Re: *.openmoko.org infrastructure
Am 08.05.2012 um 21:55 schrieb Bob Ham: On Tue, 2012-05-08 at 20:33 +0200, Harald Welte wrote: On Tue, May 08, 2012 at 07:20:36PM +0100, Bob Ham wrote: On Mon, 2012-05-07 at 00:08 +0100, Bob Ham wrote: I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. Harald, Maximilian and Joachim, as the present administrators what do you think about this idea? I think it's an excellent idea. This seems like a good idea, given that there are people who are already taking care of hosting other projects and thus already do quite a bit of sysadmin. So just to be clear on this: From my point of view the problem is not to pay for some hosting/traffic every month. If anything else fails, I'd even pay that out of my own pocket. The main issue is to have somebody who will actually take care of updating the debian packages, run backups, make sure things run smoothly and do a dist-upgrade every couple of years. Indeed, that is the main issue. The way I see it, freeing the community from the burden of administering infrastructure can only be a good thing. I think one point hasn't been addressed much in this discussion so far. How important are all the *.openmoko.org services to us? So I suggest to everyone on this mailing list to give an indication like this: I regard as important: ( ) lists.openmoko.org (incl. the mailing lists) ( ) wiki.openmoko.org ( ) docs.openmoko.org ( ) svn.openmoko.org ( ) people.openmoko.org ( ) projects.openmoko.org ( ) Others: __ This may shed a new light on this topic. Nikolaus PS: looking the mailing lists I have seen only three with major activity: 1. this list (community) 2. devel (which is more or less a dfu-util discussion list) 3. openmoko-kernel ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re : *.openmoko.org infrastructure
Order of importance of subdomains of openmoko.org is not the key point. I take notice that we can rely on those partners you named, but still we must find someone to manage contents. I don't know much of the publishing system, but may have a closer look at it if you want (if PHP or CGI it could be ok). Also, take good notice I'm coming in Berlin from 22 to 28 of May, and hope to see you there. I will have suggestions and pro-active routes to develop the Openmoko project. There is a document (soon to be translated in English) describing my intentions at : https://docs.google.com/open?id=0BwsNltshAe2hMzZOa1l0V3ozdFU It's Google doc, so if you want things to be discussed more privately, tell me. Regards, -- Guillaume Farny CCAS de Bagnolet Place Salvador Allende 93170 Bagnolet FRANCE 06.16.52.42.32 farny.guilla...@yahoo.fr - Mail original - De : Dr. H. Nikolaus Schaller h...@goldelico.com À : List for Openmoko community discussion community@lists.openmoko.org Cc : Maximilian Bauer gi...@openmoko.org; Michael 'Mickey' Lauer mic...@openmoko.org; Paul Fertser fercer...@gmail.com; Harald Welte lafo...@gnumonks.org; Joachim Steiger r...@openmoko.org; Joerg Reisenweber jo...@openmoko.org Envoyé le : Jeudi 17 mai 2012 13h49 Objet : Re: *.openmoko.org infrastructure Am 08.05.2012 um 21:55 schrieb Bob Ham: On Tue, 2012-05-08 at 20:33 +0200, Harald Welte wrote: On Tue, May 08, 2012 at 07:20:36PM +0100, Bob Ham wrote: On Mon, 2012-05-07 at 00:08 +0100, Bob Ham wrote: I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. Harald, Maximilian and Joachim, as the present administrators what do you think about this idea? I think it's an excellent idea. This seems like a good idea, given that there are people who are already taking care of hosting other projects and thus already do quite a bit of sysadmin. So just to be clear on this: From my point of view the problem is not to pay for some hosting/traffic every month. If anything else fails, I'd even pay that out of my own pocket. The main issue is to have somebody who will actually take care of updating the debian packages, run backups, make sure things run smoothly and do a dist-upgrade every couple of years. Indeed, that is the main issue. The way I see it, freeing the community from the burden of administering infrastructure can only be a good thing. I think one point hasn't been addressed much in this discussion so far. How important are all the *.openmoko.org services to us? So I suggest to everyone on this mailing list to give an indication like this: I regard as important: ( ) lists.openmoko.org (incl. the mailing lists) ( ) wiki.openmoko.org ( ) docs.openmoko.org ( ) svn.openmoko.org ( ) people.openmoko.org ( ) projects.openmoko.org ( ) Others: __ This may shed a new light on this topic. Nikolaus PS: looking the mailing lists I have seen only three with major activity: 1. this list (community) 2. devel (which is more or less a dfu-util discussion list) 3. openmoko-kernel ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
Do 17. Mai 2012 So I suggest to everyone on this mailing list to give an indication like this: FIRST: look at my mail addr! Guess where it's hosted. SECOND: there's quite some stuff in /home/* and other places of several of those servers, that's definitely not meant to go public. There are internal ML, datasheets and even sources under NDA, whatnot. I don't like this to get handed out to random admin folks. Sean as well will rise objections (if he's aware of the problem) [Sean, you know I always been one of the most vivid proponents of disclosing *everything* to public, but there are definitely both legal and privacy limitations in doing so here. Harald's assumption that stuff is nicely separated on several vservers might not hold true, e.g already passwords for wiki user accounts must not get disclosed. I'm extremely sceptic about the direction this thing moves here] THIRD: svn.openmoko.org has been missing since weeks now, and that's where the whole hassle started. So assuming we can solve the problem by shutting down even more of the infra is the wrong angle to tackle the issue. Even if we find ${WHATEVER}.openmoko.org is a service not actually used anymore, even if we spot 80% of the ML being zombie state, we still need infra for the remaining stuff that's evidently used. So no use in such a poll right now - distracting. /j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
Hi, Am 17.05.2012 um 15:52 schrieb Joerg Reisenweber: Do 17. Mai 2012 So I suggest to everyone on this mailing list to give an indication like this: FIRST: look at my mail addr! Guess where it's hosted. Yes, I know. But there are estimatedly 100k other addresses to host a (private) mailbox. And, how many people are using this service? I doubt that there are more than 5... SECOND: there's quite some stuff in /home/* and other places of several of those servers, that's definitely not meant to go public. There are internal ML, datasheets and even sources under NDA, whatnot. IMHO, that should be moved to openmoko.com if it is owned by Openmoko, Inc. I don't like this to get handed out to random admin folks. Sean as well will I 150% agree for sensible information. Nobody would trust us any more if that happens. rise objections (if he's aware of the problem) [Sean, you know I always been one of the most vivid proponents of disclosing *everything* to public, but there are definitely both legal and privacy limitations in doing so here. Harald's assumption that stuff is nicely separated on several vservers might not hold true, e.g already passwords for wiki user accounts must not get disclosed. I'm extremely sceptic about the direction this thing moves here] THIRD: svn.openmoko.org has been missing since weeks now, and that's where the I am just wondering who is using the SVN? If not the community, who else? whole hassle started. So assuming we can solve the problem by shutting down even more of the infra is the wrong angle to tackle the issue. Even if we find It would simplify things if areas that nobody is using any more can be closed. ${WHATEVER}.openmoko.org is a service not actually used anymore, even if we spot 80% of the ML being zombie state, we still need infra for the remaining stuff that's evidently used. So no use in such a poll right now - distracting. My observation (as a pure voluntary community member and not related at all to Openmoko, Inc.) is that there isn't much being evidently used. But I don't claim to know everything: therefore the poll is useful. BR, Nikolaus BTW: there is also git.openmoko.org which is up and running. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
Do 17. Mai 2012 Hi, Am 17.05.2012 um 15:52 schrieb Joerg Reisenweber: Do 17. Mai 2012 So I suggest to everyone on this mailing list to give an indication like this: FIRST: look at my mail addr! Guess where it's hosted. Yes, I know. But there are estimatedly 100k other addresses to host a (private) mailbox. That's not the point though. I'm already preparing to move my mail addr somewhere else, resp propagate another main mail addr. This will take some time. But the main concern is my *private* mails are on those servers! And, how many people are using this service? I doubt that there are more than 5... Maximum, or less. See above though. SECOND: there's quite some stuff in /home/* and other places of several of those servers, that's definitely not meant to go public. There are internal ML, datasheets and even sources under NDA, whatnot. IMHO, that should be moved to openmoko.com if it is owned by Openmoko, Inc. There's no such thing like openmoko.com anymore for such stuff, afaik. Anyway, again not the point. *WHO* is doing that move/cleanup? I don't like this to get handed out to random admin folks. Sean as well will I 150% agree for sensible information. Nobody would trust us any more if that happens. Exactly! THIRD: svn.openmoko.org has been missing since weeks now, and that's where the I am just wondering who is using the SVN? If not the community, who else? it's always the community ;-D For svn there is a number of indispensable stuff in there, e.g. all the nifty tools of Werner Almesberger. He recently noticed Dang, svn down! /join #qi- hardware and ask him, see chanlogs. whole hassle started. So assuming we can solve the problem by shutting down even more of the infra is the wrong angle to tackle the issue. Even if we find It would simplify things if areas that nobody is using any more can be closed. Sure thing. But introducing the process of finding out _what_ can get shut down now is actually temporarily complicating things a lot. ${WHATEVER}.openmoko.org is a service not actually used anymore, even if we spot 80% of the ML being zombie state, we still need infra for the remaining stuff that's evidently used. So no use in such a poll right now - distracting. My observation (as a pure voluntary community member and not related at all to Openmoko, Inc.) is that there isn't much being evidently used. But I don't claim to know everything: therefore the poll is useful. The poll will again just reach those that are aware of it. So you don't find out good data about who's really using the stuff. Heck, are we discussing *deletion of data* here? You know the internet isn't supposed to ever forget a singly byte ;-D. We should try to make this urban legend hold true, at least for OM. E.G. I'm not 'using' *anything* of the svn et al stuff right now, nevertheless I think I don't have a local copy of uBoot-dirtymoko and Werner's nand-tools to flash it to a device in a year or two, when I come up with some fancy project I want to do with these devices in my shelf. /j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
On Mon, 2012-05-07 at 00:08 +0100, Bob Ham wrote: I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. They're open to the idea of migrating openmoko.org services and/or server VMs to their infrastructure. I believe this would be the most sustainable and practical solution. Harald, Maximilian and Joachim, as the present administrators what do you think about this idea? -- Bob Ham r...@settrans.net for (;;) { ++pancakes; } signature.asc Description: This is a digitally signed message part ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
Hi Bob, On Tue, May 08, 2012 at 07:20:36PM +0100, Bob Ham wrote: On Mon, 2012-05-07 at 00:08 +0100, Bob Ham wrote: I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. Harald, Maximilian and Joachim, as the present administrators what do you think about this idea? I think it's an excellent idea. This seems like a good idea, given that there are people who are already taking care of hosting other projects and thus already do quite a bit of sysadmin. So just to be clear on this: From my point of view the problem is not to pay for some hosting/traffic every month. If anything else fails, I'd even pay that out of my own pocket. The main issue is to have somebody who will actually take care of updating the debian packages, run backups, make sure things run smoothly and do a dist-upgrade every couple of years. But in the end, it is Sean that you have to convince, not me ;) Regads, Harald -- - Harald Welte lafo...@gnumonks.org http://laforge.gnumonks.org/ Privacy in residential applications is a desirable marketing option. (ETSI EN 300 175-7 Ch. A6) ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
On Tue, 2012-05-08 at 20:33 +0200, Harald Welte wrote: On Tue, May 08, 2012 at 07:20:36PM +0100, Bob Ham wrote: On Mon, 2012-05-07 at 00:08 +0100, Bob Ham wrote: I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. Harald, Maximilian and Joachim, as the present administrators what do you think about this idea? I think it's an excellent idea. This seems like a good idea, given that there are people who are already taking care of hosting other projects and thus already do quite a bit of sysadmin. So just to be clear on this: From my point of view the problem is not to pay for some hosting/traffic every month. If anything else fails, I'd even pay that out of my own pocket. The main issue is to have somebody who will actually take care of updating the debian packages, run backups, make sure things run smoothly and do a dist-upgrade every couple of years. Indeed, that is the main issue. The way I see it, freeing the community from the burden of administering infrastructure can only be a good thing. But in the end, it is Sean that you have to convince, not me ;) Sean, what do you think? -- Bob Ham r...@settrans.net for (;;) { ++pancakes; } signature.asc Description: This is a digitally signed message part ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
roh, gismo and myself (at least) have root on all the boxes. There is no prescribed contact procedure, and there is nobody who is even officially supposed to maintain the machines and/or the software and configurations on them. roh and gismo used to be doing paid syssadmin work for Openmoko Inc., until the company pulled out of mobile phones. After that, roh and gismo have still been able to continuing to donate some of their spare time to it, but with a lot of other involvements, the amount of time available for that is probably just the bare minimum to keep things alive, and not any major migrations. Hi, I am system administrator and OpenMoko/GTA02 user/developer. I am working in Magento inc. and my primary job is web and related areas. I have experience with RHEL/Debian/FreeBSD/OpenSolaris. Also i am participating in some Open Source projects, e.g. smartmontools or FreeBSD. I would be glad to help with system administration of OpenMoko servers, because i am very interested in this project and want to help it somehow. If you need more information about me - please contact me directly, using samm [at] os2.kiev.ua address. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
*.openmoko.org infrastructure
Hello, There was considerable tension in the community lately regarding the Openmoko community resources, namely docs.openmoko.org , wiki.openwoko.org , svn.openmoko.org , people.openmoko.org . The problem is that not only the servers were unstable and apparently required a kick every now and then but also that there's no clearly visible contact person and the whole situation regarding the resources is causing honest concern and inconvenience in community. Right at the moment svn.openmoko.org and docs.openmoko.org are still inaccessible (since weeks) and nobody really knows how to proceed. For the community to be able to deduce a way to manage this situation i would like to ask everybody involved to clarify these questions: 1. Who owns the servers and who pays which bills? How to get a proxy or transfer the responsibility for the openmoko.org domain and infrastructure to other person in case the original owner becomes unavailable? Whom can you propose to take this role? 2. Who actually has admin credentials, is supposed to care about things like server outages, and what is the suggested procedure to contact that person? Is there a proxy for this position? How to transfer the credentials to another person in case the original maintainer becomes unavailable? 3. What is the practical and reliable way to do administrative tasks like e.g. obtaining the dump of the wiki and docs.openmoko.org bugtracker database? 4. What suggestions do you have to improve the status quo? Thanks in advance for your replies and commitment. Sincerely, Paul Fertser -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
Hi Paul and others, On Mon, May 07, 2012 at 12:53:15AM +0400, Paul Fertser wrote: For the community to be able to deduce a way to manage this situation i would like to ask everybody involved to clarify these questions: 1. Who owns the servers and who pays which bills? How to get a proxy or transfer the responsibility for the openmoko.org domain and infrastructure to other person in case the original owner becomes unavailable? Whom can you propose to take this role? The servers are rented from Hetzner in Germany, the bills are paid by Sean Moss-Pultz / Openmoko Inc. 2. Who actually has admin credentials, is supposed to care about things like server outages, and what is the suggested procedure to contact that person? roh, gismo and myself (at least) have root on all the boxes. There is no prescribed contact procedure, and there is nobody who is even officially supposed to maintain the machines and/or the software and configurations on them. roh and gismo used to be doing paid syssadmin work for Openmoko Inc., until the company pulled out of mobile phones. After that, roh and gismo have still been able to continuing to donate some of their spare time to it, but with a lot of other involvements, the amount of time available for that is probably just the bare minimum to keep things alive, and not any major migrations. As it is a considerably complex setup involving about a dozen of virtual machines on outdated Linux distributions (without security updates), maintenance probably is a nightmare. As it seems, while all of us still feel extremely sympathetic about Openmoko, none of the poeple with admin privileges are still actively invovled with the Openmoko community. If I wasn't involved in a gazillion of other projects and constantly overworked and behind any possible schedule, I would even jump in myself ot help. But I simply can't :( Is there a proxy for this position? How to transfer the credentials to another person in case the original maintainer becomes unavailable? 4. What suggestions do you have to improve the status quo? From my point of view, in order to proceed, what would need to be done is: 1) to remove/split any data that is openmoko proprietary/internal from the public information. Traditionally, there was a clean separation of such data in different virtual machines, so I think this is more or less a hypothetical problem. Does any of the ex-openmoko guys disagree? 2) finding somebody with the required skills, dedication and trust by all involved entities/people to actually maintain/upgrade those systems for the time to come. 3) ensure that Openmoko, Inc. still has an adequate amount of last resort control over what happens there, as it (after all) still is their trademark. So if somebody sues them over content on those sites, they need to be able to take action by removing such content without having to go to anyone else. 4) migrate the VMs or at least the content to new, more reliable (and also less expensive) rented server hardware. I would like to use this opportunity to thank Sean for continuing to fund the server hosting / traffic, and roh + gismo for what they have done so far in keeping things alive. Regards, Harald -- - Harald Welte lafo...@gnumonks.org http://laforge.gnumonks.org/ Privacy in residential applications is a desirable marketing option. (ETSI EN 300 175-7 Ch. A6) ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: *.openmoko.org infrastructure
On Sun, 2012-05-06 at 23:23 +0200, Harald Welte wrote: From my point of view, in order to proceed, what would need to be done is: 2) finding somebody with the required skills, dedication and trust by all involved entities/people to actually maintain/upgrade those systems for the time to come. 3) ensure that Openmoko, Inc. still has an adequate amount of last resort control over what happens there, as it (after all) still is their trademark. So if somebody sues them over content on those sites, they need to be able to take action by removing such content without having to go to anyone else. 4) migrate the VMs or at least the content to new, more reliable (and also less expensive) rented server hardware. I've spoken through IRC to some of the guys at the Open Source Lab at Oregon State University¹. This group hosts, for example, the Meego project, as well as helping host parts of the Apache, Debian, CentOS and Gentoo infrastructures, to name a few. They're open to the idea of migrating openmoko.org services and/or server VMs to their infrastructure. I believe this would be the most sustainable and practical solution. Bob Ham ¹ http://osuosl.org/about-osuosl -- Bob Ham r...@settrans.net for (;;) { ++pancakes; } signature.asc Description: This is a digitally signed message part ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community