[00:19] join #ubuntu-testing === txwikinger is now known as Guest31757 === txwikinger_ is now known as txwikinger === MTecknology is now known as mteck-mobile === MTeck is now known as MTecknology === _neversfelde is now known as neversfelde [15:34] !schedule [15:34] Ubuntu releases a new version every 6 months. Each version is supported for 18 months to 5 years. More info at http://www.ubuntu.com/ubuntu/releases & http://wiki.ubuntu.com/TimeBasedReleases [15:59] o/ [15:59] o\ [15:59] yo [15:59] \o [16:00] * mathiaz waves [16:00] \o [16:01] o\ [16:01] let's get the server team meeting started [16:01] #startmeeting [16:01] Meeting started at 10:01. The chair is mathiaz. [16:01] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [16:01] o~ [16:01] ~o~ [16:01] !"#¤ [16:01] once everyone stopped dancing we can proceed to today's agenda: https://wiki.ubuntu.com/ServerTeam/Meeting [16:02] Last week minutes: https://wiki.ubuntu.com/MeetingLogs/Server/20090210 [16:02] [TOPIC] Update ServerGuide for Jaunty [16:03] New Topic: Update ServerGuide for Jaunty [16:03] sommer: how is doc.ubuntu.com going? [16:04] mathiaz: well, most of the new sections are done [16:04] mathiaz: and most of the bugs are fixed [16:04] some reviews would be cool :) [16:05] mathiaz: I did have a question about the OpenLDAP DIT? [16:05] sommer: https://wiki.ubuntu.com/JauntyServerGuide lists all the sections that needs review [16:05] yep [16:06] sommer: what's your question about the DIT? [16:06] mathiaz: is it going to be part of the slapd package? or I guess how is the base going to be setup? [16:06] sommer: for jaunty nothing changes [16:07] mathiaz: because I think it should simplify the ldap section [16:07] sommer: I'll try to get a new upstream version of openldap uploaded before FF [16:07] mathiaz: oh, I thought there was going to be something [16:07] sommer: well - nothing more than what was already there [16:07] sommer: ie a very basic tree structure. [16:08] mathiaz: I see, that works... answers my question, thanks [16:09] sommer: great. Anything else apart asking for reviews of the new sections? [16:09] mathiaz: nope, that should be it [16:10] sommer: oh - is doc.ubuntu.com up-to-date? [16:10] mathiaz: ah, I don't think it is yet, but I did hear back from mdke and he's working on it [16:10] sommer: ok great. [16:10] let's move on. [16:11] [TOPIC] Postfix and Dovecot integration [16:11] New Topic: Postfix and Dovecot integration [16:11] right on time :) [16:11] ivoks: good work on that one. [16:11] \o/ [16:11] mathiaz: it's not done yet :) [16:11] ivoks: there are still some things to be fixed [16:11] mathiaz: we have one bug and couple of issues [16:11] A post on the ubuntuserver blog has been published: http://ubuntuserver.wordpress.com/2009/02/13/an-improved-mail-server-stack-in-jaunty-dovecot-and-postfix-integration/ [16:11] i've prepared a debdiff, but i would like a review on it, since this is my first binary-indep package [16:12] lamont: ScottK: do you have comments on the new dovecot-postfix package? [16:12] mathiaz: I haven't had a chance to really look at it. [16:12] ditto [16:12] don't, untill new version is uploaded :) [16:12] ivoks: attach it to a bug, and subscribe ubuntu-main-sponsor to ask for sponsorship [16:13] will do [16:13] ivoks: apart from testing is there anything else to be done on this topic for jaunty? [16:14] not really [16:14] there are plans for jaunty+1, but that's another story [16:14] since we are in FF [16:14] ivoks: sure - where are these plan kept? [16:14] nowhere atm [16:14] ivoks: is there a wiki page somehwere? [16:14] not yet [16:15] basicaly, integrating amavis with these two [16:15] and myabe having a 'supported' webmail infrastructure [16:15] ivoks: ok - it may be worth creating a wiki page under w.u.c/ServerTeam/MailServer [16:16] good idea [16:16] ivoks: as a braindump page to keep track of interesting ideas [16:16] ivoks: just bullet points - the interesting ideas can then be turned into specs for the next cycle [16:16] right [16:17] [ACTION] ivoks to create a wiki page for ideas about improving the mail server task [16:17] ACTION received: ivoks to create a wiki page for ideas about improving the mail server task [16:17] ivoks: ScottK: anything else to report in the mail server area for Jaunty? [16:17] it will rock. [16:18] :) [16:18] It might be worth a mention that dkim-milter now uses unbound for DNS. [16:19] Unbound is a relatively new DNS library that supports DNSSEC. [16:19] Also just as a warning, clamav 0.95 is due out soon. [16:19] It, of course, breaks all the rdepends. [16:20] ScottK: hare you blogged about it? IIRC I saw something mentioning about it [16:20] Yes [16:20] ScottK: it may have been in the changelog though. [16:20] They promised though that this is the one true interface redesign and they won't have to do it again. [16:20] They've never promised this before. [16:20] So expect some post-FF fun with this one. [16:20] ScottK: ok - we'll see how things will turn out. [16:21] anything else to add on this front? [16:21] In related news clamav 0.94.2 is now available in hardy-backports [16:21] We're about ready to push it to dapper-backports too. [16:21] That's it. [16:21] ScottK: great - thanks for the good work! [16:21] ivoks: ^^ ! [16:21] [TOPIC] Power management [16:21] New Topic: Power management [16:21] kirkland: anything new on this front? [16:24] ok - kirkland seems to have lost power [16:24] let's move on [16:24] [TOPIC] virtualization [16:24] New Topic: virtualization [16:24] \o/ [16:24] soren: what's new there? [16:25] Lots of stuff going on. We've got a very fresh kvm, libvirt, virt-manager (currently uninstallable due to old virtinst... working on it).. [16:25] yippee [16:25] ...I'm working on some updates for VMBuilder. Among the new things is support for VMWare ESXi. [16:25] ..I hope to manage to rework a few things that need some love before FF. One thing I'd like to do is to make it able to build for multiple hypervisors in the same run. [16:26] Also, we've been pushing a lot of cloud computing stuff. [16:26] A month ago (or thereabouts), I uploaded OpenNebula, and we're currently working hard to get Eucalyptus uploaded before FeatureFreeze. [16:27] soren: are there plans to blog about all this new things? [16:27] Eucalyptus has a lot of Java dependencies, so it has been a bit of a rough ride, but upstream is very happy to work with us and are making excellent progress. [16:27] mathiaz: Just after featurefreeze, yes. [16:27] soren: or give an overview about all these new things? [16:27] Yes. [16:28] I want people to test all of this much more, so I'll be posting about clever ways you can use OpenNebula for testing (I'd like to work with you on that). [16:28] soren: awesome - we'll have to wait for next week to get all details then:) [16:28] I've been a bit hesitant to blog about Eucalyptus, since it's still vapourware. [16:28] But eucalyptus is in the archive now.. [16:29] Most of it's build-dependencies are, too, so we're inches away from a usable package :) [16:29] Koon is doing fantastic work on the Java dependencies. [16:29] * soren hugs Koon [16:29] s/it's/its/ a few lines ago.. [16:30] Hmm... I think that's all for now. I'll be blogging about stuff as it becomes available. [16:30] * ivoks claps [16:31] soren: awesome. Thanks for giving this short status update. [16:31] Certainly [16:31] wow [16:31] soren: should we have info in the serverguide about the cloud stuff, or is it mainly ec2 at this point? [16:32] mathiaz: sorry, my network dropped; back now [16:33] sommer: We definitely should. === dholbach_ is now known as dholbach [16:33] we cant mention the ec2-ami-tools in the eucalpytus stuff though [16:33] sommer: Things should be stabilizing enough within the next week for it to actually make sense to start writing docs for it :) [16:34] soren: sounds good, I'll be patient [16:35] ok - let's move on then [16:35] zul: gotcha, we'll just make it generic [16:35] since we'll have to wait for FF to get a good overview of the new virtualization features in jaunty :) [16:35] [TOPIC] Power management [16:35] New Topic: Power management [16:35] kirkland: now that you're back - any news on this front? [16:36] mathiaz: i have pwrkap and powerman are two new packages, added to the archive today [16:36] mathiaz: powernowd has been added to the server seed, to enable cpu freq scaling by default on ubuntu servers [16:37] mathiaz: pm-utils and ethtool have been added to the server-ship seed, which are what's needed for suspend/hibernate and resume by WoL [16:37] mathiaz: as well as wakeonlan, promoted to main, and on the server-ship seed [16:37] mathiaz: i need to write up some instructions on how to test this [16:37] mathiaz: i'm interested in feedback on hardware where suspend/hibernate can be verified as working [16:38] mathiaz: hw which is *not* working, i'm interested in that too... but we probably won't focus on fixing those until next cycle [16:38] kirkland: great - so the current plan for jaunty is to have powernowd installed by default and all the other suspend/hibernate components available from the cd? [16:38] mathiaz: correct [16:38] mathiaz: suspend is really cool, btw. [16:39] mathiaz: for my home network, i created a ssh wrapper, which pings first to see if alive [16:39] mathiaz: if not, sends a wakeonlan, waits 3 seconds, and ssh's again [16:39] mathiaz: my machines wake from suspend when i need them :-) [16:39] sweet [16:39] mathiaz: and it only takes about 3 seconds for them to wake up which is lot faster than ~60seconds to boot [16:40] mathiaz: i put a power meter on them too [16:40] kirkland: awesome! [16:40] mathiaz: running, they're about 50W-80W (depending on cpu freq, load, hd, cd, etc) [16:40] mathiaz: suspended, they're at about 11w [16:41] kirkland: that seems like a good topic for a blog post [16:41] kirkland: a good use case for supsend/resume for servers [16:41] mathiaz: cool, will do. [16:42] [ACTION] kirkland to blog about using suspend/resume for servers [16:42] ACTION received: kirkland to blog about using suspend/resume for servers [16:42] ubuntu - saving energy and the climate! [16:42] penguins [16:42] lol [16:42] I don't suspend, but I do set hdparm to power down the harddrives... [16:43] my server doesn't do suspend :( [16:43] ivoks: did you report it to kirkland ? [16:43] kirkland: fix it [16:43] mathiaz: just did :) [16:44] ivoks: let me set up a wiki page to track this [16:44] kirkland: how do you keep track which hw is works/doesn't? [16:44] ivoks: i'll ping you to fill in your info [16:44] kirkland: sure [16:44] mathiaz: i'm going to do it by wiki, for starters [16:44] kirkland: ok - one wiki page seems a good place to start. [16:44] mathiaz: i'll link off of the laptop suspend page [16:45] kirkland: ah ok - there is a suspend page for laptop already. [16:45] another way of waking up server is IPMI [16:45] There's also multi-core power savings: 'echo 1 > /sys/devices/system/cpu/sched_mc_power_savings' [16:45] any chance to see ipmitool in main? [16:45] and laptop-mode, which I do use on my server: 'echo 5 > /proc/sys/vm/laptop_mode' [16:46] ivoks: i'm not too familiar with ipmi [16:46] dell servers support ipmi; i'm not sure about wake on lan [16:46] ivoks: ipmitools aren't part of openipmi? [16:46] mathiaz: no, ipmitool can use openipmi, but doesn't depend on it (it can use real hardware) [16:46] apparently no [16:47] openipmi is just support for ipmi on local machine [16:47] ivoks: does ipmitool provide cmd lines that openipmi can't? [16:47] ipmitool can be used to connect to other machines [16:47] mathiaz: yes; serial over lan, for example [16:48] nijaba: would ipmitool be useful in main? [16:48] with openipmi you can't do anything outside local machine [16:48] they don't exclude each other [16:48] mathiaz: i would definitely think so [16:49] ok - we can always move ipmitool to main after FF [16:49] to get started, a MainInclusionReport has to be written [16:49] right, but it's not being developed for 2 years :/ [16:50] ivoks: hm - right. [16:50] ivoks: if upstream is not very active it may not be worth [16:50] ivoks: and we can't really do anything about it :/ [16:50] true [16:50] sorry, my bad [16:50] there are new releases [16:50] http://sourceforge.net/projects/ipmitool/ [16:50] LINK received: http://sourceforge.net/projects/ipmitool/ [16:50] ivoks: anyway - we can always think about it after FF [16:50] ok [16:51] [TOPIC] Open Discussion [16:51] New Topic: Open Discussion [16:51] anything else to add? [16:52] mathiaz: there was that reply about an Ubuntu Directory Service [16:53] sommer: right - I was going to reply to point out to the ubuntu-directory@ ml list and LP team [16:53] ah, that works :) [16:53] sommer: and also mention that we're about to enter FF so now is not the most appropriate time to start a discussion for jaunty [16:55] true... I didn't have anything else [16:56] ok - let's move on to the last item on the agenda: [16:57] [TOPIC] Agree on next meeting date and time [16:57] New Topic: Agree on next meeting date and time [16:57] next week, same time, same place? [16:57] + [16:57] sure [16:59] ok - then. [16:59] see you all next week, same time, same place [16:59] get ready for FF and make jaunty rock! [16:59] #endmeeting [16:59] Meeting finished at 10:59. [16:59] thanks for hosting mathiaz [17:00] later all [17:01] * smb_tp settles down [17:01] Hello All! Time for the kernel team meeting [17:01] #startmeeting [17:01] Meeting started at 11:01. The chair is pgraner. [17:01] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [17:01] * rtg_ waves [17:01] * apw phases in ... [17:02] * lieb here [17:02] * amitk arriveth [17:02] [TOPIC] Security & bugfix kernels [17:02] New Topic: Security & bugfix kernels [17:02] Greg keeps me busy. [17:02] Intrepid .... [17:02] smb_tp: how are we looking? [17:03] Anyhow, for Intrepid I am done with .17 and try to get this uploaded next day before Greg comes with .18 [17:03] smb_tp: when do we hit the 4 month cutoff for Intrepid? [17:04] End of this month [17:04] I'd say. [17:04] smb_tp: ok, so .18 might not make it then...? [17:04] * rtg_ agrees [17:04] i think we mooted by beta freeze [17:04] Not completely no. Depends on the time scale [17:04] smb_tp: how does Hardy look? [17:05] Also near upload. I added a few things including some xen patches and the vmware tsc updates [17:05] So this time both hardy and intrepid will bump abi [17:05] sorry, running late [17:05] Intrepid again [17:06] Unfortunately I am still fighting to catch the Aspire bug on Hardy, but this will have to get in after that. Actually by the nature of it it might just resolve itself [17:06] Anythings else for this topic? [17:07] [TOPIC] Jaunty [17:07] New Topic: Jaunty [17:07] Reminder Feature Freeze is the 19th (this week) [17:08] I'd like to do a quick rundown of the big features to see where we are: [17:08] lieb: kerneloops & crashdump? [17:08] have to test running non-root. no one else does that config. [17:09] crashdump, its a bigger problem req updates to 3 packages [17:09] turns out the kexec kernel also oom panics [17:09] ;( [17:10] lieb: status of MIR? [17:10] how big is the 'crash dump window' by default? [17:10] pgraner, awaiting my test results +debdiff [17:11] apw, I'd have to check. [17:11] lieb: we are coming down to the wire, we need to get this locked up ASAP. This should be your #1 priority [17:11] lieb: do you need help? [17:11] it is. the kexec stuff will probably not make it due to too many packages [17:12] the kerneloops stuff should be settled today once I do web sumit testting to confirm [17:12] lieb: what packages specifically needs to be updated? [17:13] no more code, just re-working the debian package to ubuntu (init scripts) cuz they don't do it either [17:13] lieb: is that something we can get the foundations team to help with? [17:13] wrt kerneloops, just the one. wrt kexec, the kernel (or whatever configs the kdump image), makedump [17:13] crash, and (possibly) kexec-tools [17:14] pgraner, maybe but what is really needed is per the uds topic, these tools/daemons need closer proj mgmt coupling w/ the kernel [17:15] I mentioned that at the sprint but I haven't had time to compose the web page yet [17:15] we arn't going to change the coupling at this late stage in the jaunty cycle i don't recon [17:15] lieb: I'm having a hard time understanding why this is so difficult, we can take this offline\ [17:15] apw, that is true. [17:15] pgraner, ok [17:16] apw: how are we on suspend/resume? [17:16] we are starting to get some real user reports now === fader_ is now known as fader|lunch [17:16] i am monitoring those separatly from the normal bugs [17:17] and trying to get some feedback to every one [17:17] there is a bunch of people who are reporting things because apport [17:17] woke up, but who don't really know what it means. not sure [17:17] how we will handle the ones where they arn't capable of helping us yet [17:17] did you notice an uptick once it hit planet.ubuntu? [17:17] was that what it wa [17:18] was, i am also seeing people sliding over to jaunty [17:18] in general and triggering the reports [17:18] so far its about 5 a day [17:18] good deal [17:18] i'll have a better feel by the end of the week [17:18] we have done some more work on the apport support [17:18] to get more information from the users system without needing [17:19] to talk to them, and also helping make the reports easier to sift [17:19] apw: great for next week can you give us a rundown of what the prob look like and how many etc...? [17:19] ogasawara, i should note that the bugs now have apport-kerneloops, resume and _one_of_ suspend or hibernate now [17:19] pgraner, sure thing, most right now are getting a cut-n-paste response [17:19] apw: good, that'll make searching easier [17:19] as they have 0 information on what they were doing. whenwe have interacted with a few [17:20] more we'll know better [17:20] apw: ack, good deal [17:20] apw, you capturing hibernate issues - i thought it was just suspend/resume? [17:20] the testing stuff is just suspend, but the apport stuff detects both [17:20] and differentiates them in the bugs as reported (now) [17:20] ahah [17:20] apw: one more... vanilla kernel builds... what the status, and how are we on ppa hosting? [17:20] i am seeing an uptick in those reports [17:21] vanilla builds are now up to date, latest of each tree is built [17:21] we never planned to PPA host them. [17:21] in the ~apw version on kernel. the old one on people now points there [17:21] rtg_: sabdfl wants them in a ppa [17:21] rtg is now testing the more official kernel-ppa user for them [17:21] rtg_: for the ease of end user testing [17:21] i haven't had time to test loading them into a PPA yet [17:22] hmm [17:22] but the thought there was to see if we could keep just the latest of each in a PPA [17:22] not try anything clever to get them all in [17:22] apw: in order to do so, we have to upload a source package [17:22] but that is not started yet [17:22] rtg_ yeah, but with the lpia experience under my belt, i think i know waht needs twiddling [17:23] to get at least one of each into a ppa. anyhow, we committed to trying it and i have that on my todo [17:23] hope to be able to report on that next week [17:23] its the package namig thats gonna be PITA [17:23] yes. i think as long as we only keep the latest per release [17:23] ie a .24 in hardy, a .27 in intrepid and a .28 in jaunty [17:23] then we may be able to shove them all into a ppa [17:24] beyond that, you know what you are doing and can install a .deb [17:24] from the kernel-ppa web site [17:24] k [17:24] as i say its got to be investigated, and it may be [17:24] 'to do that we need these 3 things from a PPA please' [17:25] so it may be a KK think [17:25] thing [17:26] apw: we're also gonna have to get some storage exceptions. these kernels start to eat a lot of space [17:26] [TOPIC] Jaunty Bugs [17:26] New Topic: Jaunty Bugs [17:26] ogasawara: what are the top ones this week? [17:26] rtg_, yeah speacially as we have had to keep the source [17:26] so the majority are from previous weeks . . . [17:27] from last week bug 326891, bug 255886, and bug 323256 - all got feedback [17:27] we didn't make much dent in the previous 2 weeks worth which dropped week-1 [17:27] Launchpad bug 326891 in linux "2.6.27.11 kernel breaks r8169 support for rtl8102e" [High,Fix committed] https://launchpad.net/bugs/326891 [17:27] Bug 255886 on http://launchpad.net/bugs/255886 is private [17:27] Launchpad bug 323256 in linux "2.6.27-11 Intel Ethernet e100e Remains active after shutdown" [High,In progress] https://launchpad.net/bugs/323256 [17:28] The e1000e is looked after by inter and can be worked around [17:28] intel, even [17:29] there are some older ones marked as High: [17:29] bug 323134 [17:29] Launchpad bug 323134 in linux "oopses with sr0 in armel versatile image" [High,Triaged] https://launchpad.net/bugs/323134 [17:29] ogra confirmed though that if he disabled CONFIG_BLK_DEV_SR oops went away [17:29] ogasawara: are we tagging the arm bugs differently? [17:30] pgraner: I haven't been but can start [17:30] i do think it would helpful to have them marked [17:30] ogasawara: amitk would that help you? [17:30] ogasawara, thats fixed partially, amitk made it a module ... though now it oopses if the module is loaded [17:30] so i think we need to drop it completely [17:30] pgraner: yes it would help to have arm bugs tagged separately [17:31] ogra: noted [17:31] ogasawara, could we colour them different or something for background [17:31] apw: yup, am planning on coloring regressions already, so would be a similar change [17:31] always one step ahead of me [17:31] or editing the summary to add [arm]? [17:32] amitk: I'll edit the summary/title and tag them "arm" [17:32] ogasawara: thanks [17:32] listing just a few other bugs . . . bug 322879 is another 2.6.27-11 regression [17:32] Launchpad bug 322879 in linux "Verizon CDMA card no longer works" [High,Triaged] https://launchpad.net/bugs/322879 [17:32] amitk, you shouldnt say [noted] .... you should say ogra you goddamn slacker, get your ass up and write that in the bug :P [17:33] along with bug 319840 (hardy regression) [17:33] Launchpad bug 319840 in linux "Kernel panic when attempting to netboot hardy 20090119 on HP Proliant ML350 G5" [High,Triaged] https://launchpad.net/bugs/319840 [17:33] ogra: whatever you said ^^^^ [17:33] :) [17:33] :) [17:34] ogasawara: are the bugs assigned to individuals are still with kernel-team? [17:34] i wonder if there is milage to keeping the regressions in a new list with the this weeks ones, or always keep them on 'this week' [17:34] so they don't slide down and get lost [17:34] keeping them on 'this week regardless of week might work nicly [17:35] apw: good point [17:35] pgraner: bugs are no longer assigned to kernel-team, and is currently the devs responsibility to take ownership (is assign it to themselves) [17:35] apw: I was thinking it would be better to pull regressions out into their own section [17:35] * pgraner passes on ogasawara's reminder to take ownership of the ones she put out today :-) [17:35] ogasawara, that would work just as well as long its at the top [17:36] they are the most important bugs for smb_tp and his kin [17:36] * ogasawara nods [17:36] * smb_tp agrees [17:36] ogasawara: Anything else on bugs? [17:36] pgraner: there are few other high ones but I want to get them to test latest Jaunty first. I'll raise them in next weeks meeting if necessary. [17:37] ogasawara: ack, thanks [17:37] Moving on then [17:37] Since sconklin is out today we will skip LPIA, if anyone has questions or issues take it to the list pls... === kaptengu is now known as kaptengu_ [17:37] amitk: armel ... how goes it? === kaptengu_ is now known as kaptengu [17:38] [TOPIC] armel [17:38] New Topic: armel [17:38] versatile and iop4xx (slug) are now working [17:38] ogra beat me into fixing the kernels for both [17:38] amitk, ogra: do we have instructions for loading the image up anywhere? I have a few slugs I could test [17:39] I've left iop32x and orion5x flavours as 'don't cares' for now since there isn't a whole lot of community [17:39] amitk: ACK [17:39] pgraner: ogra is doing final testing now and should write up a wiki page soon [17:39] that's it from me [17:40] ogra: let me know when you have it done :-) [17:40]