Re: [fedora-arm] Who's using Kirkwood?
On 10/10/2012 05:55 PM, Derek Atkins wrote: Hi Folks, I'm interested to know who is using Kirkwood, and who would miss it if it went away. For now, we won't kill off ARMv5 because it is used in the official rPi builds but that doesn't mean I'm not interested to know whether we should put testing effort into Kirkwood for F18. My thought is that the latest plugs are moving to ARMv7, and so as the cutting edge Linux distro, we should make plans for deprecating support over the coming releases. This is not a call to drop support today. If I can get numbers on how many people care, that will help. All my Arm devices are Kirkwoods, including Sheeva and Guru Plug devices, and I was considering acquiring some Dreamplug devices, too. I use them in production (with Fedora), and honestly I'd feel very put out if Fedora dropped support for them. I know a bunch of other people who have other kirkwood devices, too. If you read the full thread it's not about dropping the support in the short term. I did read the thread, but our definitions of "short term" appear to be different. The thread appeared to be a question of support for F18 or F19. IMNSHO I feel Kirkwood support should probably remain until, oh, F25 or 26, at a minimum. There are just too many (IMHO) Kirkwoods out in production. More to the point, they are still being made and sold in reasonable quantity. I know that RPi looks interesting, but they are still very hard to acquire. (Limit 1, then wait a few months??) That's no longer the case. In most cases I believe it should now be relatively instant shipping and they're certainly no longer limited to single unit. Glad to hear that. However I'm loathe to throw away my investment of Kirkwoods. I cannot answer you how many others bought them. Have you tried asking them for approximate numbers? 512MB of usable RAM on a SheevaPlug is also a lot easier to live with than 192MB of usable RAM on the Pi. If the VIA APC was cited as an alternative, then maybe I could almost get behind that in due course (512MB of RAM, *TX form factor). But running one of the default desktop environments with a browser that actually works reasonably well for most commonly used websites (i.e. not Midori) in 192MB of RAM? While swapping to an average SD card? Do be serious. The x86 port still supports a Pentium, I don't see any reason to drop support for kirkwood. Is it really that much extra effort? It is surprisingly quite a lot of effort. Oh? Could you elaborate on that? What "quite a lot of effort" does it take? From my experience of rolling a similar distribution, if the kernel code works as it's supposed to, a day or so of tweaking the configs, followed by about a day of compiling (in a 1.2GHz Kirkwood). If there are issues? Much longer because the compile takes so long. Fedora no longer supports Pentium actually. It was dropped some time ago (around Fedora 12 from memory). The lowest level of support in Fedora for x86 is now Pentium Pro (Basically i586 + CMOV) which allows support for the OLPC XO-1 (AMD Geode Processor) and the only reason it's still at that level is because there's around 1.5 million XO-1 united deployed and still be actively used and upgraded to current Fedora releases (The just released 12.1.0 is based on Fedora 17, the under development 13.1.0 release is based on Fedora 18). I know mainline Fedora would like to drop the support for that too if they could. So what you're saying is that Fedora *still* supports an x32 CPU that was released well over a decade ago... The important point to be made is that both Kirkwood and i686 class machines are still in production and available to buy new today. Gordan ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On 10/10/2012 06:47 PM, Derek Atkins wrote: Peter Robinson writes: Might as well wait until the whole 32-bit branch can be dropped. Practically all x86 CPU made in most of the past decade is x86-64. Half decade maybe as Intel first introduced 64 bit CPUs in early 2005 and it took a while to spread through their product set, and there was a lot of Atom CPUs that weren't 64 bit capable. But I agree the reasons for 32 is slowly receding. Sure, but we're a decade later. Kirkwood devices were just released what? 3 years ago? I certainly got mine more recently than that. DreamPlug (v1) was only released about 18 months ago. Gordan ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On Thu, Oct 11, 2012 at 10:42 AM, Gordan Bobic wrote: > On 10/10/2012 05:55 PM, Derek Atkins wrote: > > Hi Folks, > > I'm interested to know who is using Kirkwood, and who would miss it if > it went away. For now, we won't kill off ARMv5 because it is used in > the > official rPi builds but that doesn't mean I'm not interested to know > whether we should put testing effort into Kirkwood for F18. > > My thought is that the latest plugs are moving to ARMv7, and so as the > cutting edge Linux distro, we should make plans for deprecating support > over the coming releases. This is not a call to drop support today. If > I > can get numbers on how many people care, that will help. All my Arm devices are Kirkwoods, including Sheeva and Guru Plug devices, and I was considering acquiring some Dreamplug devices, too. I use them in production (with Fedora), and honestly I'd feel very put out if Fedora dropped support for them. I know a bunch of other people who have other kirkwood devices, too. >>> >>> >>> If you read the full thread it's not about dropping the support in the >>> short term. >> >> >> I did read the thread, but our definitions of "short term" appear to be >> different. The thread appeared to be a question of support for F18 or >> F19. IMNSHO I feel Kirkwood support should probably remain until, oh, >> F25 or 26, at a minimum. There are just too many (IMHO) Kirkwoods out >> in production. > > > More to the point, they are still being made and sold in reasonable > quantity. > > I know that RPi looks interesting, but they are still very hard to acquire. (Limit 1, then wait a few months??) >>> >>> >>> That's no longer the case. In most cases I believe it should now be >>> relatively instant shipping and they're certainly no longer limited to >>> single unit. >> >> >> Glad to hear that. However I'm loathe to throw away my investment of >> Kirkwoods. I cannot answer you how many others bought them. Have you >> tried asking them for approximate numbers? > > > 512MB of usable RAM on a SheevaPlug is also a lot easier to live with than > 192MB of usable RAM on the Pi. > > If the VIA APC was cited as an alternative, then maybe I could almost get > behind that in due course (512MB of RAM, *TX form factor). But running one > of the default desktop environments with a browser that actually works > reasonably well for most commonly used websites (i.e. not Midori) in 192MB > of RAM? While swapping to an average SD card? Do be serious. I've never said 192Mb of RAM is reasonable so I think you'll find I'm completely serious, but then neither is 512Mb. With devices like the cubieboard, gooseberry, wandboard and numerous others coming out with 1Gb of RAM I personally don't see the kirkwood nor the RPi as any for of serious. What's more the cubieboard will be only $14 more than the RPi. The x86 port still supports a Pentium, I don't see any reason to drop support for kirkwood. Is it really that much extra effort? >>> >>> >>> It is surprisingly quite a lot of effort. >> >> >> Oh? Could you elaborate on that? What "quite a lot of effort" does it >> take? > > > From my experience of rolling a similar distribution, if the kernel code > works as it's supposed to, a day or so of tweaking the configs, followed by > about a day of compiling (in a 1.2GHz Kirkwood). > > If there are issues? Much longer because the compile takes so long. I don't have 2 days to spare to deal with that. If someone else does that is absolutely fabulous. I'm yet to see them actually step up to the plate and do the work. Clearly you're not interested in doing any work what so ever, I've not actually seen a contribution from you at all. >>> Fedora no longer supports Pentium actually. It was dropped some time >>> ago (around Fedora 12 from memory). The lowest level of support in >>> Fedora for x86 is now Pentium Pro (Basically i586 + CMOV) which allows >>> support for the OLPC XO-1 (AMD Geode Processor) and the only reason >>> it's still at that level is because there's around 1.5 million XO-1 >>> united deployed and still be actively used and upgraded to current >>> Fedora releases (The just released 12.1.0 is based on Fedora 17, the >>> under development 13.1.0 release is based on Fedora 18). I know >>> mainline Fedora would like to drop the support for that too if they >>> could. >> >> >> So what you're saying is that Fedora *still* supports an x32 CPU that >> was released well over a decade ago... > > > The important point to be made is that both Kirkwood and i686 class machines > are still in production and available to buy new today. You've made that point and the point that I've made numerous times is the decision isn't being made today so it's somewhat of a mute point. Peter ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On 10/11/2012 10:51 AM, Peter Robinson wrote: On Thu, Oct 11, 2012 at 10:42 AM, Gordan Bobic wrote: On 10/10/2012 05:55 PM, Derek Atkins wrote: Hi Folks, I'm interested to know who is using Kirkwood, and who would miss it if it went away. For now, we won't kill off ARMv5 because it is used in the official rPi builds but that doesn't mean I'm not interested to know whether we should put testing effort into Kirkwood for F18. My thought is that the latest plugs are moving to ARMv7, and so as the cutting edge Linux distro, we should make plans for deprecating support over the coming releases. This is not a call to drop support today. If I can get numbers on how many people care, that will help. All my Arm devices are Kirkwoods, including Sheeva and Guru Plug devices, and I was considering acquiring some Dreamplug devices, too. I use them in production (with Fedora), and honestly I'd feel very put out if Fedora dropped support for them. I know a bunch of other people who have other kirkwood devices, too. If you read the full thread it's not about dropping the support in the short term. I did read the thread, but our definitions of "short term" appear to be different. The thread appeared to be a question of support for F18 or F19. IMNSHO I feel Kirkwood support should probably remain until, oh, F25 or 26, at a minimum. There are just too many (IMHO) Kirkwoods out in production. More to the point, they are still being made and sold in reasonable quantity. I know that RPi looks interesting, but they are still very hard to acquire. (Limit 1, then wait a few months??) That's no longer the case. In most cases I believe it should now be relatively instant shipping and they're certainly no longer limited to single unit. Glad to hear that. However I'm loathe to throw away my investment of Kirkwoods. I cannot answer you how many others bought them. Have you tried asking them for approximate numbers? 512MB of usable RAM on a SheevaPlug is also a lot easier to live with than 192MB of usable RAM on the Pi. If the VIA APC was cited as an alternative, then maybe I could almost get behind that in due course (512MB of RAM, *TX form factor). But running one of the default desktop environments with a browser that actually works reasonably well for most commonly used websites (i.e. not Midori) in 192MB of RAM? While swapping to an average SD card? Do be serious. I've never said 192Mb of RAM is reasonable so I think you'll find I'm completely serious, but then neither is 512Mb. With devices like the cubieboard, gooseberry, wandboard and numerous others coming out with 1Gb of RAM I personally don't see the kirkwood nor the RPi as any for of serious. What's more the cubieboard will be only $14 more than the RPi. Two points: 1) If that's what you think, I'd really like to stop seeing the Pi as an excuse for dropping or including anything and pandering to it. 2) 500MB-ish of RAM is actually enough for a decent user experience. I am a daily user of a Toshiba AC100, and use it daily with KDE as my desktop environment and Firefox as my browser. With 480MB of RAM, the experience is comfortable. With a few tweaks the experience stretches to pleasant: http://www.altechnative.net/2012/01/04/alleviating-memory-pressure-on-toshiba-ac100/ The x86 port still supports a Pentium, I don't see any reason to drop support for kirkwood. Is it really that much extra effort? It is surprisingly quite a lot of effort. Oh? Could you elaborate on that? What "quite a lot of effort" does it take? From my experience of rolling a similar distribution, if the kernel code works as it's supposed to, a day or so of tweaking the configs, followed by about a day of compiling (in a 1.2GHz Kirkwood). If there are issues? Much longer because the compile takes so long. I don't have 2 days to spare to deal with that. If someone else does that is absolutely fabulous. I'm yet to see them actually step up to the plate and do the work. Clearly you're not interested in doing any work what so ever, I've not actually seen a contribution from you at all. I've had an issue with the attitude for pursuing the bleeding edge in Fedora for a while - that's why I decided to roll a different distribution. When most of your bug reports expire due to the release running EOL it rather puts a downer on the motivation to bother contributing with the goal posts moving so fast at the expense of stability. Gordan ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On Thu, Oct 11, 2012 at 10:59 AM, Gordan Bobic wrote: > On 10/11/2012 10:51 AM, Peter Robinson wrote: >> >> On Thu, Oct 11, 2012 at 10:42 AM, Gordan Bobic wrote: >>> >>> On 10/10/2012 05:55 PM, Derek Atkins wrote: >>> >>> Hi Folks, >>> >>> I'm interested to know who is using Kirkwood, and who would miss it >>> if >>> it went away. For now, we won't kill off ARMv5 because it is used in >>> the >>> official rPi builds but that doesn't mean I'm not interested to know >>> whether we should put testing effort into Kirkwood for F18. >>> >>> My thought is that the latest plugs are moving to ARMv7, and so as >>> the >>> cutting edge Linux distro, we should make plans for deprecating >>> support >>> over the coming releases. This is not a call to drop support today. >>> If >>> I >>> can get numbers on how many people care, that will help. >> >> >> >> All my Arm devices are Kirkwoods, including Sheeva and Guru Plug >> devices, and I was considering acquiring some Dreamplug devices, too. >> I >> use them in production (with Fedora), and honestly I'd feel very put >> out >> if Fedora dropped support for them. I know a bunch of other people >> who >> have other kirkwood devices, too. > > > > If you read the full thread it's not about dropping the support in the > short term. I did read the thread, but our definitions of "short term" appear to be different. The thread appeared to be a question of support for F18 or F19. IMNSHO I feel Kirkwood support should probably remain until, oh, F25 or 26, at a minimum. There are just too many (IMHO) Kirkwoods out in production. >>> >>> >>> >>> More to the point, they are still being made and sold in reasonable >>> quantity. >>> >>> >> I know that RPi looks interesting, but they are still very hard to >> acquire. (Limit 1, then wait a few months??) > > > > That's no longer the case. In most cases I believe it should now be > relatively instant shipping and they're certainly no longer limited to > single unit. Glad to hear that. However I'm loathe to throw away my investment of Kirkwoods. I cannot answer you how many others bought them. Have you tried asking them for approximate numbers? >>> >>> >>> >>> 512MB of usable RAM on a SheevaPlug is also a lot easier to live with >>> than >>> 192MB of usable RAM on the Pi. >>> >>> If the VIA APC was cited as an alternative, then maybe I could almost get >>> behind that in due course (512MB of RAM, *TX form factor). But running >>> one >>> of the default desktop environments with a browser that actually works >>> reasonably well for most commonly used websites (i.e. not Midori) in >>> 192MB >>> of RAM? While swapping to an average SD card? Do be serious. >> >> >> I've never said 192Mb of RAM is reasonable so I think you'll find I'm >> completely serious, but then neither is 512Mb. With devices like the >> cubieboard, gooseberry, wandboard and numerous others coming out with >> 1Gb of RAM I personally don't see the kirkwood nor the RPi as any for >> of serious. What's more the cubieboard will be only $14 more than the >> RPi. > > > Two points: > 1) If that's what you think, I'd really like to stop seeing the Pi as an > excuse for dropping or including anything and pandering to it. Believe me I'm not pandering to the RPi _AT_ALL_ so again your point is completely boundless and useless. > 2) 500MB-ish of RAM is actually enough for a decent user experience. I am a > daily user of a Toshiba AC100, and use it daily with KDE as my desktop > environment and Firefox as my browser. With 480MB of RAM, the experience is > comfortable. With a few tweaks the experience stretches to pleasant: > http://www.altechnative.net/2012/01/04/alleviating-memory-pressure-on-toshiba-ac100/ Great! We're not talking about dropping support for the AC100. I have one as well that one day I'll get the time to configure to my liking. > >> The x86 port still supports a Pentium, I don't see any reason to drop >> support for kirkwood. Is it really that much extra effort? > > > > It is surprisingly quite a lot of effort. Oh? Could you elaborate on that? What "quite a lot of effort" does it take? >>> >>> >>> >>> From my experience of rolling a similar distribution, if the kernel code >>> works as it's supposed to, a day or so of tweaking the configs, followed >>> by >>> about a day of compiling (in a 1.2GHz Kirkwood). >>> >>> If there are issues? Much longer because the compile takes so >>> long.bich.net> wrote: > On 10/11/2012 10:51 AM, Peter Robinson wrote: >> >> >> I don't have 2 days to spare to deal with that. If someone else does >> that is absolutely fabulous. I'm yet to see them actually step up to >> the plate and do the work. Clearly you're not interested in doing any >> wo
Re: [fedora-arm] Who's using Kirkwood?
On 10/10/2012 09:46 PM, Peter Robinson wrote: I know that RPi looks interesting, but they are still very hard to acquire. (Limit 1, then wait a few months??) That's no longer the case. In most cases I believe it should now be relatively instant shipping and they're certainly no longer limited to single unit. Glad to hear that. However I'm loathe to throw away my investment of Kirkwoods. I cannot answer you how many others bought them. Have you tried asking them for approximate numbers? Marvell? Asking who in particular? And what configuration. There's a lot of kirkwood chips with 128Mb or less RAM which makes it a little pointless for a Fedora image and hence IMO not relevant. Asking Globalscale how many Kirkwood *Plugs they sold might be a good start. Fedora no longer supports Pentium actually. It was dropped some time ago (around Fedora 12 from memory). The lowest level of support in Fedora for x86 is now Pentium Pro (Basically i586 + CMOV) which allows support for the OLPC XO-1 (AMD Geode Processor) and the only reason it's still at that level is because there's around 1.5 million XO-1 united deployed and still be actively used and upgraded to current Fedora releases (The just released 12.1.0 is based on Fedora 17, the under development 13.1.0 release is based on Fedora 18). I know mainline Fedora would like to drop the support for that too if they could. So what you're saying is that Fedora *still* supports an x32 CPU that was released well over a decade ago... No. The XO-1 was released in 2007. That's half a decade ago. Given the project came out of MIT and you have a @mit.edu address I would hope you would be able to count, are you in politics by chance? I'm pretty sure that sort of attitude isn't doing much to attract users to the community. DreamPlug1 is Kirkwood based and it only hit the shelves less than 18 months ago. Gordan ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On 10/10/2012 07:59 PM, Brendan Conoboy wrote: On 10/10/2012 10:47 AM, Derek Atkins wrote: Sure, but we're a decade later. Kirkwood devices were just released what? 3 years ago? I certainly got mine more recently than that. I admit I've been running F12 on it, but that's only because there hadn't been another fedora release until F17. The comparison to i686 isn't really very apt. Kirkwood is more like i386, but even that's stretching the simile. There several problems with armv5tel support over the long term. 1. It's not self hosting. We have to use armv7 hosts to build most of the armv5 packages because only they have enough RAM, enough CPU time, fast enough swap. Building UP packages on SMP systems causes issues for a number of multithreaded packages. Transient failures, "bugs" that aren't really bugs, just packages written in the belief that armv5 code will be built and run on armv5 hosts. This problem gets worse with every release. Just out of interest, which packages are you referring to? I am assuming it is LibreOffice + a small subset of whatever is in Fedora that isn't in EL; mainly because I had no RAM/swap/CPU issues building any the 2000 or so packages that overlap. Takes about 3-4 weeks on a _single_ SheevaPlug. 3. Certain features such as atomic operations aren't available on armv5, reducing the number of packages that can be built for ARM in total: If it fails on armv5 but works on armv7, we still don't get it for armv7. In _most_ packages that require this, there are patches that address it. 5. On the whole, it's not a popular Fedora ARM target. Raspberry pi, OMAP, highbank, this is where most (not all) of our known users have hardware and interest. There are some Kirkwood users, clearly, but there are a lot more users of everything else. We should get some updated download stats on this to demonstrate, but last I saw kirkwood was maybe 3% of usage. Perhaps a poll might be a good way to ascertain this, rather than a discussion? Gordan ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On 10/11/2012 11:03 AM, Peter Robinson wrote: The x86 port still supports a Pentium, I don't see any reason to drop support for kirkwood. Is it really that much extra effort? It is surprisingly quite a lot of effort. Oh? Could you elaborate on that? What "quite a lot of effort" does it take? From my experience of rolling a similar distribution, if the kernel code works as it's supposed to, a day or so of tweaking the configs, followed by about a day of compiling (in a 1.2GHz Kirkwood). If there are issues? Much longer because the compile takes so long. I don't have 2 days to spare to deal with that. If someone else does that is absolutely fabulous. I'm yet to see them actually step up to the plate and do the work. Clearly you're not interested in doing any work what so ever, I've not actually seen a contribution from you at all. I've had an issue with the attitude for pursuing the bleeding edge in Fedora for a while - that's why I decided to roll a different distribution. That's fine, you're free to take your toys along with your opinions and play in what ever sand pit you wish. Sure. I'm also happy to also invite others who like my sandpit better to come play in it, too. When most of your bug reports expire due to the release running EOL it rather puts a downer on the motivation to bother contributing with the goal posts moving so fast at the expense of stability. Do your bugs get fixed any quicker in your sandpit? No, unless you fix them yourself. Same outcome really! Sort of - but at least it removes the pretense that the distro maintainers care. ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
[fedora-arm] Daily Koji Compare Stats
Thu Oct 11 09:05:01 EDT 2012 f17-updates : arm vs PA Same |Newer |Older |Local | Remote | Missing | -- 2523 |2 | 41 |1 | 242 | 49 | http://142.204.133.82/jon/koji/kc.f17-updates.diff.html f18 : arm vs PA Same |Newer |Older |Local | Remote | Missing | -- 11810 |6 | 236 |1 | 309 | 324 | http://142.204.133.82/jon/koji/kc.f18.diff.html f18-updates-testing : arm vs PA Same |Newer |Older |Local | Remote | Missing | -- 1702 |3 | 141 |1 | 150 | 150 | http://142.204.133.82/jon/koji/kc.f18-updates-testing.diff.html f19 : arm vs PA Same |Newer |Older |Local | Remote | Missing | -- 11330 | 14 | 793 |1 | 332 | 1379 | http://142.204.133.82/jon/koji/kc.f19.diff.html ARM Build Status Wiki: https://fedoraproject.org/wiki/Architectures/ARM https://fedoraproject.org/wiki/Architectures/ARM/Fedora17_rawhide Thu Oct 11 09:29:12 EDT 2012 ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
[fedora-arm] HEADS UP: unplanned koji outage
Hi All, Just a heads up. While looking at the RAID rebuild which has been inhibiting us from doing composes as well as slowing builds down Dennis and I tested stopping various things to see the impact on the RAID rebuild. It was sitting at around 13,000 minutes when we started. Killing off the mash/composes running dropped it by half. Shutting off NFS altogether has cut the rebuild time to around 3 hours. So we've taken the executive decision to leave NFS disable to let the rebuild complete. The direct result of this is that while koji is still available for queries etc new builds won't be able to be submitted for the duration. We figured the 3 hours would easily be caught up with the performance being quickly back to normal. We'll follow up once it's back in approximately 3 hours. Peter ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
Gordan Bobic writes: >>> So what you're saying is that Fedora *still* supports an x32 CPU that >>> was released well over a decade ago... >> >> No. The XO-1 was released in 2007. That's half a decade ago. Given the >> project came out of MIT and you have a @mit.edu address I would hope >> you would be able to count, are you in politics by chance? > > I'm pretty sure that sort of attitude isn't doing much to attract > users to the community. DreamPlug1 is Kirkwood based and it only hit > the shelves less than 18 months ago. Moreover I'm talking about the i686 CPU, not the XO-1 product. The i686 was released a very VERY long time ago, definitely well before 2007. It was released while I was actually *at* MIT back in the early 1990s. > Gordan -derek -- Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory Member, MIT Student Information Processing Board (SIPB) URL: http://web.mit.edu/warlord/PP-ASEL-IA N1NWH warl...@mit.eduPGP key available ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] HEADS UP: unplanned koji outage
And we're back The rebuild is complete and things are in the process of returning to normal. Dennis is going to priortise the F-18 compose / mash / sync so we can get all the latest F-18 loveliness on it's way to the mirrors and after that we'll resume complete service as normal. Peter On Thu, Oct 11, 2012 at 3:34 PM, Peter Robinson wrote: > Hi All, > > Just a heads up. While looking at the RAID rebuild which has been > inhibiting us from doing composes as well as slowing builds down > Dennis and I tested stopping various things to see the impact on the > RAID rebuild. It was sitting at around 13,000 minutes when we started. > > Killing off the mash/composes running dropped it by half. Shutting off > NFS altogether has cut the rebuild time to around 3 hours. > > So we've taken the executive decision to leave NFS disable to let the > rebuild complete. The direct result of this is that while koji is > still available for queries etc new builds won't be able to be > submitted for the duration. We figured the 3 hours would easily be > caught up with the performance being quickly back to normal. > > We'll follow up once it's back in approximately 3 hours. > > Peter ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
[fedora-arm] Rasp Pi F17v5 Status Update
Hey everyone, Reading over the latest meeting logs, I see that there is frustration with what we are doing (or not doing) at Seneca. No excuses or anything but our team has been fairly busy since the summer: new co-worker, returning students, fsoss, presentations to higher-ups, weird working hours, build farm maintenance, machines failing and so on. Chris updated and built a new rasp-pi-vc package and we haven't tested it yet to see if other things break or not. I need to re-build my rasp pi splash screen and scratch seems to build, however, it has a bug where you can't open or save files. The weird part is that rasbians build of scratch works fine where ours doesn't and this has blocked us a bit because we would like it to work before release since the pi was primarily supposed to be a learning tool. I was going to run a compose possibly today and we might have to release with some potentially broken packages but this may cause more confusion and just cause more people asking in channel. I'm sorry if it seems as if we're not following the release soon/often rule, maybe I'm just not used to doing things like that yet. I think our team has too much on our plates right now in terms of work and I think that's why we've been delayed by this. I'll try to get in the habit of posting more updates about this, Thanks, Jon ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Re: [fedora-arm] Who's using Kirkwood?
On 10/11/2012 03:10 AM, Gordan Bobic wrote: Just out of interest, which packages are you referring to? I am assuming it is LibreOffice + a small subset of whatever is in Fedora that isn't in EL; mainly because I had no RAM/swap/CPU issues building any the 2000 or so packages that overlap. Takes about 3-4 weeks on a _single_ SheevaPlug. You're building 2000 packages, we're building 12000. Libreoffice is definitely one one of the problem packages where an armv7hl builder is called for. The koji server has a special 'heavybuilder' group which handles such packages. Are you using USB storage on your sheevaplug? It surprises me that you can get through even 2000 in 3 weeks unless half of them are noarch ;-) 3. Certain features such as atomic operations aren't available on armv5, reducing the number of packages that can be built for ARM in total: If it fails on armv5 but works on armv7, we still don't get it for armv7. In _most_ packages that require this, there are patches that address it. According to http://fedoraproject.org/wiki/Architectures/ARM/Fedora17_rawhide openmpi, pixie, mongodb are all currently broken due to atomics. This blocks condor, iwhd, perl-MongoDB, netcdf*, espresso, gdl, gdal, gromacs, ScientificPython, towhee, pypar, orsa, R-RScaLAPACK, nco, which in turn blocks even more packages. This is not an exhaustive list. This also doesn't consider that some package builds are transiently successful and transiently fail due to thread-safe issues which aren't coded for in armv5tel. With 12000 packages you never known when an armv5tel build is going to hit an SMP builder and expose such a bug. It happens all the time, but koji-shadow just reissues these builds so they work on a subsequent build... sometimes. Or they block hundreds of packages because of a transient failure. 5. On the whole, it's not a popular Fedora ARM target. Raspberry pi, OMAP, highbank, this is where most (not all) of our known users have hardware and interest. There are some Kirkwood users, clearly, but there are a lot more users of everything else. We should get some updated download stats on this to demonstrate, but last I saw kirkwood was maybe 3% of usage. Perhaps a poll might be a good way to ascertain this, rather than a discussion? Feel free to organize one, but what will you do with the resulting data? Regardless of the result, the bottom line is that the people who volunteer to do the work get to set the direction. The plug devices are perfectly useful, still in production, but there just isn't enough manpower interested and capable of supporting them over the long term. The way to keep kirkwood alive isn't to justify its existence, it's to do the work to keep it running. -- Brendan Conoboy / Red Hat, Inc. / b...@redhat.com ___ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm