a user. That is not even a problem of old hardware ... but one of having a fast link to the web + mirrors.
stevef:
--- Quote ---With third-party I am talking about the packages I tend to install ... I don't think it makes sense to post a full list of all my deps.
--- End quote ---
Do you notice the problem with the boot times of the machines prior to installing your third party packages ?
If I suspected third party package may be responsible I'd be doing analysis of boot ups before and after the third party stuff installation.
Jerry:
If this is your PC - http://www.hp.com/hpinfo/newsroom/press_kits/2008/connecting/ds_bn_6730b.pdf I wouldn't expect miracle boot times once the networking issue is resolved. Windows Vista :) Throw in an SSD (if possible) and tweak the boot, static IP helps a lot, and you should be in business.
Navigation
[0] Message Index
[#] Next page
Go to full version