Re: Supporting armel/armhf in wheezy-lts
Hi folks, 2016-04-25 9:43 GMT+02:00 Santiago Ruano Rincón : > From the other side, what about armel/armhf LTS support? No objecting > voices? > I enjoy my debian wheezy on my iop32x, which wheezy is the last supported release for my arch, so +1 for consider LTS support on my n2100. thanks, regards happy hacking federico
Re: Supporting armel/armhf in wheezy-lts
Hi, (fwiw, I am dropping all CCs except debian-lts) On 23/04/2016 14:41, Raphael Hertzog wrote: > Concerning the other concerns that you brought up, they do not seem > specific to the support of the armel/armhf architectures. They > rather question the LTS project as a whole and the usage of money > within Debian. But the LTS project has been running smoothly for > two years now and thus there's nothing new to discover here. I do not think Tollef is questioning the whole LTS project, but only the financial part of it, AFAIK. Ideally, we should be able to find a solution within Debian or with help of entities somehow bound to Debian, like a TO. While I support that idea, I find the timing unfortunate and the question about adding armel/armhf support in wheezy-lts should not be blocked by that. As showed below, Raphaël has been proactive and is trying to push this (not so easy) subject forward. > But your questions are legitimate and we should discuss them. I have > submitted a BoF for DebConf16 to discuss how to use money to fund > "Debian projects" (and not only sprints) and I invite you to join > this discussion if you can attend. FWIW, the goal of the above mentioned BoF is to discuss openly how to fund such initiatives within Debian. Talking money should not be taboo, even if we are a volunteer organisation. We can find solutions where we avoid conflict of interest and dominant positions… and stay coherent wrt. what we do and how we do it. We can have areas where project members are encouraged to help with money. We can start collecting ideas on how such activities could be organized, but before working on the implementation details, I think it would be more useful to clarify a few things so that everyone share the same level of information. Based on the following points, we could describe the current situation, the target and see how we transition: 1) Who will organize this project? A delegation can be made, if we have a clear idea of the role and what it entails. Are there some conditions to put here to avoid dominant positions or such unpleasant situations? 2) Who can contribute to the LTS effort? 3) How money is raised to fund this effort? 4) How contributors are paid? Is it mandatory that participants are able to emit an invoice? 5) How to guarantee some equity between contributors? 6) Any more question/condition? Regards, -- Mehdi
Re: Supporting armel/armhf in wheezy-lts
On Mon, 25 Apr 2016, Antoine Beaupré wrote: > On 2016-04-25 09:27:34, Raphael Hertzog wrote: > > - I don't think that the bounty model gives the correct incentive for > > the security work, and you would have a hard time covering the hard > > packages... > > I think this is a critical part of it. Bounties are fine and fun if you > want to scratch an itch and someone happens to want to pay for it. But > then you'd probably do it anyways if there was no bounty either. It's a > small incentive, often not sufficient to get hard things done, and most > of the time not enough to pay the rent. My point was not about this. First, bounty does not imply an amount that does not match the work required... but it usually implies a fixed amount defined up-front which is close to impossible to correctly guess. And it usually implies that you get to pick what you work on and that you can ignore the bounties that do not interest you. People will pick easy updates quickly and the hard ones will languish indefinitely. Also given a fixed amount you will want to do as many as possible as quickly as possible and you are more likely to badly test an update. Here we need to be able to say, ok we're going to pay you for X hours this month, but you will have to tackle also the hard updates that have been hanging around and those that will come... (and also LTS frontdesk duties) Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
On 2016-04-25 09:27:34, Raphael Hertzog wrote: > - I don't think that the bounty model gives the correct incentive for > the security work, and you would have a hard time covering the hard > packages... I think this is a critical part of it. Bounties are fine and fun if you want to scratch an itch and someone happens to want to pay for it. But then you'd probably do it anyways if there was no bounty either. It's a small incentive, often not sufficient to get hard things done, and most of the time not enough to pay the rent. Security work is basically the opposite of that. You need to triage painfully through obscure issues in programming languages you are not necessarily comfortable with. There's a lot of legwork that needs to happen before a patch actually comes through: sometimes, most of the work is just that: triaging and closing issues... And even if you actually close a CVE, you are actually porting an already existing patch most of the time: it's not original work. So in the end, why should *you* get that bounty and not the original author? It gets weird real quickly IMHO. A. -- Premature optimization is the root of all evil - Donald Knuth
Re: Supporting armel/armhf in wheezy-lts
Hi, this is really getting off-topic from the initial discussion, so I'm dropping all lists except LTS and I add the leader in the loop (he was already following it but through debian-wb-team AFAIK). On Mon, 25 Apr 2016, Luca Filipozzi wrote: > On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > > Do you have some concrete suggestions? > > Decrease the separation by moving the funds management into Debian proper (via > a TO like SPI) and move to a bounty model for working on LTS. Make sure we're > transparent with our language regarding Debian being produced by volunteers > (eg: "The Debian Project consists of volunteers, and our products are > developed > entirely by volunteers." on [1]) by commenting on how bounties are available > (or something). Consider making LTS management a delegated team. I'm happy to investigate how to go in this direction but there are many problems to solve: - I don't think that the bounty model gives the correct incentive for the security work, and you would have a hard time covering the hard packages... - for wheezy LTS, we are looking at hiring external support to maintain for example Xen 4.1.x for us. That's just not possible in a bounty model. The experts we hire want a contract. - managing the bounty program is a lot of work and I don't see anyone volunteering to handle this work... I do it for now because my time is paid by the small cut that Freexian is taking on the money collected. I would not do it for free. > Increase the separation by removing the fundraising statements / links from > the > LTS pages previously mentioned, making Freexian just another consultancy > listed > on the consultancy pages. If your problem is only with http://wiki.debian.org/LTS/Funding then I'm happy to fix this but what about all the references where Freexian's offer is mentioned? In our press releases, eg https://www.debian.org/News/2014/20140616 https://www.debian.org/News/2015/20150424 https://www.debian.org/News/2016/20160212 (note how over time we moved to link the Funding page instead of freexian.com to avoid putting Freexian too much into light since I'm just the intermediary here) On Planet Debian, through the monthly reports I put on my blog: http://raphaelhertzog.com/tag/Freexian+LTS/ On Debian Project News which regularly link to the above monthly reports (or the reports of paid contributors directly which tend to link back to Freexian). If all those appearence of Freexian's offer are problematic and if you want to hide all those as well, then I fear that we would not have reached the level of support that we need to get LTS to a sustainable level. So while I'm happy to clarify the relationship, I also don't want to hide its existence, that would defeat its purpose. So maybe the right path is a third solution. Create a Debian Development Partner program defining rules for intermediaries like me... in fact this sounds quite close to the "Debian Labs" we once had[1]. The rules could include: - monthly reports to explain how the collected money has been put to use - define a percentage to give back to a trusted organization - maximum percentage that can be taken to cover "administrative work" - validation by the DPL and/or the partner team And then we would have less concerns in mentioning such commercial offers on our official communications. And less problems in including the sponsors of LTS in our official sponsor page (under rules to be defined). What do you think? [1] https://lists.debian.org/debian-project/2003/09/msg00020.html the thread starts here: https://lists.debian.org/debian-project/2003/09/threads.html#00020 and continues the month after: https://lists.debian.org/debian-project/2003/10/threads.html#00058 > None of this is meant to diminish or tarnish the very significant contribution > that you or Freexian are making, which are both extensive and impressive. I'm > seeking greater definition of the role and the language used. Thanks for saying this explicitly, it's appreciated. Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
On Mon, 2016-04-25 at 02:07 +, Luca Filipozzi wrote: > On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > > > > Do you have some concrete suggestions? > Decrease the separation by moving the funds management into Debian proper (via > a TO like SPI) and move to a bounty model for working on LTS. [...] I seriously doubt my employer would let me work on LTS on this basis. Ben. -- Ben Hutchings All the simple programs have been written, and all the good names taken. signature.asc Description: This is a digitally signed message part
Re: Supporting armel/armhf in wheezy-lts
Hi,
just nitpicking about a single detail here…
On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote:
> The reason why I did it within Freexian is that it was just the simplest
> way to get it started and to prove that given some sane rules it's
> possible to not harm the Debian community. Now that it's proven, I believe
> we can and should discuss how to handle it at the Debian level directly.
a.) this is nothing which can be proven and then it's done. Instead it
needs to be continously proven that this doesnt harm Debians volunteer
work model or otherwise doesn't hurt the community.
b.) doing this within Freexian is entirely different than doing this
within Debian.
--
cheers,
Holger
p.s.: it just occurred me that doing LTS differently might also hurt
the community in unexpected ways: LTS itself might work less well, and
that would also be harm. ("LTS stopped working".)
signature.asc
Description: Digital signature
Re: Supporting armel/armhf in wheezy-lts
El 25/04/16 a las 02:07, Luca Filipozzi escribió: > On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > > Do you have some concrete suggestions? > > Decrease the separation by moving the funds management into Debian proper (via > a TO like SPI) and move to a bounty model for working on LTS. Make sure we're > transparent with our language regarding Debian being produced by volunteers > (eg: "The Debian Project consists of volunteers, and our products are > developed > entirely by volunteers." on [1]) by commenting on how bounties are available > (or something). Consider making LTS management a delegated team. > > OR > > Increase the separation by removing the fundraising statements / links from > the > LTS pages previously mentioned, making Freexian just another consultancy > listed > on the consultancy pages. > > None of this is meant to diminish or tarnish the very significant contribution > that you or Freexian are making, which are both extensive and impressive. I'm > seeking greater definition of the role and the language used. As a Debian LTS paid contributor, I would disagree to get money from Debian own funds or directly by Debian in exchange of a my work. And this is for any of my work on the project. I agree with the consultancy model and I obviously support the Freexian umbrella, which I think is highly valuable. So if the LTS Funding statement is not clear enough, I prefer to increase the separation in https://wiki.debian.org/LTS/Funding. Thanks for bringing this up! From the other side, what about armel/armhf LTS support? No objecting voices? Cheers, Santiago signature.asc Description: PGP signature
Re: Supporting armel/armhf in wheezy-lts
On Monday, April 25, 2016 02:07:01 AM Luca Filipozzi wrote: > On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > > Do you have some concrete suggestions? > > Decrease the separation by moving the funds management into Debian proper > (via a TO like SPI) and move to a bounty model for working on LTS. Make > sure we're transparent with our language regarding Debian being produced by > volunteers (eg: "The Debian Project consists of volunteers, and our > products are developed entirely by volunteers." on [1]) by commenting on > how bounties are available (or something). Consider making LTS management > a delegated team. > > OR > > Increase the separation by removing the fundraising statements / links from > the LTS pages previously mentioned, making Freexian just another > consultancy listed on the consultancy pages. > > None of this is meant to diminish or tarnish the very significant > contribution that you or Freexian are making, which are both extensive and > impressive. I'm seeking greater definition of the role and the language > used. > > [1]: https://www.debian.org/devel/join/ Any suggestions on how to get that done in the next two days before wheezy-lts starts? It might be a bit more practical to defer the idea of completely changing the LTS program to a moment when there's a bit more time (maybe Debconf). So far, I don't think anyone has specifically objected to the addition of armel/armhf. Scott K signature.asc Description: This is a digitally signed message part.
Re: Supporting armel/armhf in wheezy-lts
On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > Do you have some concrete suggestions? Decrease the separation by moving the funds management into Debian proper (via a TO like SPI) and move to a bounty model for working on LTS. Make sure we're transparent with our language regarding Debian being produced by volunteers (eg: "The Debian Project consists of volunteers, and our products are developed entirely by volunteers." on [1]) by commenting on how bounties are available (or something). Consider making LTS management a delegated team. OR Increase the separation by removing the fundraising statements / links from the LTS pages previously mentioned, making Freexian just another consultancy listed on the consultancy pages. None of this is meant to diminish or tarnish the very significant contribution that you or Freexian are making, which are both extensive and impressive. I'm seeking greater definition of the role and the language used. [1]: https://www.debian.org/devel/join/ -- Luca Filipozzi http://www.crowdrise.com/SupportDebian
Re: Supporting armel/armhf in wheezy-lts
On Sun, 2016-04-24 at 22:24 +0200, Ben Hutchings wrote: > On Mon, 2016-04-18 at 09:45 +0200, Markus Koschany wrote: > > > > Am 18.04.2016 um 08:45 schrieb Guido Günther: > > [...] > > > > > > > > > I'm all for it (although it's easy to say for me since the most burden > > > will probably be on the kernel team) and having it as experimental with > > > a single sponsor seems sensible. > > +1 from my side too. I guess I'm one of those armel hobbyists and I > > could test the software even on real hardware. > > > > > > > > > > > I assume the level of sponsorship offered is reasonable to support an > > > arm port? I still wonder how we could would make it simpler to have this > > > support end up at the right places (i.e. LTS gets the sponsorship while > > > other teams like release team, security team also have additional work)? > > I also think that the opinions of the kernel team / Ben are crucial if > > we want to support ARM in the future. Otherwise I would expect that > > supporting ARM scales pretty well and that it mainly requires more time > > for testing the software. > [...] > > Openblocks ships its own kernel packages for wheezy, so they won't even I meant the company, Plat'Home. Ben. > care about the linux package. I also don't remember spending much time > on architecture-specific issues in stable updates (other than x86). > > Ben. > -- Ben Hutchings Larkinson's Law: All laws are basically false. signature.asc Description: This is a digitally signed message part
Re: Supporting armel/armhf in wheezy-lts
On Mon, 2016-04-18 at 09:45 +0200, Markus Koschany wrote: > Am 18.04.2016 um 08:45 schrieb Guido Günther: > [...] > > > > I'm all for it (although it's easy to say for me since the most burden > > will probably be on the kernel team) and having it as experimental with > > a single sponsor seems sensible. > +1 from my side too. I guess I'm one of those armel hobbyists and I > could test the software even on real hardware. > > > > > I assume the level of sponsorship offered is reasonable to support an > > arm port? I still wonder how we could would make it simpler to have this > > support end up at the right places (i.e. LTS gets the sponsorship while > > other teams like release team, security team also have additional work)? > I also think that the opinions of the kernel team / Ben are crucial if > we want to support ARM in the future. Otherwise I would expect that > supporting ARM scales pretty well and that it mainly requires more time > for testing the software. [...] Openblocks ships its own kernel packages for wheezy, so they won't even care about the linux package. I also don't remember spending much time on architecture-specific issues in stable updates (other than x86). Ben. -- Ben Hutchings Larkinson's Law: All laws are basically false. signature.asc Description: This is a digitally signed message part
Re: Supporting armel/armhf in wheezy-lts
On Sun, 24 Apr 2016, Raphael Hertzog wrote: > > Consequently, I find the use of Debian resources such as the advertising > > above > > and/or the use of Debian machines as being problematic. > > It is really worse that indicating that some Debian services are handled > by a given (commercial) CDN? Fastly and MaxCDN aren't asking for donations to keep Debian running. They are the sponsors here. -- | .''`. ** Debian ** Peter Palfrader | : :' : The universal https://www.palfrader.org/ | `. `' Operating System | `-https://www.debian.org/
Re: Supporting armel/armhf in wheezy-lts
On Sun, Apr 24, 2016 at 09:55:10AM +0200, Raphael Hertzog wrote: > > https://wiki.debian.org/LTS/ makes it appear that LTS is an official Debian > > effort. > > And it is. There are multiple Debian developers who have initiated this > project, have been organizing it on debian-lts@lists.debian.org (and not > all of them have been paid by Freexian, including many members of the > security team). Indeed. In addition there's also quite a few DDs who have worked on LTS updates who are not payed through the funds collected by Freexian (usually for packages maintained by them). Cheers, Moritz
Re: Supporting armel/armhf in wheezy-lts
On 04/24/2016 09:55 AM, Raphael Hertzog wrote: As a Debian developer (and not only as a freelancer/company owner) I do care about Debian LTS because it is important for Debian's long term relevance (at least according to me). An off-topic: As a Debian user (not a developer/programmer, at least not yet) I strongly support Debian LTS project. Occasionally I do some voluntary job by promoting Debian (and Linux in general) in various technical conferences and academic visits. Having said that, I wonder if there are some resources for promoters of that kind that you could recommend? For example, I am particularly interested in non-expensive publishing houses that are willingly to publish books in lesser-used disciplines, such as using Linux as a component of amateur radio networks in various educational environments, ... Best regards, Miroslav Skoric -- tutorial instructor: http://www.comcas.org/ http://sdiwc.net/conferences/icctim2015/ http://wocn2014.org/wtutorials.html http://sdiwc.net/conferences/2014/iceee2014/tutorial/ http://www.juit.ac.in/ISPCC_2013/ http://www.buruniv.ac.in/ICCS-2013/index.html http://www.sdiwc.net/thi/ http://www.sdiwc.net/kl/ http://www.iaeng.org/IMECS2011 http://www.dirf.org/ndt2010 http://www.icact.org/ http://www.iaeng.org/IMECS2009 http://www.iaria.org/conferences2008/ICWMC08.html http://www.wseas.org/conferences/2008/greece/education/ http://www.wseas.us/conferences/2009/rodos/education http://eurocon2007.isep.pw.edu.pl/index.php?id=tutorials.php http://tldp.org/HOWTO/FBB.html workshop / visiting lecturer: IIITM Gwalior, India SRM University, Chennai, India Vardhaman College of Engineering, Hyderabad, India GRIET, Hyderabad, India NIT Surat, India IIT, Banaras Hindu University, Varanasi, India Techno India NJR Institute of Technology, Udaipur, India BMS College of Engineering, Bangalore, India IIITM-K, Technopark, Trivandrum, India BU, Bangkok, Thailand IIUM, Kuala Lumpur, Malaysia book chapter author: Handbook of Research on Human Performance and Instructional Technology ISBN: 978-1-60566-782-9; 678 pp; October 2009 Published under Information Science Reference, an imprint of IGI Global http://www.igi-global.com/reference/details.asp?id=34814 Simulation in Computer Network Design and Modeling: Use and Analysis ISBN: 978-1-46660-191-8; 582 pages; February 2012 Published by IGI Global, USA http://www.igi-global.com/book/simulation-computer-network-design-modeling/58282 Wireless Networks and Security: Issues, Challenges and Research Trends ISBN: 978-3-642-36168-5; 510 pages; February 2013 Published by Springer, Germany http://www.springer.com/engineering/signals/book/978-3-642-36168-5 Handbook of Research on Progressive Trends in Wireless Communications and Networking ISBN: 9781466651708; 592 pages; February 2014 Published by IGI Global, USA http://www.igi-global.com/book/wireless-communications-networking/90600
Re: Supporting armel/armhf in wheezy-lts
Hi, thanks for the feedback. On Sat, 23 Apr 2016, Julien Cristau wrote: > I think one of the contentious points is how "Freexian raising funds to > work on Debian LTS" is already too close to calling itself "Debian LTS > fundraising", so I'm not sure bringing them closer would alleviate > anyone's concerns. On Sat, 23 Apr 2016, Luca Filipozzi wrote: > I'm interested in working on Debian Partners but not if it includes Debian LTS > fundraising as currently structured: either Debian LTS fundraising falls under > the same umbrella as other Debian fundraising (and subject to the same rules, > including that funds not be used for reimbursing effort, potentially) or it is > separate and branded as a Freexian service. Just to be clear, I don't mind if the money (and the corresponding work that I don't particularly enjoy) is handled by a trusted organization. But as long as we assume that Debian is not willing to "reimburse effort" as you put it, then it just does not make sense since it would mean that Debian LTS would not exist. The reason why I did it within Freexian is that it was just the simplest way to get it started and to prove that given some sane rules it's possible to not harm the Debian community. Now that it's proven, I believe we can and should discuss how to handle it at the Debian level directly. As a Debian developer (and not only as a freelancer/company owner) I do care about Debian LTS because it is important for Debian's long term relevance (at least according to me). > https://wiki.debian.org/LTS/ makes it appear that LTS is an official Debian > effort. And it is. There are multiple Debian developers who have initiated this project, have been organizing it on debian-lts@lists.debian.org (and not all of them have been paid by Freexian, including many members of the security team). > However, https://wiki.debian.org/LTS/Funding directs those interested in > providing funding in support LTS to do so through Freexian rather than through > Software in the Public Interest. This page directs to Debian developers who are willing to do the work for money. Most of them preferred to join the initiative I started behind the Freexian umbrella... mainly because it makes sense to have a common offer, clear rules, etc. While you might find that the distinction is not important, I find it important. Since Freexian is external to Debian, there's no reason for it to have some exclusive relationship concerning Debian LTS. We could have other similar structures. I do make a distinction between the Debian LTS project and Freexian as an administrative facilitator. > Consequently, I find the use of Debian resources such as the advertising above > and/or the use of Debian machines as being problematic. It is really worse that indicating that some Debian services are handled by a given (commercial) CDN? > Make the distinction clearer, and the problem goes away. Do you have some concrete suggestions? Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
On Sat, Apr 23, 2016 at 02:41:30PM +0200, Raphael Hertzog wrote: > That's why I have been advocating for a change of the DMUP. It has been used > far too often to annoy persons who are being paid (or who are accepting > donations) to work on Debian instead of causing real troubles that could > annoy sponsors or create problems to the DSA team. I think that the distinction between Debian LTS and Freexian is too indefinite. https://wiki.debian.org/LTS/ makes it appear that LTS is an official Debian effort. However, https://wiki.debian.org/LTS/Funding directs those interested in providing funding in support LTS to do so through Freexian rather than through Software in the Public Interest. Freexian isn't a TO and the funds it collects aren't subject to the same disbursement conditions (approval by DPL; execution by SPI Treasurer; review by Debian Auditor) as funds collected by Debian. Consequently, I find the use of Debian resources such as the advertising above and/or the use of Debian machines as being problematic. Make the distinction clearer, and the problem goes away. > I have also been looking at ways to bring the "LTS funding" closer to Debian > and to find a way to join all this in the Debian Partner program but we don't > have many volunteers interested in this work. We discussed it a bit last year > during Debconf with Luca Filippozi, Martin Krafft and Neil McGovern, but this > never went further. And I obviously don't want to be leading this project due > to the clear conflict of interest that I would have... I'm interested in working on Debian Partners but not if it includes Debian LTS fundraising as currently structured: either Debian LTS fundraising falls under the same umbrella as other Debian fundraising (and subject to the same rules, including that funds not be used for reimbursing effort, potentially) or it is separate and branded as a Freexian service. I don't begrudge people making a living supporting Debian. I do mind the indistinction between Debian LTS and Freexian. It's situations like these that cause organizations to have things like DMUP. (You also have a clear conflict of interest in arguing for changes to the DMUP. That said, periodic re-examination of policies and procedures is healthy for organizations.) (The majority of Debian equipment is still hosted by post-secondary institutions where commercial activity using university resources is frowned upon.) -- Luca Filipozzi http://www.crowdrise.com/SupportDebian
Re: Supporting armel/armhf in wheezy-lts
On 23/04/16 13:41, Raphael Hertzog wrote: Hi, On Fri, 22 Apr 2016, Tollef Fog Heen wrote: I am not speaking on behalf of DSA here. Thanks for making this clear. I also want to explain why I included DSA in the discussion: I wanted to make sure that the fact that we run wheezy armel/armhf buildd for two more years do not go against some DSA plans to decommission some machines running those buildd and that you had no other problems to keep those machine running during the LTS timeframe. I'm not DSA but my understanding is that all suites are built on the same buildds, since jessie supports armel and armhf and squeeze will almost certainly support at least one and hopefully both of those we will need autobuilder hardware for the forseeable future.
Re: Supporting armel/armhf in wheezy-lts
On Sat, Apr 23, 2016 at 14:41:30 +0200, Raphael Hertzog wrote: > I have also been looking at ways to bring the "LTS funding" closer to Debian > and to find a way to join all this in the Debian Partner program but we > don't have many volunteers interested in this work. We discussed it a bit > last year during Debconf with Luca Filippozi, Martin Krafft and Neil > McGovern, but this never went further. And I obviously don't want to be > leading this project due to the clear conflict of interest that I would > have... > I think one of the contentious points is how "Freexian raising funds to work on Debian LTS" is already too close to calling itself "Debian LTS fundraising", so I'm not sure bringing them closer would alleviate anyone's concerns. Cheers, Julien
Re: Supporting armel/armhf in wheezy-lts
Hi, On Fri, 22 Apr 2016, Tollef Fog Heen wrote: > I am not speaking on behalf of DSA here. Thanks for making this clear. I also want to explain why I included DSA in the discussion: I wanted to make sure that the fact that we run wheezy armel/armhf buildd for two more years do not go against some DSA plans to decommission some machines running those buildd and that you had no other problems to keep those machine running during the LTS timeframe. Concerning the other concerns that you brought up, they do not seem specific to the support of the armel/armhf architectures. They rather question the LTS project as a whole and the usage of money within Debian. But the LTS project has been running smoothly for two years now and thus there's nothing new to discover here. But your questions are legitimate and we should discuss them. I have submitted a BoF for DebConf16 to discuss how to use money to fund "Debian projects" (and not only sprints) and I invite you to join this discussion if you can attend. That said the DMUP is not a foundation document defining our values, it's a set of rules that DSA edicted long ago to: - avoid problems for the sponsors that are hosting our servers (that were quite often public institutions like universities) - give some ground to DSA to block accounts for obvious misuses of the Debian resources (hosting warez, hosting of webpages unrelated to Debian, running daemons unrelated to Debian, for example to crack passwords or cryptography challenges, etc.) So while the text contains wording again "commercial purpose", it should not include work that benefits to Debian. Otherwise we are all in troubles. Many people are paid to work on Debian (even if we do not always know it) and we don't want to refuse their work. I tried to discuss with James Troup who wrote the text with other members a long time ago and while he did not remember the exact reasons why they came up with this, he clearly mentionned the fact that this was inspired by rules used in universities and that it made sense to align to those since we were often hosted in such places. But honestly, the fact that some people are paid to work on Debian LTS and that they upload packages that end up on the servers of universities sponsoring us is not a real problem. That's why I have been advocating for a change of the DMUP. It has been used far too often to annoy persons who are being paid (or who are accepting donations) to work on Debian instead of causing real troubles that could annoy sponsors or create problems to the DSA team. I have also been looking at ways to bring the "LTS funding" closer to Debian and to find a way to join all this in the Debian Partner program but we don't have many volunteers interested in this work. We discussed it a bit last year during Debconf with Luca Filippozi, Martin Krafft and Neil McGovern, but this never went further. And I obviously don't want to be leading this project due to the clear conflict of interest that I would have... Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
On Fri, Apr 22, 2016 at 05:16:19PM +0200, Tollef Fog Heen wrote: > > The LTS period is about to start soon and we will send out an announce > > soon... it would be nice to be able to say a word about armel/armhf, so > > an official confirmation from ftpmasters/DSA would be nice. > > So, I'm a bit divided on all this. While the DMUP is heavy artillery > and we should be careful about invoking it, I think it crosses the line > in «Don't use Debian Facilities for private financial gain or for > commercial purposes, including consultancy or any other work outside the > scope of official duties or functions for the time being, without > specific authorization to do so.», so you need explicit authorization > before you start. I'm a bit surprised you bring this up now, when the question at hand is, whether to add armel+armhf as new supported architectures to Wheezy LTS. Because, also for doing Squeeze LTS for i386 and amd64, Debian machines were used, with - I'd say - consens of the project that this is ok, IOW: there was specific authorisation to do so. There's another Debian project which is in a similar situation: official Debian images for cloud services, made on Debian hardware. There's a commercial purpose too. (And then we get into the blurry area of all the consultants and freelancers and employees working on paid time on Debian stuff using Debian ressources…) And while I think we absolutly should discuss these things (dc16?) and probably revisited the DMUP (written how long ago?) this will all take time and right now we have a somewhat pressing issue at hand: Can we add armel/armhf to Wheezy LTS *now*? LTS starts in 3 days. -- cheers, Holger signature.asc Description: Digital signature
Re: Supporting armel/armhf in wheezy-lts
]] Raphael Hertzog I am not speaking on behalf of DSA here. > The LTS period is about to start soon and we will send out an announce > soon... it would be nice to be able to say a word about armel/armhf, so > an official confirmation from ftpmasters/DSA would be nice. So, I'm a bit divided on all this. While the DMUP is heavy artillery and we should be careful about invoking it, I think it crosses the line in «Don't use Debian Facilities for private financial gain or for commercial purposes, including consultancy or any other work outside the scope of official duties or functions for the time being, without specific authorization to do so.», so you need explicit authorization before you start. We have to balance multiple issues here: on one hand, LTS is clearly useful. On the other hand, Debian is a volunteer organisation and we don't pay people to work on Debian. By asking for donations for «Debian LTS» and then paying folks to work on it, we're at least very, very close to that line, if we don't cross it, somewhat depending on who «we» are and how it's marketed and presented, and I think that is not particularly clearly communicated today. JFTR, for me at least, this isn't about wanting a piece of the action; I don't want to be paid for my DSA work. One option I've been toying about with (but which I'm not sure is a good one) is that some portion of the Debian LTS income go to Debian, rather than individuals. -- Tollef Fog Heen UNIX is user friendly, it's just picky about who its friends are
Re: Supporting armel/armhf in wheezy-lts
On Mon, 18 Apr 2016, Raphael Hertzog wrote: > In the mean time, the sponsor clarified that they will join as "gold > sponsor" so they are effectively sponsoring 8 hours of work per month, > which seems to be enough to cover for the increased work that those > ports might represent. The sponsorship is now confirmed since they have paid for a full year already. Plat'Home is the fourth gold LTS sponsor. >From IRC discussions it looks like that the buildd maintainers (at least Neil Williams and Riku Voipio) have no objections to supporting armel/armhf (Riku is rather pleased by it). On the ftpmaster side, I had no official feedback but at least Ansgar seemed to be fine with it when I asked him on IRC. On the Debian Admin side, I have not had any feedback yet. The LTS period is about to start soon and we will send out an announce soon... it would be nice to be able to say a word about armel/armhf, so an official confirmation from ftpmasters/DSA would be nice. Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
Hey guys, Might be a longshot, but you could contact the Scaleway team (scaleway.com). They provide physical arm servers with (among others) Wheezy images, as well as Jessie and Sid images. I couldn't find a direct contact email, but they do have a ticketing system on their website. (or maybe someone here knows someone there?) Floris. -- Stoica-Marcu "FlowRiser" Floris-Andrei floris...@gmail.com
Re: Supporting armel/armhf in wheezy-lts
On Mon, Apr 18, 2016 at 09:44:43AM +0200, Raphael Hertzog wrote: > Hi, > > On Mon, 18 Apr 2016, Guido Günther wrote: > > I assume the level of sponsorship offered is reasonable to support an > > arm port? > > Yes, I believe so. They will join as gold sponsor. > > > I still wonder how we could would make it simpler to have this > > support end up at the right places (i.e. LTS gets the sponsorship while > > other teams like release team, security team also have additional work)? > > What additional work are you referring to specifically? > > The security and release teams are no longer involved in the LTS part of > the work. > > That said if other teams believe that they have extra-work to support > the LTS initiative, I'm entirely open to the idea of funding persons > outside of the LTS team provided that they work benefit the LTS project > and provided that they follow the same rules of transparency and > accountability. With possibly extra-caution like a check with > other members on their team that none opposes to the paid work and > that nobody else would want to do the same work in their free time. We're fixing quiet some issues in Wheezy either via the security team or the release team. All of the work is reviewed by members of these teams. This is work that might not have happened if we weren't to support Wheezy. Also the switch from wheezy on security master will probably involve ftpmaster / security team time. > > > It seems the raspberry pi foundation removed the wheezy based raspbian > > from the download page but I wonder if we should reach out to them since > > I assume there are lots of arm based devices still running wheezy and > > they would certainly benefit from arm support (although they have to > > recompile). Does anybody have any contacts there? > > What would you ask to the raspberry pi foundation? If they want to join as sponsors. Cheers, -- Guido
Re: Supporting armel/armhf in wheezy-lts
Am 18.04.2016 um 08:45 schrieb Guido Günther: [...] > I'm all for it (although it's easy to say for me since the most burden > will probably be on the kernel team) and having it as experimental with > a single sponsor seems sensible. +1 from my side too. I guess I'm one of those armel hobbyists and I could test the software even on real hardware. > I assume the level of sponsorship offered is reasonable to support an > arm port? I still wonder how we could would make it simpler to have this > support end up at the right places (i.e. LTS gets the sponsorship while > other teams like release team, security team also have additional work)? I also think that the opinions of the kernel team / Ben are crucial if we want to support ARM in the future. Otherwise I would expect that supporting ARM scales pretty well and that it mainly requires more time for testing the software. Speaking of the security team and release team, what about the ftp team and the upcoming switch to Wheezy-LTS. Raphaël wrote [1] that the switch would be easy according to one of the ftp masters. Is there anything that we should take care of / coordinate with the ftp team before April, 26th? Regards, Markus [1] https://lists.debian.org/debian-lts/2016/02/msg00130.html signature.asc Description: OpenPGP digital signature
Re: Supporting armel/armhf in wheezy-lts
Hi, On Mon, 18 Apr 2016, Guido Günther wrote: > I assume the level of sponsorship offered is reasonable to support an > arm port? Yes, I believe so. They will join as gold sponsor. > I still wonder how we could would make it simpler to have this > support end up at the right places (i.e. LTS gets the sponsorship while > other teams like release team, security team also have additional work)? What additional work are you referring to specifically? The security and release teams are no longer involved in the LTS part of the work. That said if other teams believe that they have extra-work to support the LTS initiative, I'm entirely open to the idea of funding persons outside of the LTS team provided that they work benefit the LTS project and provided that they follow the same rules of transparency and accountability. With possibly extra-caution like a check with other members on their team that none opposes to the paid work and that nobody else would want to do the same work in their free time. > It seems the raspberry pi foundation removed the wheezy based raspbian > from the download page but I wonder if we should reach out to them since > I assume there are lots of arm based devices still running wheezy and > they would certainly benefit from arm support (although they have to > recompile). Does anybody have any contacts there? What would you ask to the raspberry pi foundation? Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
Hello, On Fri, 15 Apr 2016, Raphael Hertzog wrote: > armhf users to run jessie and armel users to be only hobbyists). But here > we have a sponsor willing to sign on a contract saying us « our customers wish > to use Debian 7 OS continuously so that we are planning to sponsor the LTS > efforts for two years. » In the mean time, the sponsor clarified that they will join as "gold sponsor" so they are effectively sponsoring 8 hours of work per month, which seems to be enough to cover for the increased work that those ports might represent. Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: http://www.freexian.com/services/debian-lts.html Learn to master Debian: http://debian-handbook.info/get/
Re: Supporting armel/armhf in wheezy-lts
(trimming the cc list a bit since this is mostly a +1) On Fri, Apr 15, 2016 at 10:03:45AM +0200, Raphael Hertzog wrote: > Hello, > > I know that we decided to not support arm* for wheezy-lts during last > Debconf but it turns out that Freexian has been contacted by a potential LTS > sponsor selling arm* products: > > Openblocks A7 (armel) > http://openblocks.plathome.co.jp/products/obs_a/a7/spec.html > Opneblocks AX3 (armhf) > http://openblocks.plathome.co.jp/products/obs_a/ax3/ > > The decision has been taken on the fact that we knew of nobody that was > really interested on running wheezy on those architectures (we expected > armhf users to run jessie and armel users to be only hobbyists). But here > we have a sponsor willing to sign on a contract saying us « our customers wish > to use Debian 7 OS continuously so that we are planning to sponsor the LTS > efforts for two years. » > > Are there teams/persons with objections to adding armel/armhf to the set of > supported architectures for wheezy-lts ? > > I have put in copy: > - ftpmasters, they have to agree to keep hosting armel/armhf on > security.debian.org > - ARM porters, mostly to ensure that buildd maintainers are OK with this > - Debian System Administrators as this might have an impact on the lifetime > of the arm systems running the wheezy auto-builders > > Obviously this does have an impact on the work of LTS team members too since > we will have to monitor ARM* builds too and it might mean more work for the > kernel maintenance too. > > Since we have a single sponsor interested in ARM support, maybe we can > introduce > it on a provisional/experimental basis so that we can drop it if the > sponsorship > does not materialize (or is not enough to cover for the supplementary work). > It > might also be an opportunity to make a call for other armel/armhf users to > join > as sponsors (if you know some that we should contact, please let us know). I'm all for it (although it's easy to say for me since the most burden will probably be on the kernel team) and having it as experimental with a single sponsor seems sensible. I assume the level of sponsorship offered is reasonable to support an arm port? I still wonder how we could would make it simpler to have this support end up at the right places (i.e. LTS gets the sponsorship while other teams like release team, security team also have additional work)? It seems the raspberry pi foundation removed the wheezy based raspbian from the download page but I wonder if we should reach out to them since I assume there are lots of arm based devices still running wheezy and they would certainly benefit from arm support (although they have to recompile). Does anybody have any contacts there? Cheers, -- Guido
