Hi,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
What are the various goals of the LiveCD project ?
- A way for people to see what hardware is supported by CentOS - with drivers included by upstream only - with drivers that can be added with little effort (to eg. accomodate eeepc and other purposes) - including a tool to send a hardware status report to a central database
- A tool for recovering a broken system - include only a few basic tools for debugging
- A tool for installing CentOS on a system
- A workable environment for people to use day by day - do we want persistent storage - does it need to be installable from Windows - does it include wireless firmware
- Provide the tools for people to make their own spins - all sorts of appliances, eg. DVR, rescue, ...
I am sure we can list more applicable use-cases for this project. And in the end we have to agree which fit the LiveCD project (eg. whether we make an official CentOS LiveCD and a seperate CentOS-plus LiveCD) and what the main priorities are.
Also, to stimulate the LiveCD development there should be no reason why we couldn't release a LiveCD 5.3.1 or LiveCD 5.3.2 if there are compelling reasons to make a new respin (and update the mirrors).
In the end I see a bright future for the LiveCD as the project where driver-development (dasha, elrepo), CentOS-based appliances, hardware testing, and much more come together. To me it is one of those unique opportunities to drive different other projects with a very positive attitude as it brings (even more) meaning to CentOS to a lot of (new) people.
Where do we want to go with this and what is realistically achievable ?
I am thrilled to find out :)
Dear Dag,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
I don't think that there is a need to divide from upstream atm. and am not really willed to break compatibility for features.
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
On the other hand we should make it easier for ppl to remaster the CentOS LiveCDs to allow creation of appliances and specialized LiveCDs.
We should then create some name/logo usage guidelines, to make sure that a rebuild is marked as 'un-official' or part of a CentOS project (e.g. a Xfce LiveCD, an Edu media or a installable one)
Best Regards Marcus
Marcus Moeller wrote:
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
On the other hand we should make it easier for ppl to remaster the CentOS LiveCDs to allow creation of appliances and specialized LiveCDs.
Yes, that would be great. I'd love to have a bootable CD that came up running backuppc so I could use it and a USB-connected copy of my backup archive to restore files immediately in a disaster scenario. I'm setting up something in a vmware image now, but booting straight into it would be even better.
On Fri, 1 May 2009, Marcus Moeller wrote:
Dear Dag,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
I don't think that there is a need to divide from upstream atm. and am not really willed to break compatibility for features.
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
Hi Dag, ...
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
There is nothing wrong in including (compatible) upstream resources, if it's useful and this is what Patrice is already doing afaik.
I have also noticed that there are also some basic remaster instructions available on the project page.
Best Regards Marcus
On 01/05/2009, Marcus Moeller mail@marcus-moeller.de wrote:
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Novice user: I'd like to see if a reliable, well respected Linux kernel based OS runs on my laptop / notebook.
Venerable wizard: Obtain copies of the various LiveCDs that are available and test.
Novice user: This works, that works, CentOS doesn't work ( ---> into the trash it goes), Ubuntu works, . . .
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
Alan.
on 5-1-2009 12:19 PM Alan Bartlett spake the following:
On 01/05/2009, Marcus Moeller mail-7BlZPJ8e1eab+SiqwsCprbNAH6kLmebB@public.gmane.org wrote:
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Novice user: I'd like to see if a reliable, well respected Linux kernel based OS runs on my laptop / notebook.
Venerable wizard: Obtain copies of the various LiveCDs that are available and test.
Novice user: This works, that works, CentOS doesn't work ( ---> into the trash it goes), Ubuntu works, . . .
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
Alan.
The opposite would also be true... Novice tries CentOS live CD and says "Wow, this works great. I want to install CentOS!". Several hours later as he DL's the CD's to install it ,"Now this piece of @#&% doesn't work". Also into the trash...
On Fri, 1 May 2009, Scott Silva wrote:
on 5-1-2009 12:19 PM Alan Bartlett spake the following:
On 01/05/2009, Marcus Moeller mail-7BlZPJ8e1eab+SiqwsCprbNAH6kLmebB@public.gmane.org wrote:
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
The opposite would also be true... Novice tries CentOS live CD and says "Wow, this works great. I want to install CentOS!". Several hours later as he DL's the CD's to install it ,"Now this piece of @#&% doesn't work". Also into the trash...
So you prefer someone whose hardware would work on CentOS not use CentOS because he is under the impression that it does not work.
This eg. includes wireless. If they use the LiveCD, no wireless.
Seems counter-productive. And yes there is a good reason to run an enterprise linux on a netbook or a laptop. Saying there are better alternatives is a very personal statement.
On Sat, May 2, 2009 at 11:01 AM, Dag Wieers dag@centos.org wrote:
On Fri, 1 May 2009, Scott Silva wrote:
on 5-1-2009 12:19 PM Alan Bartlett spake the following:
On 01/05/2009, Marcus Moeller mail-7BlZPJ8e1eab+SiqwsCprbNAH6kLmebB@public.gmane.org wrote:
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
The opposite would also be true... Novice tries CentOS live CD and says "Wow, this works great. I want to install CentOS!". Several hours later as he DL's the CD's to install it ,"Now this piece of @#&% doesn't work". Also into the trash...
So you prefer someone whose hardware would work on CentOS not use CentOS because he is under the impression that it does not work.
Some people I know have said something around the lines of: "I tried the CentOS live CD and it didn't, work so I used Ubuntu"
Seems counter-productive. And yes there is a good reason to run an enterprise linux on a netbook or a laptop. Saying there are better alternatives is a very personal statement.
I use CentOS on my laptop and I would never want to change :)
The problem I see with having a live CD that includes all the 3rd party repros is that it will be a lot of work and people will not notice that there is a difference between "base" and "extended". So why not go with 2 different CDs one like it is now the "base" and then have a second one that has all the 3rd party packages enabled and is clearly labeled as such. I am aware that this will be more work to produce.
Cheers Didi
on 5-2-2009 3:01 AM Dag Wieers spake the following:
On Fri, 1 May 2009, Scott Silva wrote:
on 5-1-2009 12:19 PM Alan Bartlett spake the following:
On 01/05/2009, Marcus Moeller mail-7BlZPJ8e1eab+SiqwsCprbNAH6kLmebB-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org wrote:
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
The opposite would also be true... Novice tries CentOS live CD and says "Wow, this works great. I want to install CentOS!". Several hours later as he DL's the CD's to install it ,"Now this piece of @#&% doesn't work". Also into the trash...
So you prefer someone whose hardware would work on CentOS not use CentOS because he is under the impression that it does not work.
This eg. includes wireless. If they use the LiveCD, no wireless.
Seems counter-productive. And yes there is a good reason to run an enterprise linux on a netbook or a laptop. Saying there are better alternatives is a very personal statement.
No. I think they should both work the same, or at least have the live CD have 2 boot modes. A "test my hardware for CentOS" mode, and a "Give me the works" mode. Otherwise, you would have to dl the first CD and try a minimal install, possibly working, possibly not. It is much easier to test a piece of hardware with a live CD that mimics the intended installation exactly BEFORE you buy something rather than scramble for drivers after the fact.
Dear Alan,
> So it becomes effectively useless for everyone with a netbook/laptop and > needs wireless ? I cannot use the LiveCD unless I somehow transfer the > firmware (or remake the LiveCD), you loose users, hurt the project.
I am not sure if missing netbook support leads in loosing our audience.
Novice user: I'd like to see if a reliable, well respected Linux kernel based OS runs on my laptop / notebook.
Venerable wizard: Obtain copies of the various LiveCDs that are available and test.
Novice user: This works, that works, CentOS doesn't work ( ---> into the trash it goes), Ubuntu works, . . .
Result: Another potential user lost.
Surely that didn't really need to be spelt out for you, Marcus?
So, what? CentOS is not the best solution for everyone. E.g. UNR is a very good choice for netbooks, which comes with a app-launcher optimized for small screens and a lot of necessary drivers. I personally see no need to switch these to CentOS.
But Patrice is right. We could offer something like an installable 'Plus' CD which contains some useful driver addons and maybe 3rd party applications (not my decision).
The 'official LiveCD' should just reflect the project development status.
Best Regards Marcus
Dag Wieers wrote:
On Fri, 1 May 2009, Marcus Moeller wrote:
Dear Dag,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
I don't think that there is a need to divide from upstream atm. and am not really willed to break compatibility for features.
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
Creating to different LiveCD (official vs plus) does make sense. With the official LiveCD, you experience the equivalent CentOS release and see how your hardware is supported.
The LiveCD 'plus' includes packages from 3rd party repositories. It could include monitoring software, rescue utilities. It could also demonstrate how CentOS could be improved to support additional hardware. However, trying to find a consensus about which packages/features should be included in this LiveCD is nearly impossible.
A collection of kickstart configuration files could be placed on the project page with the corresponding name for the resulting ISO image. These supported configurations could be updated and maintained. The creation of the ISO images could be done by the interested users.
-- Patrice
Patrice Guay wrote:
Creating to different LiveCD (official vs plus) does make sense. With the official LiveCD, you experience the equivalent CentOS release and see how your hardware is supported.
This pretty much follows what I was thinking as well, with one difference. Lets not call it 'official' and 'plus'. The 'plus' name is good, since it would reflect on the state of the media well, however the word 'official' is a bit misleading. Anything we put on mirror.centos.org and announce publicly is 'official', wether its plus or otherwise. Anyone have other recommendations for what it might be called ? where it == livecd built from the distro only, using pkgs in the distro only [1]
The LiveCD 'plus' includes packages from 3rd party repositories. It could include monitoring software, rescue utilities. It could also demonstrate how CentOS could be improved to support additional hardware.
Including packages from third party repo's is something that makes me a bit uncomfortable, since there is little or no feedback loop that most people would have. I'd like to look at things from the flip side if we can - how about we bring those packages into the centos repo's or XX repo and use that to base things from, that would also allow us to have and test an update path for individual packages. In most cases, its just being able to speak with the existing packagers in the various repos and asking to shadow their work, few if any would object. Many would gladly offer to help directly.
However, trying to find a consensus about which packages/features should be included in this LiveCD is nearly impossible.
I think we just need to be a bit more focused on what the real goal is, and not worry too much about the implementation bits. eg: we should not throw away the idea of a liveDVD, if so required. and/or publishing pxe images that would boot the live[CD]VD's either. Package consensus will come with the efforts and work cycles.
- KB
[1]: we still need the livecd tools from 'outside' the [base].
Karanbir Singh wrote:
Patrice Guay wrote:
Creating to different LiveCD (official vs plus) does make sense. With the official LiveCD, you experience the equivalent CentOS release and see how your hardware is supported.
This pretty much follows what I was thinking as well, with one difference. Lets not call it 'official' and 'plus'. The 'plus' name is good, since it would reflect on the state of the media well, however the word 'official' is a bit misleading. Anything we put on mirror.centos.org and announce publicly is 'official', wether its plus or otherwise. Anyone have other recommendations for what it might be called ? where it == livecd built from the distro only, using pkgs in the distro only [1]
Core or Base maybe?
Or just LiveCD and LiveCD-Plus (or whatever the "Plus" version is to be called)
Ned Slider wrote:
Karanbir Singh wrote:
Patrice Guay wrote:
Creating to different LiveCD (official vs plus) does make sense. With the official LiveCD, you experience the equivalent CentOS release and see how your hardware is supported.
This pretty much follows what I was thinking as well, with one difference. Lets not call it 'official' and 'plus'. The 'plus' name is good, since it would reflect on the state of the media well, however the word 'official' is a bit misleading. Anything we put on mirror.centos.org and announce publicly is 'official', wether its plus or otherwise. Anyone have other recommendations for what it might be called ? where it == livecd built from the distro only, using pkgs in the distro only [1]
Core or Base maybe?
Or just LiveCD and LiveCD-Plus (or whatever the "Plus" version is to be called)
+1 to LiveCD vs LiveCD-Plus
Patrice Guay wrote:
Or just LiveCD and LiveCD-Plus (or whatever the "Plus" version is to be called)
+1 to LiveCD vs LiveCD-Plus
That works for me as well.
Karanbir Singh wrote:
Patrice Guay wrote:
The LiveCD 'plus' includes packages from 3rd party repositories. It could include monitoring software, rescue utilities. It could also demonstrate how CentOS could be improved to support additional hardware.
Including packages from third party repo's is something that makes me a bit uncomfortable, since there is little or no feedback loop that most people would have. I'd like to look at things from the flip side if we can - how about we bring those packages into the centos repo's or XX repo and use that to base things from, that would also allow us to have and test an update path for individual packages. In most cases, its just being able to speak with the existing packagers in the various repos and asking to shadow their work, few if any would object. Many would gladly offer to help directly.
We could start by moving the livecd tools repository from my own webserver to an official CentOS yum repository. The same yum repository could be used to host 3rd party packages that are officially supported with the LiveCD release.
-- Patrice
Patrice Guay wrote:
We could start by moving the livecd tools repository from my own webserver to an official CentOS yum repository. The same yum repository could be used to host 3rd party packages that are officially supported with the LiveCD release.
how about we import all these into the Extras/ and contrib/ repo's instead ?
Dag Wieers wrote:
On Fri, 1 May 2009, Marcus Moeller wrote:
Dear Dag,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
I don't think that there is a need to divide from upstream atm. and am not really willed to break compatibility for features.
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
+1.
The only reason I can see for sticking religiously with the upstream/CentOS base is to use the LiveCD as a tool to test hardware compatibility. IMHO that's a lost opportunity as others have noted and a LiveCD deserves to be so much more than just that. Besides, we all install additional drivers on our real systems when hardware isn't detected or supported by the base offering - just that's somewhat more difficult using a LiveCD which by it's nature is intended (in many users opinion) to be quick and easy to use, not more difficult. We will likely just lose that potential userbase to Ubuntu.
I appreciate this IS a difficult call as when you add functionality you also lose the ability to use it as a strict testing tool for out of the box distro compatibility. Is there any way we can have the best of both worlds?
on 5-1-2009 12:34 PM Ned Slider spake the following:
Dag Wieers wrote:
On Fri, 1 May 2009, Marcus Moeller wrote:
Dear Dag,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
I don't think that there is a need to divide from upstream atm. and am not really willed to break compatibility for features.
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
+1.
The only reason I can see for sticking religiously with the upstream/CentOS base is to use the LiveCD as a tool to test hardware compatibility. IMHO that's a lost opportunity as others have noted and a LiveCD deserves to be so much more than just that. Besides, we all install additional drivers on our real systems when hardware isn't detected or supported by the base offering - just that's somewhat more difficult using a LiveCD which by it's nature is intended (in many users opinion) to be quick and easy to use, not more difficult. We will likely just lose that potential userbase to Ubuntu.
I appreciate this IS a difficult call as when you add functionality you also lose the ability to use it as a strict testing tool for out of the box distro compatibility. Is there any way we can have the best of both worlds?
Unless it could have 2 sets of kernels and initrd's. One bone stock, and one "Enhanced" with extra drivers.
<snip>
Concerning the Live-CD, I would suggest to offer a stable version that reflects the CentOS release with all dis-advantages it may have (not installable, e.g.)
So it becomes effectively useless for everyone with a netbook/laptop and needs wireless ? I cannot use the LiveCD unless I somehow transfer the firmware (or remake the LiveCD), you loose users, hurt the project.
Upstream doesn't have a LiveCD, so I don't see a good point in maintaining the same hardware support in that respect. It only hurts the LiveCD effort. (Same for additional drivers for netbooks/laptops/desktops)
What's even more, upstream does have wireless firmware in their addon repository, so in effect we are not offering the same as they are offering to customers.
Then we would have to stop telling people to test their hardware by DL'ing the live cd since if the live cd had extra drivers, it wouldn't be a fair test.
Recalling that we have to put in the line of the kernel (boot)
noapic pnpbios = off
To operate the video driver in some noteboot HP / Dell.
2009/5/1 Dag Wieers dag@centos.org:
Hi,
I am amazed by the usefulness of the CentOS LiveCD and I would like to discuss the scope of the project. Some of the recommendations I made go against the original idea of LiveCD project.
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
What are the various goals of the LiveCD project ?
- A way for people to see what hardware is supported by CentOS - with drivers included by upstream only - with drivers that can be added with little effort (to eg. accomodate eeepc and other purposes) - including a tool to send a hardware status report to a central database
- A tool for recovering a broken system - include only a few basic tools for debugging
- A tool for installing CentOS on a system
- A workable environment for people to use day by day - do we want persistent storage - does it need to be installable from Windows - does it include wireless firmware
- Provide the tools for people to make their own spins - all sorts of appliances, eg. DVR, rescue, ...
I am sure we can list more applicable use-cases for this project. And in the end we have to agree which fit the LiveCD project (eg. whether we make an official CentOS LiveCD and a seperate CentOS-plus LiveCD) and what the main priorities are.
Also, to stimulate the LiveCD development there should be no reason why we couldn't release a LiveCD 5.3.1 or LiveCD 5.3.2 if there are compelling reasons to make a new respin (and update the mirrors).
In the end I see a bright future for the LiveCD as the project where driver-development (dasha, elrepo), CentOS-based appliances, hardware testing, and much more come together. To me it is one of those unique opportunities to drive different other projects with a very positive attitude as it brings (even more) meaning to CentOS to a lot of (new) people.
Where do we want to go with this and what is realistically achievable ?
I am thrilled to find out :)
-- dag wieers, dag@wieers.com, http://dag.wieers.com/ -- [Any errors in spelling, tact or fact are transmission errors] _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
On Sat, 2 May 2009, Thiago Avelino wrote:
Recalling that we have to put in the line of the kernel (boot)
noapic pnpbios = off
To operate the video driver in some noteboot HP / Dell.
With syslinux it is possible to provide boot-options automatically based on hardware. So we could build a small database with known boot-options for CentOS kernels (and keep contact-info for new kernel respins) so it would automagically work.
It does require either a specialized module, or better, lua-scripting support, which is not officially supported yet.
Dag Wieers wrote:
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
Sounds good. Are there any legal issues around the wireless firmware files ? If its a case of just talking to the providers of these firmware to get permission - thats something which we could potentially do.
I do agree that as soon as we leave the deliver-what-upstream-has path, we may open a can of worms (do we also want to fix known bugs ? replace upstream software ? legality ?), so we have to decide what is desirable, what is possible and where the project's effort ends.
This is an interesting subject to me, and I think we ( dag and I ) have spoken about this many times in different venues. Exactly where does the line get drawn and exactly what is considered 'CentOS brandable' and what isnt. Since the topic is now being actively discussed I'll start a new thread just focused on this one idea and we can thrash it out there.
What are the various goals of the LiveCD project ?
Rather than goals, can we narrow it down to 'use cases'. I think there is only one goal - to build either one or many livecd's based on and around the CentOS distro. If we focus on use-case, it allows wider and more elaboate cases to come in, and we can then look at mechanisms to solve those issues in some sort of a priority order.
Perhaps I am being pedantic, but it will help in the long run.
I am sure we can list more applicable use-cases for this project. And in the end we have to agree which fit the LiveCD project (eg. whether we make an official CentOS LiveCD and a seperate CentOS-plus LiveCD) and what the main priorities are.
Another thing I'd like to propose is that we seperate out the livecd 'project' into upstream for the CentOS provided livecd(s). That would allow a separate entity and therefore a different crew to focus only on the tooling components and perhaps work with parallel branches.[1]
Also, to stimulate the LiveCD development there should be no reason why we couldn't release a LiveCD 5.3.1 or LiveCD 5.3.2 if there are compelling reasons to make a new respin (and update the mirrors).
Nope, but it would be 5.3-1 and 5.3-2! So as to not confuse with upstreams 5.3.1 and 5.3.2 etc. Although most people who dont really know the whole story will get confused anyway
Where do we want to go with this and what is realistically achievable ?
Where I would like to see this go, and this is my personal take on this is:
- Send out a call for user-cases - Setup a feedback process, email / lists work well - Nominate a person ( Patrice ? ) to be the 'lead' on the efforts - Setup whatever infrastructure we need to make this happen - Build foo!
Karanbir Singh wrote:
Dag Wieers wrote:
But my (outside) stance on the LiveCD is that it should give the best achievable experience possible for people when trying CentOS. To me that includes adding drivers that are available in other repositories (which are missing from upstream, including wireless firmware, etc...)
Sounds good. Are there any legal issues around the wireless firmware files ? If its a case of just talking to the providers of these firmware to get permission - thats something which we could potentially do.
Intel iwl firmwares are licensed as: Redistributable, no modification permitted. Upstream redistribute them in one of their channels as do rpmforge.
Ned Slider wrote:
Intel iwl firmwares are licensed as: Redistributable, no modification permitted. Upstream redistribute them in one of their channels as do rpmforge.
is that the only firmware in question here ? whats the real license on that package, does it have conflicts being packaged with GPL stuff ?
Karanbir Singh wrote:
Ned Slider wrote:
Intel iwl firmwares are licensed as: Redistributable, no modification permitted. Upstream redistribute them in one of their channels as do rpmforge.
is that the only firmware in question here ? whats the real license on that package, does it have conflicts being packaged with GPL stuff ?
1. Probably not, but it's the only wireless interface in my machine so I provided the information in the hope it might be useful.
2. http://intellinuxwireless.org/?n=Downloads
Copyright (c) 2006-2009, Intel Corporation. All rights reserved.
Redistribution. Redistribution and use in binary form, without modification, are permitted provided that the following conditions are met:
* Redistributions must reproduce the above copyright notice and the following disclaimer in the documentation and/or other materials provided with the distribution. * Neither the name of Intel Corporation nor the names of its suppliers may be used to endorse or promote products derived from this software without specific prior written permission. * No reverse engineering, decompilation, or disassembly of this software is permitted.
Limited patent license. Intel Corporation grants a world-wide, royalty-free, non-exclusive license under patents it now or hereafter owns or controls to make, have made, use, import, offer to sell and sell ("Utilize") this software, but solely to the extent that any such patent is necessary to Utilize the software alone, or in combination with an operating system licensed under an approved Open Source license as listed by the Open Source Initiative at http://opensource.org/licenses. The patent license shall not apply to any other combinations which include this software. No hardware per se is licensed hereunder.
DISCLAIMER. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
3. Packaged or redistributed? Why would you want to package it along with anything else? No issue redistributing it on a LiveCD or other media though afaics. Only an issue if you want to stay wholly GPL.
The packages are in rpmforge so help yourself. They probably belong in Extras anyway (upstream has them in their Supplementary channel).
Ned Slider wrote:
The packages are in rpmforge so help yourself. They probably belong in Extras anyway (upstream has them in their Supplementary channel).
we cant just add everything from Red Hat's supplimentary sections to CentOS - most of those packages ( I still seem to be under the impression that all of them ) are in Supplimentary since Red Hat are not allowed to make them public, and in 9 times out of 10 it will be due to license restrictions.
Ofcourse there is also the big question of exactly where 'we' exist, and what rules/laws are applicable in what way wherever that 'we exist' is.
Karanbir Singh wrote:
Ned Slider wrote:
The packages are in rpmforge so help yourself. They probably belong in Extras anyway (upstream has them in their Supplementary channel).
we cant just add everything from Red Hat's supplimentary sections to CentOS - most of those packages ( I still seem to be under the impression that all of them ) are in Supplimentary since Red Hat are not allowed to make them public, and in 9 times out of 10 it will be due to license restrictions.
Presumably they make them available to anyone with a RHN subscription otherwise why would the channel exist, and they are perfectly entitled to redistribute them (the Intel firmwares) if they want under the terms of the Intel license. My assumption was that RH only distributes as part of the main distro stuff that is open source/GPL or whatever (I don't believe the source is available for these binary firmwares).
I don't know what else is in Supplementary so can't comment on that, but my guess is it's mostly closed source binary distributables (eg, Acrobat Reader, flash-plugin, RealPlayer, Java? etc) that RH can't ship as part of a fully open source distro.
Ned Slider wrote:
Presumably they make them available to anyone with a RHN subscription otherwise why would the channel exist,
yup! but most people who get that far have paid something for the access, so if Red Hat were to take on some cost for the packages, it would / could be offset against that.
I don't know what else is in Supplementary so can't comment on that, but my guess is it's mostly closed source binary distributables (eg, Acrobat Reader, flash-plugin, RealPlayer, Java? etc) that RH can't ship as part of a fully open source distro.
That is also my understanding. Licensing issues. There was a long thread about this when the Sun guys came around talking about CentOS doing a Sun's built rpm set in the Extras/ pkgs. I guess most of that would still be relevant. We could revisit if required though. Or if the situation has changed enough. I feel in the specific case of Java, openjdk address's much of whats needed.
Karanbir Singh wrote:
What are the various goals of the LiveCD project ?
Rather than goals, can we narrow it down to 'use cases'. I think there is only one goal - to build either one or many livecd's based on and around the CentOS distro. If we focus on use-case, it allows wider and more elaboate cases to come in, and we can then look at mechanisms to solve those issues in some sort of a priority order.
Perhaps I am being pedantic, but it will help in the long run.
If you make the rebuilding mechanism a priority so anyone can easily build a custom bootable iso (cd or dvd) or usb device, the version you start with or other use cases won't matter that much.
Are there any existing tools that can generically move a system back and forth from a running install (perhaps a virtual machine) to a usb device with support for booting on different hardware, or to a bootable iso image? Or, perhaps even more generic: is there a tool to move an existing installed system to different hardware?
-
Alan Bartlett -
Dag Wieers -
Didi -
Karanbir Singh -
Les Mikesell -
Marcus Moeller -
Ned Slider -
Patrice Guay -
Scott Silva -
Thiago Avelino