[opensuse-factory] help me out: what are the TOP features for 12.3?
Hey all, After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release! Cheers, Jos
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone! - GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science, as well as the python, ruby, perl programming languages Nothing of breaking importance (within the new pkgs), business chugging along as usual. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 22.02.2013 12:31, schrieb Jan Engelhardt:
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone!
- GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science, as well as the python, ruby, perl programming languages
Nothing of breaking importance (within the new pkgs), business chugging along as usual.
Err how about OpenStack Folsom? -- With kind regards, Sascha Peilicke SUSE Linux GmbH, Maxfeldstr. 5, D-90409 Nuernberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer HRB 16746 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 22 February 2013 12:38:20 Sascha Peilicke wrote:
Am 22.02.2013 12:31, schrieb Jan Engelhardt:
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone!
- GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science,
as well as the python, ruby, perl programming languages
Nothing of breaking importance (within the new pkgs), business chugging along as usual.
Err how about OpenStack Folsom?
Yup, good one. This one I already had, but keep em coming ;-)
On Friday 22 February 2013 12:31:48 Jan Engelhardt wrote:
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone!
- GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science, as well as the python, ruby, perl programming languages
If you (or somebody else) could write something more concrete about either (or both) of those, I can probably use that... ;-)
Nothing of breaking importance (within the new pkgs), business chugging along as usual.
On Friday 2013-02-22 13:24, Jos Poortvliet wrote:
On Friday 22 February 2013 12:31:48 Jan Engelhardt wrote:
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone!
- GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science, as well as the python, ruby, perl programming languages
If you (or somebody else) could write something more concrete about either (or both) of those, I can probably use that... ;-)
Anthy, and input method editor for the Japanese language, has had its last tarball release at the start of 2009. It is being supplanted by mozc, developed by and with Google. From time to time, Anthy would eat characters during output of Kanji; I have not seen this happening with mozc. Anthy had its own set of keybindings, whereas mozc uses the same (default) bindings as Windows JAIME, so OS convertists have it really easy. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 22 February 2013 15:45:01 Jan Engelhardt wrote:
On Friday 2013-02-22 13:24, Jos Poortvliet wrote:
On Friday 22 February 2013 12:31:48 Jan Engelhardt wrote:
On Friday 2013-02-22 12:18, Jos Poortvliet wrote:
Hey all,
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
Almost 700 new (source) packages. There ought to be something it for someone!
- GHC Haskell - mozc IME - some utilities for NFC hardware - a bunch of devel packages for the topic of science,
as well as the python, ruby, perl programming languages
If you (or somebody else) could write something more concrete about either (or both) of those, I can probably use that... ;-)
Anthy, and input method editor for the Japanese language, has had its last tarball release at the start of 2009. It is being supplanted by mozc, developed by and with Google. From time to time, Anthy would eat characters during output of Kanji; I have not seen this happening with mozc. Anthy had its own set of keybindings, whereas mozc uses the same (default) bindings as Windows JAIME, so OS convertists have it really easy. So oS just became better for our Japanese userbase... Good, thanks.
Jos Poortvliet wrote:
After assembling the Features list, we need a list of top features for openSUSE 12.3 now. I'm having a little trouble compiling this, so I'd like to ask all you to tell me about the features that excite you most in the upcoming release!
systemd now has fully replaced sysvinit. cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.de/ SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 22 of February 2013 13:43EN, Ludwig Nussel wrote:
Jos Poortvliet wrote:
After assembling the Features list, we need a list of top features for openSUSE 12.3 now.
systemd now has fully replaced sysvinit.
Taking an option away from users listed as a "top feature"? That would be really absurd... Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2013-02-22 19:48 (GMT+0100) Michal Kubeček composed:
Ludwig Nussel wrote:
Jos Poortvliet wrote:
After assembling the Features list, we need a list of top features for openSUSE 12.3 now.
systemd now has fully replaced sysvinit.
Taking an option away from users listed as a "top feature"? That would be really absurd...
Speaking of taking away features, 11.4 was the last release in which panning[1] worked, forcing me to stick to Evergreen for daily use. :~( [1] not to imply openSUSE specific: https://bugs.freedesktop.org/show_bug.cgi?id=39949 https://bugzilla.novell.com/show_bug.cgi?id=771521 https://bugzilla.redhat.com/show_bug.cgi?id=710191 -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 2013-02-22 19:59, Felix Miata wrote:
Speaking of taking away features, 11.4 was the last release in which panning[1] worked, forcing me to stick to Evergreen for daily use. :~(
[1] not to imply openSUSE specific: https://bugs.freedesktop.org/show_bug.cgi?id=39949
Humm I have working panning in a openSUSE 12.1 system around here. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2013-02-22 21:38 (GMT+0100) Jan Engelhardt composed:
Felix Miata wrote:
Speaking of taking away features, 11.4 was the last release in which panning[1] worked, forcing me to stick to Evergreen for daily use. :~(
[1] not to imply openSUSE specific: https://bugs.freedesktop.org/show_bug.cgi?id=39949
Humm I have working panning in a openSUSE 12.1 system around here.
I was thinking in terms of supported releases. AIUI, Evergreen will stay 11.4 after 12.1 support ends in about 3 months. https://bugzilla.novell.com/show_bug.cgi?id=771521 was filed against 12.2 and remains unfixed in 12.3. BS has patched Xorg to include it for 12.2, but I haven't seen same for 12.3 yet. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 2013-02-22 21:53, Felix Miata wrote:
On 2013-02-22 21:38 (GMT+0100) Jan Engelhardt composed:
Felix Miata wrote:
Speaking of taking away features, 11.4 was the last release in which panning[1] worked, forcing me to stick to Evergreen for daily use. :~(
[1] not to imply openSUSE specific: https://bugs.freedesktop.org/show_bug.cgi?id=39949
Humm I have working panning in a openSUSE 12.1 system around here.
I was thinking in terms of supported releases. AIUI, Evergreen will stay 11.4 after 12.1 support ends in about 3 months.
It also works in 12.2, and I suspect it also works in 12.3, though I have not tested that yet because NVIDIA has not updated their drivers (drumrol...) for Xorg ABI 13.1. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2013-02-23 12:11 (GMT+0100) Jan Engelhardt composed:
On Friday 2013-02-22 21:53, Felix Miata wrote:
On 2013-02-22 21:38 (GMT+0100) Jan Engelhardt composed:
Felix Miata wrote:
Speaking of taking away features, 11.4 was the last release in which panning[1] worked, forcing me to stick to Evergreen for daily use. :~(
[1] not to imply openSUSE specific: https://bugs.freedesktop.org/show_bug.cgi?id=39949
Humm I have working panning in a openSUSE 12.1 system around here.
It also works in 12.2, and I suspect it also works in 12.3, though I have not tested that yet because NVIDIA has not updated their drivers (drumrol...) for Xorg ABI 13.1.
I wouldn't know about what's possible using non-free drivers because I never ever use them. I have more than a dozen 12.2 and 12.3 systems using various Intel, NVidia and ATI chips, and on none of them does panning work using only the default repos and FOSS drivers, which is probably why all three bugs I referenced remain open with no comments from anyone claiming otherwise. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
It also works in 12.2, and I suspect it also works in 12.3, though I have not tested that yet because NVIDIA has not updated their drivers (drumrol...) for Xorg ABI 13.1.
You can easily compile them yourself: osc co X11:Drivers:Video nvidia-gfxG02 and then follow the README. Regards, Christian Boltz --
folgender Spam ist hier eingetroffen. Bitte ergreifen Sie gegen Ihren Kunden geeignete Maßnahmen. _____ Wir haben den User erschossen. ___|[] | abuse@germany.net \__|______| Wir machen den Weg frei /-(o_o_o_o)
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Saturday 2013-02-23 13:35, Christian Boltz wrote:
Hello,
Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
It also works in 12.2, and I suspect it also works in 12.3, though I have not tested that yet because NVIDIA has not updated their drivers (drumrol...) for Xorg ABI 13.1.
You can easily compile them yourself:
osc co X11:Drivers:Video nvidia-gfxG02
That would only build the kernel module, not nvidia_drv.so. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
On Saturday 2013-02-23 13:35, Christian Boltz wrote:
Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
It also works in 12.2, and I suspect it also works in 12.3, though I have not tested that yet because NVIDIA has not updated their drivers (drumrol...) for Xorg ABI 13.1.
You can easily compile them yourself: osc co X11:Drivers:Video nvidia-gfxG02
That would only build the kernel module, not nvidia_drv.so.
Are you sure? I have two good arguments to confirm everything is working: 1. I'm using those packages myself, and have a working nvidia driver ;-) 2. # rpm -qf /usr/lib64/xorg/modules/updates/drivers/nvidia_drv.so x11-video-nvidiaG02-304.64-0.x86_64 # rpm -qi x11-video-nvidiaG02-304.64-0.x86_64 | tail -n1 Distribution: X11:Drivers:Video / openSUSE_Factory Regards, Christian Boltz -- So does that mean you're volunteering, Pascal? No, wait - I can see the headline now: "Guru Site Down Due to Maintainer Working 24/7 on Newsletter - Millions in Agony!" [Kai Ponte in opensuse-project] -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Saturday 2013-02-23 17:11, Christian Boltz wrote:
That would only build the kernel module, not nvidia_drv.so.
Are you sure?
Pret-ty sure. $ sosc build -bd openSUSE_12.2 x86_64 2>&1|tee bl [...] [ 173s] ares07 finished "build nvidia-gfxG02.spec" at Sat Feb 23 16:55:17 UTC 2013. [...] $ grep nvidia_drv bl $ So it did not build nvidia_drv.so at all (at most it shipped some prebuilt crap), which is not surprising, given NVIDIA does not give out any sources for that. You really don't want to argue with me on that ;-)
1. I'm using those packages myself, and have a working nvidia driver ;-)
The package is useless to me because I use NV 96.x. Guess I could try NV 173 again which has Xorg 1.13, but last time I did, I was not quite satisfied with the picture. (Something _was_ different.) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Jan Engelhardt
On Saturday 2013-02-23 17:11, Christian Boltz wrote:
That would only build the kernel module, not nvidia_drv.so.
Are you sure?
Pret-ty sure.
$ sosc build -bd openSUSE_12.2 x86_64 2>&1|tee bl [...] [ 173s] ares07 finished "build nvidia-gfxG02.spec" at Sat Feb 23 16:55:17 UTC 2013. [...] $ grep nvidia_drv bl $
So it did not build nvidia_drv.so at all (at most it shipped some prebuilt crap), which is not surprising, given NVIDIA does not give out any sources for that. You really don't want to argue with me on that ;-)
1. I'm using those packages myself, and have a working nvidia driver ;-)
The package is useless to me because I use NV 96.x. Guess I could try NV 173 again which has Xorg 1.13, but last time I did, I was not quite satisfied with the picture. (Something _was_ different.)
The packages from nvidia provide: /usr/lib64/xorg/modules/drivers/nvidia_drv.so This particular one from 313.18 -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
On Saturday 2013-02-23 17:11, Christian Boltz wrote:
That would only build the kernel module, not nvidia_drv.so.
Are you sure?
Pret-ty sure.
We'll see ;-)
$ sosc build -bd openSUSE_12.2 x86_64 2>&1|tee bl [...] [ 173s] ares07 finished "build nvidia-gfxG02.spec" at Sat Feb 23 16:55:17 UTC 2013. [...] $ grep nvidia_drv bl $
So it did not build nvidia_drv.so at all (at most it shipped some prebuilt crap), which is not surprising, given NVIDIA does not give out any sources for that.
This confirms the kernel module build didn't build the module for x11. Not too surprising ;-) There's a reason why there's another spec in the OBS that builds the X11 module you are looking for. At least it does for me ;-) Did you try to osc build x11-video-nvidiaG02.spec ? The resulting package x11-video-nvidiaG02 should contain nvidia_drv.so. (Note: in this context, with "build" I mean "it does something to ensure all needed files are in the RPM afterwards". I don't know if there's always a compiler involved or if it just copies files around.)
You really don't want to argue with me on that ;-)
Wrong guess ;-)
1. I'm using those packages myself, and have a working nvidia driver ;-) The package is useless to me because I use NV 96.x. Guess I could try NV 173 again which has Xorg 1.13, but last time I did, I was not quite satisfied with the picture. (Something _was_ different.)
In the meantime, we are at version 304. However I can't guarantee for the picture quality ;-) Regards, Christian Boltz -- Naja, nicht alles von IBM ist gut. Auch Microsoft hat außer Mäusen ja noch nichts bemerkenswertes produziert. [Olaf Kaluza in d.a.f.c] -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sat, Feb 23, 2013 at 09:31:41PM +0100, Christian Boltz wrote:
Hello,
Am Samstag, 23. Februar 2013 schrieb Jan Engelhardt:
On Saturday 2013-02-23 17:11, Christian Boltz wrote:
That would only build the kernel module, not nvidia_drv.so.
Are you sure?
Pret-ty sure.
We'll see ;-)
$ sosc build -bd openSUSE_12.2 x86_64 2>&1|tee bl [...] [ 173s] ares07 finished "build nvidia-gfxG02.spec" at Sat Feb 23 16:55:17 UTC 2013. [...] $ grep nvidia_drv bl $
So it did not build nvidia_drv.so at all (at most it shipped some prebuilt crap), which is not surprising, given NVIDIA does not give out any sources for that.
This confirms the kernel module build didn't build the module for x11. Not too surprising ;-)
There's a reason why there's another spec in the OBS that builds the X11 module you are looking for. At least it does for me ;-)
Did you try to osc build x11-video-nvidiaG02.spec ? The resulting package x11-video-nvidiaG02 should contain nvidia_drv.so.
(Note: in this context, with "build" I mean "it does something to ensure all needed files are in the RPM afterwards". I don't know if there's always a compiler involved or if it just copies files around.)
These days the driver builds during RPM installation on your system I think. Ciao, Marcus -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Saturday 2013-02-23 21:31, Christian Boltz wrote:
So it did not build nvidia_drv.so at all (at most it shipped some prebuilt crap), which is not surprising, given NVIDIA does not give out any sources for that.
There's a reason why there's another spec in the OBS that builds the X11 module you are looking for. At least it does for me ;-)
Funny you. %build # nothing
(Note: in this context, with "build" I mean "it does something to ensure all needed files are in the RPM afterwards". I don't know if there's always a compiler involved or if it just copies files around.)
I think I was explicit enough about building (or the lack thereof) nvidia_drv.so, and not shuffling the file from a horrible .run to a .rpm under the umbrealla of building an RPM.
The package is useless to me because I use NV 96.x. Guess I could try NV 173 again which has Xorg 1.13, but last time I did, I was not quite satisfied with the picture. (Something _was_ different.)
In the meantime, we are at version 304. However I can't guarantee for the picture quality ;-)
304 does not support NV34. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/22/2013 01:48 PM, Michal Kubeček wrote:
On Friday 22 of February 2013 13:43EN, Ludwig Nussel wrote:
Jos Poortvliet wrote:
After assembling the Features list, we need a list of top features for openSUSE 12.3 now.
systemd now has fully replaced sysvinit.
Taking an option away from users listed as a "top feature"? That would be really absurd...
Michal Kubeček
It's necessary to let people know upfront. Yes? -- Cheers! Roman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/22/2013 09:17 PM, Roman Bysh wrote:
On 02/22/2013 01:48 PM, Michal Kubeček wrote:
On Friday 22 of February 2013 13:43EN, Ludwig Nussel wrote:
Jos Poortvliet wrote:
After assembling the Features list, we need a list of top features for openSUSE 12.3 now.
systemd now has fully replaced sysvinit.
Taking an option away from users listed as a "top feature"? That would be really absurd...
Michal Kubeček
It's necessary to let people know upfront. Yes?
That it is a top feature that they have lost the freedom of choosing . Ok cool -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 22/02/13 17:21, Togan Muftuoglu escribió:
It's necessary to let people know upfront. Yes?
That it is a top feature that they have lost the freedom of choosing . Ok cool
Nobody has lost the freedom of choosing, they still have a few distributions without systemd.. but soon only BSD will be their choice. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, Feb 22, 2013 at 5:27 PM, Cristian Rodríguez
El 22/02/13 17:21, Togan Muftuoglu escribió:
It's necessary to let people know upfront. Yes?
That it is a top feature that they have lost the freedom of choosing . Ok cool
Nobody has lost the freedom of choosing, they still have a few distributions without systemd.. but soon only BSD will be their choice.
Agree, it needs to be said. Maybe don't call it a feature, maybe call it a "change". -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, 22 Feb 2013 17:36:46 -0300, Claudio Freire wrote:
On Fri, Feb 22, 2013 at 5:27 PM, Cristian Rodríguez
wrote: El 22/02/13 17:21, Togan Muftuoglu escribió:
It's necessary to let people know upfront. Yes?
That it is a top feature that they have lost the freedom of choosing . Ok cool
Nobody has lost the freedom of choosing, they still have a few distributions without systemd.. but soon only BSD will be their choice.
Agree, it needs to be said. Maybe don't call it a feature, maybe call it a "change".
"Enhanced system boot performance through the streamlining of the initialization system to a single tool (systemd), which allows the development team to focus their efforts on the upstream tool that is receiving the most attention." It gets ridiculous trying to spin stuff with marketing-speak. I once had a manager who insisted there were "no problems, only opportunities". I think systemd is an improvement overall, but I understand that there will always be those who resist change or who seek out (or are "fortunate" enough to find) the problems that come with any change. Heck, I'm usually one who ends up finding new and creative ways to break stuff that nobody ever dreamed of. :) The change does need to be called out, because it's important that experienced prospective users know what they're going to be using as it is a significant change. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 23/02/13 21:00, Linda Walsh escribió:
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
As usual, you are wrong. systemd has nothing to do with LVM. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sun, 24 Feb 2013 02:37, Cristian Rodríguez
El 23/02/13 21:00, Linda Walsh escribió:
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
As usual, you are wrong. systemd has nothing to do with LVM.
Please, tell us: Just WHAT is responsible for the ordering of the services? Either systemd defaults, or the openSUSE team that did the migration from (working in correct order) SystemV-Init to systemd? Eitherway, please, get that RIGHT before Goldmaster, in my eyes that is a absolute show-stopper for any LVM using installation. I'd call this a TOP-MISS-feature. -- Yamaban.
Cristian Rodríguez wrote:
El 23/02/13 21:00, Linda Walsh escribió:
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
As usual, you are wrong. systemd has nothing to do with LVM.
Um... I didn't say it DID, I said it _needed_ to. The fact that it doesn't start lvm before scanning for local disks IS the problem I am speaking of. It would explain why my lvm-based local disks are NOT mounted in time for local services to come up. Under sysVinit, I see a boot order (before it goes into the rc1 or rc3 script) of: S01boot.sysctl, S01boot.udev S01boot.usr-mount S03boot.rootfsck S04boot.clock S05boot.device-mapper, S05boot.loadmodules S06boot.cgroup, S06boot.lvm S09boot.localfs S11boot.cleanup, S11boot.klog, S11boot.localnet, S11boot.swap, S11setserial S12boot.isapnp, S12boot.ldconfig, S12boot.sysstat ------------------ systemV knew enough to load lvm in cycle 6, and the localfs in cycle 9. You are saying that systemd knows nothing about LVM? It seems to have scripts for everything above except lvm and pnp. Could you elaborate on how systemd can't do LVM? Is it no longer supported? I can see it running the lvm and lvm-monitor scripts before it begins run level 3 -- you can tell the difference in the log when systemd kicks in vs. not -- but it doesn't run them early in boot. It runs rootfsck early in boot, udev sysctl, Why doesn't it run lvm volume detection right after devicemapper and loadmodules like sysV does? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 23/02/13 23:37, Linda Walsh escribió:
You are saying that systemd knows nothing about LVM?
Exactly.
It seems to have scripts for everything above except lvm and pnp. Could you elaborate on how systemd can't do LVM?
In this release cycle there is no lvm generator (http://www.freedesktop.org/wiki/Software/systemd/Generators) to interface with systemd. All is done in a combination between the initrd, udev and boot.lvm legacy init script, this is all part of packages lvm2 and device-mapper.
Is it no longer supported?
Where did you got that idea ? of course it should work.
I can see it running the lvm and lvm-monitor scripts before it begins run level 3 -- you can tell the difference in the log when systemd kicks in vs. not -- but it doesn't run them early in boot.
It runs rootfsck early in boot, udev sysctl, Why doesn't it run lvm volume detection right after devicemapper and loadmodules like sysV does?
Hard to tell if we dont know exactly what are you doing or expecting, what bug report exactly ? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/23/2013 08:37 PM, Cristian Rodríguez pecked at the keyboard and wrote:
El 23/02/13 21:00, Linda Walsh escribió:
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
As usual, you are wrong. systemd has nothing to do with LVM.
As usual, you don't have a clue Cristian. systemd IS responsible for starting system services of which the mounting of partitions is one of them. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 23/02/13 23:59, Ken Schneider - openSUSE escribió:
As usual, you don't have a clue Cristian. systemd IS responsible for starting system services of which the mounting of partitions is one of them.
Systemd does what is told by the relevant packages. lvm2, device-mapper.. the scripts and udev rules they contain. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 24.02.2013 03:59, schrieb Ken Schneider - openSUSE:
As usual, you are wrong. systemd has nothing to do with LVM.
As usual, you don't have a clue Cristian. systemd IS responsible for starting system services of which the mounting of partitions is one of them.
Cristian and Ken, can you please stop that rude tone? You're free to disagree, but not in the way you're discussing here. Greetings, Stephan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Stephan Kulow wrote:
As usual, you are wrong. systemd has nothing to do with LVM.
As usual, you don't have a clue Cristian. systemd IS responsible for starting system services of which the mounting of partitions is one of them.
Cristian and Ken, can you please stop that rude tone? You're free to disagree, but not in the way you're discussing here.
I get that tone often with Christian -- I think Ken was trying to speak to him in a way he might better understand, like speaking a native tongue ...? ;-) you should check out Christian's responses to bug reports sometime...they are...at least as 'pleasant'... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sat, 23 Feb 2013 16:00:13 -0800, Linda Walsh wrote:
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
Have you reported the bug or added to it? I haven't run into that problem myself. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Jim Henderson wrote:
Have you reported the bug or added to it? I haven't run into that problem myself. Jim
----
Haven't reported it as a bug because didn't want to just have it closed out
as 'not a supported use case' like several of my other recent bugs where
my system has gone from working very fast to 'slow' w/systemd. (it's slower
than the scripts by 30%).
I may be reading the helpful advice Christian gave, incorrectly, BUT it it
*sounds* l
ike there is an expectation that LVM will be done for systemd by prepscripts
that are hidden away on initrd just like the broken fix for having /usr mounted at
boot time.
Please note. I DID ask, *why* is it that systemd can't mount /usr and (and
now, i find out,
LVM). It sounds like systemd isn't a *boot* solution, but a service manager
that takes over
from some initrd that does the real booting with standard shell scripts.
Why SUSE insists on hiding this in initrd is a bit peculiar. It's not hard to do it
and leave it on the hard disk and make initrd for modules only.
I was given no help and told it was not possible for systemd to mount /usr within
the first few steps. That they couldn't do it, and I was welcome to submit patches.
Really?
Like these changes?
Now tell me why this wouldn't work in the field with the right modules loaded
off the ramdisk?
====================================
/etc/systemd/system> ls default.target.wants
systemd-readahead-collect.service@ systemd-readahead-replay.service@
usr-mount.service@
^^^^^^^^^^^^^^^^^
/lib/systemd> ls system/*usr*
system/usr-mount.service system/usr.mount@
/lib/systemd> cat system/usr.mount
####################################################################
# This file is linda's
#
# systemd is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
[Unit]
Description=mount /usr on root early in systemd startup
DefaultDependencies=no
Conflicts=shutdown.target
Before=sysinit.target shutdown.target
ConditionPathExists=/bin/mount
[Service]
Type=oneshot
ExecStart=/lib/systemd/systemd-usr
RemainAfterExit=no
StandardOutput=syslog+console
TimeoutSec=5
[Install]
WantedBy=default.target
#--------------------------------------------------------------------
/lib/systemd> mroe systemd-usr
#!/bin/bash
export PATH=/bin:/sbin
function datetime {
DSTRING="%b %d %H:%M:%S";
date +"$DSTRING" ${1:+-d $1}
}
function _hostname {
local hp=/proc/sys/kernel/hostname
if [[ -e $hp ]]; then hostname=$(<$hp);
else hostname=localhost;fi
echo "$hostname"
}
{
declare -i status=-1
declare -i retries=3
declare -i wait=4
while ((status!=0 && retries>0)) ; do
if /bin/mount --no-canonicalize /usr ; then
echo "$(datetime) $(_hostname) /usr mounted pass"
status=0
else
status=$?
echo "$(datetime) $(_Hostname) FAILED: /usr/mount FAIL"
fi
retries+=-1
if ((status!=0)); then
echo "Sleeping for $wait "
sleep 4
fi
done
} | (
if [[ -e /bin/tee ]]; then
tee /tmp/bootlog-helper || cat
else
while read ln;do
echo "$ln" >&2
echo "$ln" >/tmp/echolog-helper
done
fi
)
--------------------------------------------------------------------------
The above are simplistic and should be straightforward to modify.
They are geared at a lowest-common-denominator of my system....
It took about an hour or two to figure out what systemd wanted -- it mounts in
parallel
with the modules...before rootfsck
В Sat, 23 Feb 2013 21:42:17 -0800
Linda Walsh
service boot.localfs start <----***here is localfs starting
Under systemd localfs is not even started. bor@opensuse:~> systemctl status localfs.service localfs.service - Shadow /etc/init.d/boot.localfs Loaded: loaded (/lib/systemd/system/localfs.service; static) Active: inactive (dead) CGroup: name=systemd:/system/localfs.service because it is not pulled in by anything. And even if it had been, it is noop: bor@opensuse:~> grep ExecStart /lib/systemd/system/localfs.service ExecStart=/bin/true -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sat, 23 Feb 2013 21:42:17 -0800, Linda Walsh wrote:
Haven't reported it as a bug because didn't want to just have it closed out as 'not a supported use case' like several of my other recent bugs where my system has gone from working very fast to 'slow' w/systemd. (it's slower than the scripts by 30%).
I think closing it as "not a supported use case" is a decision to be made by the developers, not by the reporter. I understand what you're saying, but at the same time it does nobody any good if you don't report it as a bug. If it does get closed, then when others indicate they have the same issue, the issue can be opened again. And if it isn't, then maybe there's some information that the developers can provide as to *why* it's not a supported use case. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
В Sat, 23 Feb 2013 16:00:13 -0800
Linda Walsh
Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
My filesystems in 12.2 are on LVM. I can boot. Several test installations of 12.3 I did were on LVM. I could boot. What did I do wrong? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/24/2013 02:24 AM, Andrey Borzenkov wrote:
В Sat, 23 Feb 2013 16:00:13 -0800 Linda Walsh
пишет: Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
My filesystems in 12.2 are on LVM. I can boot. Several test installations of 12.3 I did were on LVM. I could boot. What did I do wrong?
Probably there is more people here using LVM/cryptsetup and can share experiences to see if there is anything wrong, I am not using it since ages, will do again when BTRFS replaces all this lvm thing by something sane :-) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
В Sun, 24 Feb 2013 02:51:02 -0300
Cristian Rodríguez
On 02/24/2013 02:24 AM, Andrey Borzenkov wrote:
В Sat, 23 Feb 2013 16:00:13 -0800 Linda Walsh
пишет: Jim Henderson wrote:
I think systemd is an improvement overall, but I understand that there will always be those .. who are > "fortunate" enough to find) the problems that come with any change.
Um... like having it mount local filesystems and declaring that runlevel 3 has been reached before LVM has finished finding and mounting local volumes?
My filesystems in 12.2 are on LVM. I can boot. Several test installations of 12.3 I did were on LVM. I could boot. What did I do wrong?
Probably there is more people here using LVM/cryptsetup and can share experiences to see if there is anything wrong,
One possibility is that lvm startup does not wait on udev-settle: bor@opensuse:~> systemctl --no-pager show lvm.service -p After After=udev.service device-mapper.service multipath.service md.service dmraid.service udev-trigger.service systemd-journald.socket fsck-root.service So as long as LVM is not included in initrd or not all devices are found there it's possible that LVM will miss something. But to confirm this (or make any progress) bug report with diagnostic information is required. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Andrey Borzenkov wrote:
bor@opensuse:~> systemctl --no-pager show lvm.service -p After After=udev.service device-mapper.service multipath.service md.service dmraid.service udev-trigger.service systemd-journald.socket fsck-root.service
So as long as LVM is not included in initrd or not all devices are found there it's possible that LVM will miss something.
But to confirm this (or make any progress) bug report with diagnostic information is required.
That is what is happening... What diagnostic information is wanted? Here's a bugid: https://bugzilla.novell.com/show_bug.cgi?id=805420 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Sun, Feb 24, 2013 at 1:14 PM, Linda Walsh
Andrey Borzenkov wrote:
bor@opensuse:~> systemctl --no-pager show lvm.service -p After After=udev.service device-mapper.service multipath.service md.service dmraid.service udev-trigger.service systemd-journald.socket fsck-root.service
So as long as LVM is not included in initrd or not all devices are found there it's possible that LVM will miss something.
But to confirm this (or make any progress) bug report with diagnostic information is required.
That is what is happening... What diagnostic information is wanted?
a) which openSUSE version you are really running b) whether you are using sysvinit or systemd for a start.
Here's a bugid:
this bug is filed against 12.3 RC1 and complaints about systemd. From provided log snippet it is obvious that you are using sysvinit and services shown in this log do not even exist in 12.3 RC1. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 25/02/13 05:21, Andrey Borzenkov escribió:
this bug is filed against 12.3 RC1 and complaints about systemd. From provided log snippet it is obvious that you are using sysvinit and services shown in this log do not even exist in 12.3 RC1.
That bug report does not make any sense, whatsoever and at the end.. "Note -- My system boots off of it's hard disk -- not a ram disk. Needed drivers are built-in. Additional note -- in order for /usr to be mounted at boot, 'mount' and any needed libraries need to reside in /bin and not be moved." So it is esentially INVALID for the following reasons: 1, filled against 12.3rc1, but running sysvinit (!!?) 2. complains about systemd but it is not the service manager in use. 3. running in a distro that is NOT 12.3rcx. probably 12.1 or earlier. 4. Running without an initramfs/initrd, that wont work and is totally unsupported. If someone fills a proper valid bug report will be cool :-| -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Andrey Borzenkov wrote:
On Sun, Feb 24, 2013 at 1:14 PM, Linda Walsh
wrote: Andrey Borzenkov wrote:
bor@opensuse:~> systemctl --no-pager show lvm.service -p After After=udev.service device-mapper.service multipath.service md.service dmraid.service udev-trigger.service systemd-journald.socket fsck-root.service
So as long as LVM is not included in initrd or not all devices are found there it's possible that LVM will miss something.
But to confirm this (or make any progress) bug report with diagnostic information is required.
That is what is happening... What diagnostic information is wanted?
a) which openSUSE version you are really running b) whether you are using sysvinit or systemd
for a start.
Here's a bugid:
this bug is filed against 12.3 RC1 and complaints about systemd. From provided log snippet it is obvious that you are using sysvinit and services shown in this log do not even exist in 12.3 RC1.
As stated in the the followup, the message when I tried to enable boot.local was that it was shadowed/controlled by systemd. I didn't create the message. If the message is wrong, file a bug report against whoever wrote the message. 2nd, if systemV was controlling the boot process, it would have followed the normal boot order dependencies. It's never had a problem doing so. It opens the scripts in alphanumeric sorted order -- its not rocket science. As you point out -- systemd isn't able to do dependencies for localfs -- they point to /dev/null, so basically it starts the localfs "whenever" in 12.3 and 12.1. This is completely uncoordinated with starting lvm. On an lvm with multiple snapshots going at the same time, it can take well in excess of a minute for it to finish and I've had systemV scripts timeout on it as well -- but in this case the timeout was <10 seconds. ------------------------------------------------------------------------- Cristian Rodr wrote:
That bug report does not make any sense, whatsoever and at the end..
"Note -- My system boots off of it's hard disk -- not a ram disk. Needed drivers are built-in.
Additional note -- in order for /usr to be mounted at boot, 'mount' and any needed libraries need to reside in /bin and not be moved."
So it is esentially INVALID for the following reasons:
1, filled against 12.3rc1, but running sysvinit (!!?)
--- not according to the suse software messages on the system.
2. complains about systemd but it is not the service manager in use.
That's not what chkconfig says.
3. running in a distro that is NOT 12.3rcx. probably 12.1 or earlier.
Already been told systemd can't coordinate boot processes -- you have to run initrd to run systemV initrd to boot the HW -- that's why you need initrd.
4. Running without an initramfs/initrd, that wont work and is totally unsupported.
Right -- because you need to run the SystemV style startup scripts to start the HW, before you launch systemd since for whatever reason, you haven't put the smarts in systemd to handle basics like lvm, dmeventd, localfs...etc. All that has to be done before systemd starts because the all inclusive new new boot manager can't handle it? Why did you jump to something that can't handle dependencies? As for you saying initrd-less systems are unsupported, -- I was told that you were unable to do it in systemd, yet I provided the patches to load /usr in systemd at the start. They weren't rocket science. So why the refusal on SuSE's part to support legacy systems? It certainly not because they "can't". -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wed, Feb 27, 2013 at 9:28 AM, Linda Walsh
Andrey Borzenkov wrote: 2nd, if systemV was controlling the boot process, it would have followed the normal boot order dependencies. It's never had a problem doing so.
It is possible, but your log shows INIT: entering run-level 3 I am not aware of systemd ever producing this message. And you still did not say from which openSUSE version you did paste these messages. Could we start with cat /etc/SuSE-release cat /etc/fstab systemctl --all --type mount -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Andrey Borzenkov wrote:
On Wed, Feb 27, 2013 at 9:28 AM, Linda Walsh
wrote: Andrey Borzenkov wrote: 2nd, if systemV was controlling the boot process, it would have followed the normal boot order dependencies. It's never had a problem doing so.
It is possible, but your log shows
INIT: entering run-level 3
I am not aware of systemd ever producing this message.
And you still did not say from which openSUSE version you did paste these messages.
Could we start with
cat /etc/SuSE-release cat /etc/fstab systemctl --all --type mount
Sorry, I had updated the bug to reflect that systemd was from 12.1. Where I got that systemd was handling the 'boot', was from the output of chkconfig:
chkconfig boot.local on insserv: Note: sysvinit service boot.local is shadowed by systemd local.service, Forwarding request to '/bin/systemctl --root / enable local.service'. Operation failed: No such file or directory
Is that message incorrect or misleading? I.e. from what Christian says, systemd doesn't handle lvm, dmeventd, or mounting localfs (which points to null in 12.3 as it does in 12.1). Correct me if I am confused, but it appears that the message from chkconfig is misleading: while systemd "shadow"s the boot, it doesn't really handle it -- and that sysVinit-style shell scripts located on the ram disk actually handle all the stuff in the /etc/boot.d directory -- and that systemd doesn't actually get control until all (or most) of what is in /etc/boot.d has been processed on the ramdisk (initrd). If that's the case, why make a big deal of removing sysVinit -- i.e. why not just say it's role has been reduced, leave it to handle the functions that systemd cannot (or is not), that are located in /etc/boot.d, and then transfer control to systemd? I.e. if /etc/boot.d has to be handled before giving control over to systemd, why the mandate for removing sysVinit but forcing equivalent functionality into initrd? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 27/02/13 14:59, Linda Walsh escribió:
but forcing equivalent functionality into initrd?
Because mounting functionality-critical filesystems belong to the initrd and not elsewhere. Unless your idea is to make distribution maintenance a hell. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Cristian Rodríguez wrote:
El 27/02/13 14:59, Linda Walsh escribió:
but forcing equivalent functionality into initrd?
Because mounting functionality-critical filesystems belong to the initrd and not elsewhere. Unless your idea is to make distribution maintenance a hell.
Huh? My idea is how it has been for the past 40 years. Suddenly it is broken? Correct me if I am wrong, but before 12.1, The system used to start 'critical filesystem functions' from the root disk via the scripts in /etc/boot.d Then runlevel scripts (rc{1-5}) were called to start up the system. It appears that systemd is taking the place of rc{1-5}, which I don't have huge issues with. But it doesn't take the place of the /etc/boot.d scripts -- we've ben told it is, but it doesn't -- instead, they are all being moved onto initrd, where it makes user maintenance and control a nightmare. how does moving these file critical functions out of /etc/boot.d onto an initrd solve some "distribution maintenance hell? I.e. what distribution maintenance hell, is this solving, because it has worked fine for decades, and now it's "hell"?? Why? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 22 February 2013 20:21, Togan Muftuoglu
That it is a top feature that they have lost the freedom of choosing . Ok cool
I would argue that we should celebrate our distributions embracing of systemd as a positive In addition to all the usual positives bandied about systemd (faster boot times, on-demand service loading, etc), and the less obvious ones (eg. lid-close suspend to ram working from tty), focusing on a single init system is making our lives easier as distribution hackers, one stack to integrate with, one stack to polish, one stack to test - plus we still have all the backwards compatibility that systemd-sysvinit brings for our users who still choose to do things the old way. For our non-contributing users, I'm not sure it's a top feature, but I do think it's important for our contributors to know about, so they can pitch in making the most of everything we get in systemd and tying any loose ends left over from this transition. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Richard Brown
On 22 February 2013 20:21, Togan Muftuoglu
wrote: That it is a top feature that they have lost the freedom of choosing . Ok cool
I would argue that we should celebrate our distributions embracing of systemd as a positive In addition to all the usual positives bandied about systemd (faster boot times, on-demand service loading, etc), and the less obvious ones (eg. lid-close suspend to ram working from tty), focusing on a single init system is making our lives easier as distribution hackers, one stack to integrate with, one stack to polish, one stack to test - plus we still have all the backwards compatibility that systemd-sysvinit brings for our users who still choose to do things the old way.
It couldn't be further from the truth, it neither makes my life as a distribution hackers, to the contrary it causes needless work (as a case in point look at the ugly bandaids I had to add to xfce4-power-manager and lightdm) because of API churn, idiotic defaults, and systemd being a source of countless bugs and regression that had to be addressed, nor is it compatible to systemd-sysvinit. -- Guido Berhoerster -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El 22/02/13 18:18, Guido Berhoerster escribió:
It couldn't be further from the truth, it neither makes my life as a distribution hackers, to the contrary it causes needless work (as a case in point look at the ugly bandaids I had to add to xfce4-power-manager and lightdm) because of API churn, idiotic defaults, and systemd being a source of countless bugs and regression that had to be addressed, nor is it compatible to systemd-sysvinit.
KDE and gnome are doing just fine, are you sure it is not a problem of XFCE and lightdm instead of systemd ? or more precisely an issue generated by the deprecation of consolekit ? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, Feb 22, 2013 at 08:47:58PM +0000, Richard Brown wrote:
focusing on a single init system is making our lives easier as distribution hackers
It is making my life much harder. I've been using last few versions on my desktop since something like third milestone and by the time of RC1, I've been using them on all my desktops. With 12.3, I still didn't get out of a testing only virtual machine and I really don't see where I'll find enough time to make it work properly. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, Feb 22, 2013 at 03:17:32PM -0500, Roman Bysh wrote:
On 02/22/2013 01:48 PM, Michal Kubeček wrote:
On Friday 22 of February 2013 13:43EN, Ludwig Nussel wrote:
systemd now has fully replaced sysvinit.
Taking an option away from users listed as a "top feature"? That would be really absurd...
It's necessary to let people know upfront. Yes?
Definitely. It should be mentioned in Release Notes as users should be warned. I just don't like the abuse of language when taking a feature away is presented as a feature or even a top feature. We should be honest and present it the way it is: that some group of people decided to take this option away and leave systemd as the only option. We shouldn't pretend we are _giving_ users some feature when we are actually _taking_ a feature _away_. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (22)
-
Andrey Borzenkov
-
Christian Boltz
-
Claudio Freire
-
Cristian Rodríguez
-
Felix Miata
-
Guido Berhoerster
-
Jan Engelhardt
-
Jim Henderson
-
Jos Poortvliet
-
Ken Schneider - openSUSE
-
Linda Walsh
-
Ludwig Nussel
-
Marcus Meissner
-
Michal Kubecek
-
Michal Kubeček
-
Patrick Shanahan
-
Richard Brown
-
Roman Bysh
-
Sascha Peilicke
-
Stephan Kulow
-
Togan Muftuoglu
-
Yamaban