[00:00] Okay, who's here? [00:01] * johnc4510-laptop +1 [00:01] hmmm [00:02] Well, we could start talking but maybe it's be best to wait a few minutes.. [00:02] it'd* [00:03] DPic: yes [00:03] and tyche is here for meeting too [00:04] cool [00:17] * johnc4510-laptop anyone else here for a marketing meeting? [00:17] i heard there was one from DPic [00:17] ah :) [00:17] supposedly starting... 15 minutes ago [00:17] yep [00:18] not many have shown up [00:18] yeah, disappointing :/ but we can still discuss plans [00:19] ok, we better start then, i have to leave in 40 mins. now [00:19] okay [00:20] so [00:20] so, for the hardy release, we want to have a big release campaign [00:21] so far we know we want to press release the local teams can adopt and release themselves [00:21] any other ideas? [00:21] we need to get fliers in places where people will pick them up [00:21] and for that, we need fliers [00:21] indeed [00:22] are there premade fliers from earlier releases [00:22] or [00:22] so they probably need tweaking [00:22] yeah [00:22] have we checked the art team for new ones for hardy heron [00:22] we could always start fresh and just use the past fliers for ideas [00:23] the art team doesn't seem.... very "with" it, but i'm sure if we asked them they'd make one [00:23] we haven't, does the art team make fliers and those sorts of materials? [00:23] that would be good. i'll get in contact with them [00:23] not sure, but they could point us to the newest artwork for hardy [00:24] i think it's supposed to be fairly similar to gutsy [00:24] true [00:24] the big redesign got postponed [00:24] yeah, that's probably for the best [00:24] :-\ [00:25] i would have liked the refresh, considering it's now competing against glossy UI's like OSX's and Vista's [00:25] i think they need more time to work on it, but just a simple refresh would have been nice [00:25] so, for the press release-- how can we make that effective? [00:25] we _can_ use ideas from past fliers for the content [00:26] ^ yeah, already said that :) [00:26] kk [00:26] we really want to start being a communication and resources hub for local marketing [00:27] we need to be able to contact all the LoCo's for whatever out current campaign is [00:27] which we can do through the loco contacts list [00:27] canonical puts out a press release themselves so we need something we can use to blog and digg etc and to let the loco's use [00:27] yeah, out own press release wouldn't be necessary [00:28] but we should focus on attracting more people who wouldn't normally hear about ubuntu [00:29] i created this page which we should work on a lot https://wiki.ubuntu.com/LoCoActivism [00:29] looking [00:29] as well as https://wiki.ubuntu.com/LoCoMediaCoverage [00:31] ok, as a start, *I* would like to see us getting the loco's involved in distributing fliers about the upcoming release [00:31] me too [00:31] and CD's too, of course [00:32] you are going to have a hard time with cd's i think [00:32] teams can't solely operate online, they have to get out and hit the streets to erach out to more people [00:32] Maybe create a template letter to send to local media agencies? [00:32] somerville32: canonical issues their own press release that can be used [00:32] most of the loco's however would be receptive to distributing fliers [00:33] DPic: Okay. So how about encouraging loco teams to forward that press release to local media agencies? [00:33] i think somerville32 is meaning something the loco's could send to their local media [00:34] somerville32: :) [00:34] yeah, definitely [00:34] I'm sure that some local agencies would print it if not for lack of something else :P [00:34] the loco media coverage page is a good guide for that [00:34] locos probably want to do some customization to it [00:34] why will CD's be difficult? [00:34] they might not be immediately available [00:34] Maybe we could have an easy step program? "Want to help promote Ubuntu 8.04? Step 1. Step 3. Step 4." [00:35] but when they are, they should be distributed as widely as possible [00:35] specifically about the next ubuntu release, maybe even covering a release party. [00:35] in reality, LoCos need to make their own CDs [00:35] really? [00:35] approved LoCos get a couple hundred [00:35] DPic: we could have the loco's insert times and dates of loco installfest on the fliers to give access to cd's [00:35] they probably wouldn't come out as nice as official ones tho [00:36] if you aren't approved, you might get 20 or 30 max [00:36] asking all the individual loco members to request CDs might help, but don't depend on shipit [00:36] no [00:36] well keep in mind, individuals on launchpad who are active can get ten-- and if each unapproved team has several active people on launchpad, they can get a hundred or so as well [00:36] boredandblogging: hey [00:37] fliering and postering is something we should add to the loco activism page [00:37] and we could link to our resources page when we have uploaded thos kinds of materials [00:38] as i said earlier, let the loco's use the release flier to promote installfest or release parties in their area [00:38] have you guys looked at the old spreadbuntu stuff? [00:38] i have, yeah [00:38] boredandblogging: i have not yet [00:38] :( [00:38] but there should be good ideas there [00:38] i put the useful stuff here https://wiki.ubuntu.com/MarketingTeam/Materials [00:39] as far as materials go [00:39] http://doc.ubuntu.com/~marketing/spreadubuntu/ [00:39] boredandblogging: you mean spreadUbuntu? [00:39] desertc: hah, yeah [00:39] it can be used for a starting point [00:39] or just for ideas [00:39] agreed [00:39] spreadbuntu is a windowing theme [00:40] DPic: is that a flier or a cd cover? [00:40] which? [00:40] last link, under flier [00:40] boredandblogging: yeah, spent a lot of time looking at DIY marketing, actually [00:40] need to encourage locos to think big, like get a huge banner and hang it where a lot of people will see it [00:41] DPic: https://wiki.ubuntu.com/MarketingTeam/Materials [00:41] boredandblogging: How many of these assets do you consider ready-to-use ?? [00:41] johnc4510-laptop: it's a flier [00:41] kk [00:41] and post the fliers anyplace that will let them [00:42] and they should be asking to post it anyplace the visit during the day [00:42] fliers are cheap and something anyone in a loco can do [00:43] you need a double sided print for those fliers, right? [00:43] flyers (silly spell check) ;) [00:43] desertc: yeah [00:44] is there anything else that people can do for the hardy release? [00:44] that increases the cost of production by quite a bit, just fyi [00:45] well, we'll have more soon [00:45] that one is just from the spreadUbuntu thing [00:45] contacting computer retailers might be good to request ubuntu and shot it off. but that's something that doesn't have to just be for the hardy release [00:45] show it off* [00:45] need to get the locos excited about hardy [00:45] computer retailers generally depend on windows for business [00:46] we should keep the flyers to _one_ sided productions i think [00:46] there isn't an incentive to push an alternative [00:46] johnc4510-laptop: that wouldn't be hard to do. i don't think having double-sided options would be a bad thing though [00:46] boredandblogging: could canonical get the locos an advanced press release on the features of hardy? else, is there a definitive list of new features that we can start passing around to locos? [00:47] it's at feature freeze [00:47] Maybe what we should do, is write up press releases not for the media, but for loco teams to read what we have going and what we want them to do [00:47] so yeah there should be a list up [00:47] i can check [00:47] yeah, features are set [00:47] not a bad thing, but it's not needed i don't think we want to make this as easy on the loco's as possible, and the most cost efficient [00:47] as for the canonical press release, I wouldn't really depend on them [00:48] boredandblogging: right - I understand what you're saying [00:48] think if we can come up with something general, and let the locos customize it, it would be good enough [00:49] so it sounds like we need to put together a short press release based on the hardy features - something in layman's term. one page should be ample. [00:49] boredandblogging: nice [00:49] we don't want to just have one thing. we'll craete a bunch of things that loco's can either use directly or customize for themselves [00:49] um [00:49] one thing [00:49] well - one thing would be the start of many things ;) [00:49] we should list ALL the features of ubuntu [00:49] GENERAL things [00:49] because these fliers aren't for people UPGRADING ubuntu [00:50] that's a good point [00:50] they're for people who *don't use it yet* [00:50] should point out the advantages of ubuntu and open source [00:50] or - at least include the new features AND the existing features [00:50] however, the newsworthy part is the new release, of course [00:50] i remember seeing an article on digg about what was wrong with the ubuntu press release for feisty and how it should have been different [00:50] let me try and find it [00:51] yeah me too [00:51] no version numbers [00:51] should have links to ubuntu, forums and how the community supports with help [00:51] like gnome 2.x.y [00:51] agreed [00:52] [00:52] [00:52] lol [00:52] http://blog.gobanquet.com/index.php/why-ubuntu-804-needs-better-marketing/ [00:52] no clue what happened there [00:52] hee hee [00:53] one challenge will be getting someone in the locos talking to the media. most likely they have no contacts. there might need to be some coaching for doing it [00:53] thanks for finding that, i couldn't find it for some reason [00:53] me neither [00:53] Yes, please. Coaching is definitely needed. [00:53] lol [00:53] desertc: yeah, and any people who are experienced with the media should contribute to the loco media coverage page [00:53] encourage locos to create some kind of kit, with documentation and CD and send it to some local media [00:54] we shouldn't assume that if we make a press release on a wiki that the locos will make the effort to get it in papers. [00:54] desertc: we need to get the team leads to handle that i think [00:54] in fact as we get further along we should start cultivating the team leads to get them on board with this [00:55] there is the loco contacts list [00:55] my expectation is that even getting a few locos to get this in action will be challenging, so we should start with just a few - maybe provide coaching for the approved locos and see what happens. It's going to be a learning experience for everyone [00:55] DPic: yes [00:56] it can be reached from the usteams page i think [00:56] * johnc4510-laptop sorry i've got to run :( [00:56] i'll catch up from my logs [00:56] alright thanks for showing up [00:57] well we definitely have a good amount to work with now [00:57] should we move onto talking about a meeting schedule? [00:58] :) [00:58] sure [00:59] okay, so how often should we meet, and what should we do about timing (since we're a global team)? [01:00] most of us are from the us time zones [01:00] okay so then around this time shouldn't be a problem [01:00] I know I was thinking about the us-locos in this discussion [01:00] should we meet once a month, or twice a month? [01:01] why not do another one in 2 weeks [01:01] for me, feel free to ping me whenever - if I'm online, then I'm available to help [01:01] if no one shows up [01:02] boredandblogging: in 2 weeks if no one shows up? [01:02] I will work on writing up some notes for a press release - need to work on writing a linux speech anyway this week ;) [01:02] If someone would fix the Fridge's calendar... I wouldn't miss any of the meetings because google would page my blackberry [01:02] what's wrong with the fridge's calendar? [01:03] Won't work with google calendar [01:03] ah [01:03] Some drupal module bug or something that is apparently fixed upstream [01:03] i see [01:04] so we all agree, every 2 weeks, at 00:00 UTC? [01:04] +1 [01:04] +1 [01:04] might also want stop by the -locoteams channel [01:04] and see if anyone there has any ideas [01:04] and always on wednesday? [01:04] or what they are planning to do [01:04] boredandblogging: good idea [01:05] does this day of the week always work for everyone? [01:05] sure [01:05] sure [01:06] great [01:06] fine [01:06] anything else to talk about? [01:07] i'll get working on the notes for this meeting, and drawing up a plan for us [01:09] meeting adjourned? [01:09] sure [01:10] okay thanks everyone [01:13] !seen [01:13] The seen function has not been operational for a long time. Use /msg seenserv seen nickname instead. === ember_ is now known as ember [04:07] did i miss the marketing mtg? [04:14] no, starts in 2 hours [04:14] oh wait marketing,yes [04:15] d'oh. darn UTC time!!! [04:15] thanks === _czessi is now known as Czessi === ubotu changed the topic of #ubuntu-meeting to: Current meeting: Platform Team Calendar: http://fridge.ubuntu.com/event | Logs: https://wiki.ubuntu.com/MeetingLogs/ | 27 Feb 12:00 UTC: Education Team | 27 Feb 21:00 UTC: Server Team | 28 Feb 14:00 UTC: Desktop Team | 29 Feb 12:00 UTC: MOTU | 05 Mar 07:00 UTC: Platform Team [06:51] good evening [06:52] good early morning [06:52] heya [06:52] * ogra tries to get his eyes open ... [06:52] heh [06:54] * Hobbsee throws some sand at ogra [06:54] Hobbsee, that only helps if to match a moment where i actuallly have them open :P [06:54] (and am not yawning) [06:54] * Mithrandir tickles Hobbsee [06:55] * Hobbsee tickles Mithrandir back, and stomps on his feet, before he levitates [06:56] hi [06:57] * cjwatson blinks a few times [06:57] good morning [06:58] * ArneGoetje waves [06:58] Hey folks. [06:58] asac,slangasek: ping [06:59] * TheMuso looks out the window and hopes the storm doesn't come this way... [06:59] cjwatson: hi [07:00] evening, cjwatson. here's your coffee [07:00] heh [07:01] just don't fall asleep in it :P [07:01] if its hot enough that will wake him up :) [07:01] aww! And I thought everybody on the team was a morning person. :) [07:01] ok, just missing asac - let's start anyway and hope he shows up [07:01] * ogra slowly feels the healing power of caffeine in the morning [07:02] first off, activity reports: I have four out of nine this week [07:02] mine just hit the list ... [07:02] please send these *before* the meeting [07:02] (sorry) [07:03] next the real agenda [07:03] would you prefer if I started sending mine end of day Monday? [07:03] * Actions from last week [07:03] * Result of multi-monitor discussion with desktop team? (Bryce) [07:03] bryce: how did this go? [07:03] * asac hides [07:03] slangasek: no, yours was fine, I get up a little early before the meeting so I can catch up on them [07:03] ok [07:04] cjwatson, it's come along fairly well [07:04] cjwatson: I have debs and screenshots up at: http://people.ubuntu.com/~bryce/Testing/XrandrGui/ [07:04] right now I'm mostly waiting on seb128 to review/upload things [07:05] I was hoping to get some feedback from him on the integration, but I gather he's been busy with the new gnome stuff [07:06] the gui installs and runs, and it seems to work (but I found a bug when using it on a KVM) [07:06] most of the work has been packaging - a lot of autoconf/automake hackery [07:06] aside from the text about F9 :-), that looks pretty good; no facility for choosing a different driver? [07:07] right, this only uses xrandr interfaces [07:07] screenshot 1 and 2 are before / after [07:07] I just want to make sure that if we drop displayconfig-gtk we're not reliant on anything it does [07:07] screenshot 2 corresponds to what's in the debs [07:08] ah [07:08] well, displayconfig-gtk is essentially a fancy xorg.conf editor [07:09] this new tool does everything through run-time interfaces, so anything that requires xorg.conf modification is going to be out of scope for it [07:09] perhaps we should prepare some release notes text about how to do anything that's dropped [07:09] ok [07:10] ok, thanks [07:10] * 5-a-day [07:11] Jono has asked that I remind everyone of the existence of 5-a-day (https://wiki.ubuntu.com/5-A-Day) and encourage all of you to take part [07:11] err, where relevant anyway [07:12] even if you're already doing equivalent bug work (which most of you should be, at a minimum), please also help to spread the message [07:12] will do [07:13] the wiki page has pointers to some tools to help with this [07:14] ok [07:14] hmm, so I should just call 5-a-day before a package upload? [07:14] that's one reasonable approach, yes [07:14] if you're feeling particularly enthusiastic you could hook it into dupload ;-) [07:15] ok, will do. [07:16] but the important bit is publicising to the community that this is something that the Ubuntu community does as a matter of course [07:16] I've certainly been thinking that it *could* be a way to encourage people testing a11y to file bugs, as they just about always post to mailing lists... [07:17] But other than that, I'll likely jump in and participate as well. [07:17] hmm, I think filing bugs is just about the only bug activity that's not part of 5-a-day [07:17] the point is to get the bug count down :-) [07:18] right [07:19] file more bugs so you can turn around and close them? ;-) [07:19] I think that's enough on that topic :-) [07:19] * Final call on dmraid (Luke) [07:19] Right. [07:19] If people haven't been following my reports and other IRC activity, let me bring everybody up to speed. [07:20] For several years, many consumer, and some low-end server boards have been shipping with what is known as fake raid. [07:21] Basically, the BIOS of the disk controller writes special metadata to a pair or so of disks, to make them a fake software raid, be it raid 0, 1, or 5. [07:21] Windows drivers for these controllers know what to look for, and transparently read/write this metadata as necessary, with the user not even knowing most of the time that windows is sitting on top of a RAID setup. [07:21] Dmraid is such an interface to this metadata for Linux. [07:22] However, so far from my testing, it doesn't come close to either what the BIOSs can do, or what Windows drivers do in terms of keeping the array consistant., [07:23] So, with dmraid still not being able to reliably provide event monitoring for failed/degraded arrays, and not being able to rebuild arrays, as well as at least one BIOS on a controller I have always finding array inconsistancies when Linux has touched the array, we need to really consider whether dmraid is something we want for hardy. [07:23] From my point of view, it comes down to whether I keep actively working on it, dispite these issues, or whether I move onto other things, and keep an eye on it for hardy+1. [07:24] perhaps whitelist it on a known working chip/bios-rev basis? [07:24] I have also discovered that kernel patching is required to allow for the use of RAID 5 via dmraid. [07:24] calc: Since it uses metadata on the hard disk, this could be possible, but the fact is, that at least one BIOS I have complains of array inconsistancies after I even do an install onto a dmraid array. [07:24] what's the alternative to having dmraid enabled - having no support for the controller, or having support for it as a plain IDE/SCSI controller? [07:25] Another controller I have did not appear to be as strict, but only complained when I degraded the array. [07:25] slangasek: Only supporting it as a standard IDE/SCSI controller. [07:25] /SATA even [07:25] * slangasek nods [07:25] slangasek: Which means that users who have windows on such an array could very easily destroy it. [07:26] oh... :) [07:26] but, you're saying that the current code isn't guaranteed to do much better for them? [07:26] slangasek: No it is not. [07:26] TheMuso: can we detect this situation and display a warning to the user? [07:26] Especially if the BIOS is as strict as my hpt37x controller supporting fakeraid. [07:26] is the code not working for any of the many different brands, or just not good on certain ones? [07:27] it looks to me that the dmraid program isn't really hooked into device-mapper change events, and that it only erases the metadata if you explicitly tell it to [07:27] calc: I don't know, as I only have 2. The code works, but its a matter of makings ure the array is consistant once Linux shuts down. [07:27] grr hpt is crap, i had trouble with one of those writing into cylinder 0 even when raid wasn't enabled :-\ [07:27] overwriting grub bootloader in the process [07:28] TheMuso: ok [07:28] hmm, hpt is the one that's now courting Linux [07:28] cjwatson: As I said in my report, there is work going on to get some stuff into the kernel to support dmraid better. There is also more userspace work going on for event monitoring. [07:28] does intel's matrix raid stuff use dmraid? [07:28] TheMuso: was your testing just with the installer, or did you also try editing partitions in a regular system? [07:28] getting intel's fake raid to work (if possible) would probably be a good thing [07:28] TheMuso: the reason I ask is that I notice that the installer integration does not do the same stuff that /etc/init.d/dmraid does on shutdown [07:29] cjwatson: I tried the following: 1) install windows, then Linux. 2) Install Windows, edit partitions, reboot. 3) install Linux only. [07:29] All were caught by the BIOS as possible inconsistancies on the array. [07:29] cjwatson: Well, I even tried hand-deactivating the arrays before rebooting. [07:29] though, I have to say I'm slightly surprised that the partition table is ATARAID metadata as such [07:30] I would have thought that dmraid should present a device to Linux on which parted blindly edits the sectors corresponding to the partition table, which then translate into writes to both disks [07:30] cjwatson: The partition table is totally separate to the metadata. [07:30] The metadata is at the end of the disks. The partition table is where you would normally find it. [07:31] right, so what's the metadata that needs to be updated? would that not normally only be if you added or removed disks, or if disks failed? [07:31] I'm trying to figure out why partitioning changes it [07:31] cjwatson: I don't know without digging into code. My guess something to do with state is stored in metadata, but I think this varies from controller to controller. [07:32] Well as referred to by calc, it could also be a flaky BIOS. There is no newer revisions for the board however. [07:32] And, running WIndows/DOS and doing all manner of parttion edits etc does not seem to bother the BIOS. [07:33] dmraid seems to have explicit BIOS-specific metadata handlers [07:33] Yes it does. [07:34] As all metadata is different from controller to controller. [07:34] bug 8978 refers to the issue i had with hpt controllers doing stupid things [07:34] Launchpad bug 8978 in grub "Grub - Error 21 returned" [Medium,Confirmed] https://launchpad.net/bugs/8978 [07:35] Now. I will be getting a new system in the coming weeks, that will have 2 different dmraid compatible controllers on it. I'll be interested to see how well they are handled. One will be intel, and one will be Gigabyte. [07:35] i have some edubuntu users using hpt [07:35] hmm, are there actually any bugs filed about these inconsistencies after shutdown? [07:36] * asac looking at dmraid lp bugs page [07:36] However my silicon image controller was not even bothered when Linux was installed. Never tried Windows+linux, as I wanted to move the dmraid testing to another box. [07:36] asac: No, as it is only with cone controller that I get the behavior. [07:36] one [07:36] TheMuso: I have two dmraid systems that were sent to me and can help, as I've also been tasked with similar goals. [07:37] ... so at least one controller *does* work? [07:37] evand: Right, if I were to give you some quick isntructions as to build a modified alternate disk, wuould you mind having a look? [07:37] cjwatson: Yes, for a linux only install. [07:37] I can certainly try it again, if required. Need to card swap, but thats no big deal. [07:37] TheMuso: not at all, I'll make it a top priority. [07:37] that's a little more promising [07:38] cjwatson: However, there is still the issue of not being able to rebuild the array from Linux. [07:38] basically, I'm just aware that this is an increasing target and if we don't get something in I suspect it's going to come up in point releases as a hardware support target anyway [07:38] So at this point, we could certainly blacklist the hpt37xx controller I have. [07:38] cjwatson: Right. [07:38] if it were totally non-functional, I'd agree it doesn't make sense [07:39] but it sounds to me that we have some specific bugs on which we need help from upstream [07:39] cjwatson: I do need to do a dual boot windows test with that silicon controller however to be sure. [07:40] The upstream mailing list is rather quiet. The person at Red Hat who maintains this certainly is working on it, but I think he is also involved with other things. [07:40] I'm downloading fedora 8 as we speak to see if that yields any different results. [07:40] As they've had dmraid in fedora for a few releases now. [07:40] TheMuso: blacklisting probably should be based per bios revision, but that might get complicated [07:41] I think a fairly sane approach would be to move it to main for wider testing for now (since we need it in main in order to have it in the installer), but put a prominent note in the release notes for alpha 6 and beta that we will rip it back out again if test results are unfavourable [07:41] calc: Yeah I'd say it would, but as I said, there is no newer revision for this board, and I can't download an older one to try. [07:41] maybe best to upgrade chipset to most recent bios (where possible) if it doesn't work blacklist the whole chipset for the time being [07:41] how does that sound? [07:41] sounds agreeable to me [07:41] pretty dangerous [07:41] heh [07:41] Note if any testing is to be done, it should be on a clean system so that data loss is no issue. [07:41] ogra: elaborate? [07:42] since its probably the base of your install people really need to be aware they need to reinstall if it fails [07:42] i mean [07:42] its your filesystem [07:42] TheMuso: I look forward to your recommended release notes entry ;-) [07:42] we could include a warning in partman-dmraid that this is explicitly experimental and you MUST have backups [07:42] that needs a very wordy warning [07:42] cjwatson: partman-dmraid already has a warning. [07:42] "NO, REALLY" [07:43] "Type 'Yes, do as I say' to continue" :) [07:43] i mean, i'm getting mails atm if ltsp is already final for hardy ... from people wanting to use it in production next week (if i dont shout) [07:43] The support for SATA RAID disks (using dmraid) in the installer is [07:43] experimental. You should make sure that you have a backup of any [07:43] data on your system that you do not want to lose! [07:43] [and more] [07:43] Well there is a question that gets asked by partman-dmraid. [07:43] since "its only about a month") [07:44] cjwatson: As for raid 5 support, I could easily see how well that patch could be adapted for lum perhaps? [07:44] ogra: this is balanced against the fact that you can try to install Ubuntu right now on these systems and it's *guaranteed* to do bad things to your array [07:44] right [07:44] TheMuso: it came up a little while ago, and I didn't think it was all that important? [07:45] TheMuso: do people need RAID 5 support? [07:45] cjwatson: I don't think that was dmraid related... [07:45] it was some module that had 'dmraid' in its name [07:45] cjwatson: I certainly know intel controllers support it, and some others do. [07:45] dmraid56 IIRC? [07:45] cjwatson: I didn't dig, but I'll ahve a look. It could be the one. [07:46] if it's needed, then I apologise for saying it wasn't - #ubuntu-kernel should be able to help [07:46] Yes there was a post to the kernel list about it, but I don't think its been acted upon. === arualavi_ is now known as arualavi [07:47]