You are not logged in.
Arm-the-Homeless wrote:I got the same problem with Google Earth.
I just removed it. tongue
This is no solution.
regards,
dgr48
I'm just saying that you're not the only one getting the problem.
Offline
Finit-arc-git sensitive update
Try it now.
Changes:
- new static devices added
- LVM support (thanks to Jan Spakula)
- steps order edited (udev and hal could work now)
- code clear
Last edited by adriano (2009-05-10 12:29:13)
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
Finit-arc-git sensitive update
Try it now.Changes:
- new static devices added
- LVM support (thanks to Jan Spakula)
- steps order edited (udev and hal could work now)
- code clear
udev works, and so does my keyboard
However, the clock still resets itself, causing a long list of "WRITE TIME IS IN THE FUTURE [FIXED]" errors at boot.
EDIT: I see that's on your list of known bugs. OK.
Last edited by buttons (2009-05-10 14:32:21)
Cthulhu For President!
Offline
Everything bout the known time bug seems to wor fine - X and GDM loads blazing fast!
Proud ex-maintainer of firefox-pgo
Offline
Thanks, this looks great, I will try it out.
You say UUIDs don't work. What about other /dev/disk/foo/bar devices? Like by-id names, or labels? Using "normal" /dev/sd* names can be annoying when I add a hard drive and Linux randomly decides it's /dev/sda now. Especially since I use a LUKS-encrypted swap from /etc/crypttab -- can I use /dev/disk/by-id/ there? Is this a planned feature, and if so, will it slow the init down? Or is it intrinsic to the method used to speed boot time?
Thanks again!
Offline
Hi all,
It gives kernel panic at boot (git version) (
regards,
dgr48
Offline
only two people have a kernel panic I'm trying to know why...dgr48 can you contact me via jabber?
Finit-arc-git reupdated
Last edited by adriano (2009-05-10 17:03:24)
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
You say UUIDs don't work. What about other /dev/disk/foo/bar devices? Like by-id names, or labels? Using "normal" /dev/sd* names can be annoying when I add a hard drive and Linux randomly decides it's /dev/sda now. Especially since I use a LUKS-encrypted swap from /etc/crypttab -- can I use /dev/disk/by-id/ there? Is this a planned feature, and if so, will it slow the init down? Or is it intrinsic to the method used to speed boot time?
Although I'm not adriano, I took a look at the code and the way it works is
that udev is not used at all to create the device nodes; the device nodes are
created manually from the finit-arc program, and to figure out which to
create, finit-arc parses /etc/fstab and looks for lines that begin '/dev/sd'
- and these nodes are created. That's one of the things that make so
blazingly fast - it doesn't wait for udev which looks at all the devices, and
waits for them to initialize.
One solution to this problem would be of course to run udev first... but that means a slowdown by 1.3-3 seconds (on my laptops), so *if* adriano would decide to go this way, I would vote for this to be an optional feature.
On the other hand, this was originally meant for netbooks, which rarely have more than one hard drive attached.
Offline
Hi bender. I tried your solution but it is not the solution...I tried it in a few days ago but if I start udevadm first, it doesn't create all devices, it stopped I don't know why. Then not all devices was created, for examples usb_xxx device or others...it works fine only if start it after..
EDIT:
Finit-arc-GIT newly UPDATE
- I tried to fix clock, it works for me
- Swap partition mount added
- Autostart HAL, DBUS for Xorg hotplug support
Last edited by adriano (2009-05-10 17:50:36)
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
I'm getting those annoying segfaults at boot with the -git versions (20090509 and 20090510) on my core2duo laptop. Interestingly, the -git versions work ok on my desktop P3.
I had to revert to 0.2-3 for my laptop.
Offline
I'm working on this error. Try 20090511 now
Last edited by adriano (2009-05-10 21:32:20)
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
Finit-arc-GIT newly UPDATE
- I tried to fix clock, it works for me
- Swap partition mount added
- Autostart HAL, DBUS for Xorg hotplug support
Clock WORKS. Great job!
Swap doesn't work, though.
Cthulhu For President!
Offline
I'm getting those annoying segfaults at boot with the -git versions (20090509 and 20090510) on my core2duo laptop. Interestingly, the -git versions work ok on my desktop P3.
I had to revert to 0.2-3 for my laptop.
It was segfaulting on me as well at some point; what I did is running 'strace finit-arc' on a running system. That can tell you at which point it segfaults. Of course it might not be correct, since you're not supposed to run finit-arc on a normally running system, but it did help me to discover the problem (which turned out to be my own patch
Offline
Times measured with Bootchart
Lightweigth almost stock Arch64 on Core2duo
No finit-arc: 18s (untill agetty, too lazy to configure boot into X)
finit-arc: 10s (untill X + Awesome)
after backgrounding deamons in rc.conf and minimizing mkinitcpio till bare minimum:
finit-arc: just below 10s (untill X + Awesome)
finit-arc-git 20090511: 12s (untill X + Awesome)
I've wrote http://wiki.archlinux.org/index.php/Twe … _boot_time which becomes pretty useless after finit-arc. If finit-arc becomes more stable, I'll try to make a wikipage about it too, if no one already did by than.
Offline
I'm working on this error. Try 20090511 now
This one works. Good job!
Offline
Finit-arc-git updated
try latest git now it may be perfect...
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
- Autostart HAL, DBUS for Xorg hotplug support
Does this mean I can remove hal from my rc.conf ?
Offline
yesterday yes, now no
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline
as I said before, finit-arc-git slows down the boot process by 2s here, any ideas?
Offline
Latest git doesn't load something ( keybord and mouse doesn't work) 20090511 (first one) worked just fine
.::. TigTex @ Portugal .::.
Offline
Great job. I had trouble initially with finit-arc-git (20090511), in which mouse and keyboard would not work, but it works not with 20090511-1 (just the opposite of the previous post...).
- Networks's OK.
- boot into console in 9 sec (was 16), X in 18sec (total) (was 30+). Impressive.
- swap doesn't mount. swapon -a solves it.
- Another small issue: vpnc does not work: error: "vpnc: can't initialise tunnel interface: Inappropriate ioctl for device"
Archer since 03/2009 - AUR packages
Offline
- Networks's OK.
- swap doesn't mount. swapon -a solves it.
same problem here, swapon -a helps.
oh, latest git-version here.
Offline
There is a problem of access rights with latest git:
KDE can't start for it can't write to /tmp/.ICE-unix. Indeed I have drwx-----T instead of drwxrwxrwt. Everything is fine for /tmp/.X11-unix though.
Offline
Great job. I had trouble initially with finit-arc-git (20090511), in which mouse and keyboard would not work, but it works not with 20090511-1 (just the opposite of the previous post...).
Same here, 50090512 works, though.
Cthulhu For President!
Offline
Hi, probably I will leave finit-arc becouse I'm working at a new init project. Finit-arc is very fast but very incompatible too...the project is to modify the default init system and the goal is the full compatibility with Archlinux. I will open a new post for this.
Latest bootchart with the new init system:
It is not optmized and it contains ALL parts of rc.sysinit
Italian ArchLinux User
My projects on my homepage: www.obliquo.net
Offline