=== rbz_ is now known as rbz [05:02] eggonlea: you around? === JaMa|Zzz is now known as JaMa === XorA|gone is now known as XorA [08:14] While using ubuntu-netbook installation with debian-installer I found error saying /cdrom is not able to unmount. [08:14] morning [08:14] to unmount? [08:15] Is there any alternate way for latest lucid installation on beagle board ? [08:15] unmount [08:16] Should I write complete message ?? [08:16] no [08:16] I just had a problem with ;'how to tell ubuntu installer that there is no cdrom' [08:16] "The Installer needs to commit changes to partition tables, but [08:17] cannot do so because partions on the following mount points could not be umounted: [08:17] [08:17] [08:18] "/cdrom" [08:18] I am using lucid-netbook-armel+omap.img from http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-live/current/ [08:19] my beagleboard refuses to see usb devices so I did not installed yet [08:19] What is problem ? [08:20] hubs (one on ehci, one on otg) got disconnect when rootfs is mounted [08:20] so you are using USB HDD ?? [08:21] no [08:21] pendrive, keyboard, mouse, ethernet [08:23] ok [08:31] hmm, sounds like he tried to install to the SD he is installing from [08:37] lool, heh, remember the qemu segfault i showed you yesterday ? turns out the user changed rootstock to make qemu use 512M :P [08:45] I want to test any Lucid image installation on Beagle Board. Which is the best method ? [08:52] if he wouldnt always vanish after asking i could actually answer :P [08:54] I am installing kubuntu now... but on x86-64 and from usbstick not sd [08:55] that takes out all the slowness fun [08:55] but still require some hacking if you want to have crypto lvm [08:56] yeah, use alternate for that if you are lazy :) [08:57] if you have cd [08:58] iirc usb-creator should be able to turn alternate into a USB image [08:58] anyway I had usbstick with lucid/x86-64 ubuntu which I used for tests so I just unpacked alternate cd into it [08:58] usb-creator was a tool which I found after making bootable stick ;d [08:58] heh [08:59] first one was made with usb-creator so it was just change of contents of stick [09:00] yeah [09:00] and then google for 'how to tell installer that there is no cdrom drive' [09:01] * ogra vanishes for more coffee [09:36] Amit_Karpe, hey [09:36] cwillu_at_work, ping... (thumbs up or down on the 2.6.33.3 image on your overo? ;) ) [09:39] hi [09:39] hey ogra, been talking with a couple guys at esc... they've been getting some cdrom mount issues with the Netinstall, have you seen anything similar or know what they are doing wrong? ;) [09:40] Yes, I am facing same prob ( cdrom mount ). [09:40] nope, i havent had such issues, did they use the shipped boot.scr file ? [09:40] not sure... and i can't test ideas since my beagles are 13 hours away..... [09:40] it sets a preseed option you need [09:41] if they dont use the boot.scr they should at least make sure to use the cmdline thats set in there [09:42] I use image and instruction from https://wiki.ubuntu.com/ARM/BeagleNetbookInstall [09:42] he normally runs Angstrom, so it's enabled for boot.scr... i know that much... I'm assuming he copied all three files... [09:43] Amit_Karpe, netbook != netinstall :) [09:43] And after booting I can see debian-installer screen, after partition I get error [09:43] Amit_Karpe, how is your HW setup [09:44] netbook needs to install to a USB key/disk [09:44] you cant install to the SD you boot from with it [09:44] k [09:44] Got it. [09:44] ahh.. that makes sense.. i bet he did that too... need a quick faq on the Netbook install page.. ;) [09:44] only the netinstall variant supports that [09:45] (netinstall is small enough to live completely in RAM so it doesnt need the SD after boot, netBOOK is a live system that needs the SD mounted the whole install time) [09:46] The NetbookInstall method is going to be radically different for some of these angstrom guys. ;) (using a usb stick to install to sd) [09:46] i though the notes i added at the top of the wikipages were enough .... seems i need to rephrase that [09:46] it's just new and different.. ;) [09:46] yeah [09:46] i didnt take that into account :) [09:47] i think d-i is also strange for people not used to it [09:47] now I am getting little bit :) [09:48] Yeap it is.. in the debian wiki i just ended up walking people thru every d-i step.. talking with my boss's and ti, i think we are going to do a youtube video to demostrate the NetInstall method... [09:49] yeah, might make sense [09:49] though mind you, netinstall is really slow if you want more than just a minimal system [09:50] d-i is easy [09:50] i tried an ubuntu-netbook install with it yesterday, even with local package proxy that took 6h to USB key [09:50] boot-floppies were ugly [09:50] heh, yeah [09:50] BTW: ServerInstall the same which launch d-i ?? [09:50] yeah it is... It's reliable for x11 image's at the moment.. ;) ouch... 6hours? crazy.. [09:51] Amit_Karpe, server install uses d-i as well, yes [09:51] ogra: 512m > ah that explains why it sounded familiar [09:51] i would recommend netinstall for having a minimal developer system with openssh server if you have network [09:52] k [09:52] server install for the same if you dont have network [09:52] say, still catching up the irc logs... where we luckly enough to get the rtc_ignore boot variable into lucid for today's release? [09:52] and the netbook install if you really want to play with the netbook UI [09:53] rcn-ee, nom we use a different approach ... i added a function to initramfs, just use the ficrtc bootoption [09:53] *fixrtc [09:53] it sets the clock to the last mount time of the disk before running fsck [09:54] okay fixrtc, wasn't sure on the name... just add it to the bootargs's i'm assuming? [09:54] yes [09:54] the ubuntu installers do that by default [09:54] for maverick upstream will fix fsck properly [09:55] sweet.. i'll remove my tweak and use that for the image tomorrow.. ;) [09:55] server install : will help me if my Lan to USB is not working . [09:55] right [09:55] it has all server packages in the image [09:55] I had tried netinstall but my network was not detected. [09:56] Amit_Karpe, which device by chance? [09:56] Amit_Karpe, yeah there is a bug with the linux package [09:56] it doesnt include all USB NIC drivers it should [09:56] asic and pegasus NICs will work [09:58] JP108 .... [09:58] * ogra_cmpc sees ericm|ubuntu joining and waits for ericm|opensuse and ericm|fedora [09:58] devicom DM 9601 USB Ethernet [09:59] *davicom DM9601 USB Ethernet [09:59] yeah, probably no driver in the udeb [09:59] * ericm|ubuntu sees ogra_cmpc there and waits for ogra_beagle and ogra_babbage :-) [10:00] haha [10:00] ok. [10:00] Then any other way .... [10:00] use server or netbook installs [10:00] ah yeah the evil 9601... Amit_Karpe does yours get hot enough to boil coffee? With my NetInstall script it got about half way before over heating..... [10:01] or find another NIC [10:02] dm9601 are nice cause they are cheap [10:02] but only 7Mbps transfer ;( [10:02] I have 3 of those [10:03] I have not experience, such high heat. [10:03] fine, I will try another NIC. [10:05] ogra_cmpc: use USB hard disks :) the kubuntu-netbook netinstall took me less than 2 hours, and most of the time it spend on the downloads [10:05] ogra_cmpc: Do you have qemu versatile kernel for Lucid ? [10:05] well, i have a local package proxy, download took only 10mins :) [10:05] Amit_Karpe, yeap [10:06] its linked from the rootfs from scratch page (see topic) [10:07] wget http://ports.ubuntu.com/ubuntu-ports/dists/lucid/main/installer-armel/current/images/versatile/netboot/vmlinuz [10:07] RIGHT [10:07] thx [10:07] oops [10:07] sorry for the caps [10:09] so that is uploaded on 26-Apr-2010 . [10:10] yep [10:10] current/ always points to the latest [10:11] Few days back I was trying creating armel rootfs tarball for Lucid ( by following instructions http://elinux.org/BeagleBoardUbuntu ) [10:11] k [10:12] But every time it face some & other problems. some mouthplay error while booting ..... [10:12] or some problem creating image ( saying gnome package problem )...... [10:12] mountall you mean [10:12] Are those all problems solved ? [10:13] yeah, rootstock (or rather qemu) has a bug that it hangs when installing bigger tasks [10:13] we still havent found a solution [10:13] k [10:13] so you need to install ubuntu-minimal and then add on top what you need in a VM or on real HW [10:14] yes [10:14] but that way should work fine [10:14] It will be great if I can test all apps in qemu then I would like to install on beagle ...... [10:14] i need to fix a bug rcn-ee introduced though, that will go into an SRU soon [10:15] so lets say latest Lucid prebuilt image should work ? [10:15] (if you try to use a kernel deb there will be issues atm) [10:15] if you build minimal it will work fine, yes [10:15] ogra_cmpc, which issues, it's working for the stuff I'm doing at the moment.. [10:16] rcn-ee, your determining of the kernel version uses a bash function [10:16] rootstock isnt a bash script ... its posix compliant [10:16] (we call that a "bashism" in debian/ubuntu world) [10:17] rcn-ee: use rootstock with dash [10:17] ahh.. that's very possible then... By using Angstrom's scripts on all my development PC's dash has long dissappeared, so yes, i would have missed that... crap... [10:17] rcn-ee stuck between two communitiess... [10:18] KVERS=`chroot ${ROOT} dpkg -l linux-image-2\*|grep ^i|awk '{print $2}'|sed 's/linux-image-//'` [10:18] So can I say if I used prebuild image with you suggested qemu kernel then I must able to see basic Lucid in qemu. [10:18] thats a posix compliant way we use in the livecd-rootfs script [10:18] can easily be adjusted to work in rootstock [10:19] Amit_Karpe, yes [10:19] rcn-ee: Angstrom shouldnt have any bashisms either, at one point we were actually totally clean [10:19] Thanks ogra_cmpc & rcn-ee . I will try and update you. [10:19] what shell does angstrom use ? ash/busybox ? [10:20] or also dash ? [10:20] that should work, all my builds are based off deb-pkg so the dpkg -l flag should work... pre 2.6.31 i was using make-kpkg, which would lock rootstock.. [10:20] ah [10:21] Not sure what Angstrom uses's inside the beagle.. but the build scripts check for bash/dash and recommend to use bash over dash... Now wheither the scripts are complient now and the warning isn't needed i don't know.. [10:22] the key is, user addes a random http link, we pull out the kernel image name, extract/install that, and build a uInitrd... [10:23] right [10:35] ogra_cmpc, (playing with dash) if we only do a 'dpkg -x' vs's a 'dpkg -i' does it still show up with 'dpkg -l' [10:36] nope, you need to extract it from /boot/vmlinuz [10:36] ogra_cmpc: ash from busybox mostly [10:36] ah [10:36] ogra_cmpc: but many users install bash [10:37] ash shouldnt be so much different from dash [10:37] ah true, we can get the version from that, since there should be a /boot/vmlinuz-2.6.32... [10:37] right [10:37] just split at the dash [10:39] so we just scan that, there is a possibilty of two (versatile's from qemu and the extracted deb's) so update-initramfs might have to do both.. [10:41] rcn-ee: the issue is certain upstreams kept "fixing" their libtool to be full of bashisms [10:41] well, you can always call it under bash [10:41] its just ugly but will work [10:42] Yeap, and then some developers like me don't notice that and help bring some of those into dash systems.. ;) [10:42] in ubuntu you always have both shells installed, dash is just the default for all system script since its a lot faster [10:42] as long as your script uses /bin/bash at the top its no issue at all [10:51] ogra_cmpc, Removing lines 380 and 381 and setting KERNEL_IMG_REL=`ls /boot/ | grep vmlinuz|sed 's/vmlinuz-//'` seems like it would work, unless there is more then two vmlinuz's '|tail -1' would get hopefully get the later then versatiles... [11:00] "hopefully" doesnt sound so good :) [11:00] do your kernels use the -omap suffix like ubuntus ? [11:02] They are always formed like... linux-image-2.6.34-rc5-0_1.0sid_armel.deb [11:03] hmm [11:03] i mean the vmlinuz file [11:04] in ubuntu its always /boot/vmlinuz-$version-$flavour [11:04] well [11:04] /boot/vmlinuz-$version-$abi-$flavour [11:04] they end up being: ls /boot/ | grep vmlinuz -> vmlinuz-2.6.31.5-x5.1 [11:05] where, $flavour, is my tweak bump... [11:06] although, a 'uname -r' would give us the versatile kernel, so we can grep that one out.... [11:07] KVER=$(basename $(ls -1 /boot/vmlinuz-2*-omap|tail -1)|sed 's/vmlinuz-//') [11:07] that would be a safe one for ubuntu kernels [11:12] true.. on mine i use vmlinuz-2*-x* (jaunty/karmic) vmlinuz-2*-l* (lucid and probally maverick) with a -dx* and -dl* for development kernels, but it would be nice to check for that for any other 3rd party building kernel package via deb-pkg... [11:12] though i would do it differently [11:13] kernelfile=$(basename $(ls -1 /boot/vmlinuz-2*-omap|tail -1)) [11:13] kver=${kernelfile#*-} [11:13] saving the sed invocation [11:16] actually, when running rootstock, does a /boot/vmlinuz-* for the versatile kernel actually show up in the root file system? [11:17] no [11:18] it is only used on the qemu cmdline [11:18] /boot is empty by default (note that rootstock was never supposed to handle kernels, it should only build a rootfs) [11:19] okay.. then we don't have worry about that one.. it's only if the package selection automaticilly adds one do we get a second /boot/vmlinuz-* entry.. === bjf is now known as bjf[afk] [11:20] so: http://pastebin.com/F2EB2haD should be enough.. wish i was home with a fast pc/net to test... [11:23] and watching the NetInstall, it looks like it grep's the cpu core before ever extacting any of the linux-image-(dove,freescale,omap) so /boot/vmlinuz should remain empty from those... [11:25] amitk: ping [11:26] ya [11:27] ping was not for me sorry. [11:31] Mike^: hi [11:32] amitk: I have some patches to enable cm-t35 in ubuntu kernel and really ugly patch that should make beagleboard boot with CONFIG_OMAP_MUX enabled [11:34] Mike^: send it to kernel-team@lists.ubuntu.com and we can review them [11:34] amitk: 10x [11:34] Mike^: also, file a bug in LP so we have a way to track this problem [12:01] ogra_cmpc: I got error for prebuilt image --> [12:01] "init: playmount main process (44) killed by SEGV signal [12:01] init: playmouth-splash main process (215) terminated with status 2" [12:01] Amit_Karpe: Fixes for this have been identified [12:02] Amit_Karpe: https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/567592 [12:02] patches from cjwatson [12:02] k [12:02] Will take some time til they land in ubunut [12:02] Launchpad bug 567592 in plymouth (Ubuntu) "rm: cannot remove `/var/lib/urandom/random-seed': Read-only file system (affects: 2) (heat: 14)" [Undecided,New] [12:02] it was identified as not breaking things too dramatically, but it theortically could [12:03] its definately ugly though [12:03] we should have switched to plymouth in karmic [12:03] ogra: Ok, let me travel back in time! [12:03] but who could predict how ugly it is :) [12:04] now I have to wait till things get settle down :( [12:04] Amit_Karpe: Why? Does it cause any side effect? [12:04] it shouldnt [12:05] I just stops [12:05] BTW: Did I tell, that I am using that in Qemu. [12:05] oh, k [12:05] Also same image I was using on BB, it stops [12:05] Amit_Karpe: The stopping is unrelated [12:05] Amit_Karpe: Do you have network setup? [12:06] in the graphical qemu or with --nographic ? [12:06] Amit_Karpe: /etc/network/interfaces with an eth0 stanza, or network-manager installed? [12:06] in graphical [12:07] No. auto lo [12:07] iface lo inet loopback [12:08] Amit_Karpe: You might want to add an eth0 one as well [12:08] With hardware it says filesystem check [fsck] error 4 ...... and it stops [12:08] but lo is the critical one [12:08] ah, fs check [12:08] Amit_Karpe: Do you have a RTC clock? [12:08] So while booting with BB prebuild iamge i.e. RC minimal stops coz of filesystem error. [12:09] on BB ? [12:09] Where should I check for RTC clock ? [12:10] sorry, I meant RTC clock battery [12:10] does your beagle have a soldered battery on it? [12:13] no [12:13] BT1 is free [12:14] lool, This is my BB error for prebuilt image : http://pastebin.com/jW7FMvKN [12:15] Amit_Karpe: That's an unclean fs, needs to be fsck-ed manually [12:15] fsck it on your desktop or recreate it or fix it manually? [12:15] k [12:16] Amit_Karpe, thats all with rootstock images, right ? [12:16] Amit_Karpe: You want the "fixrtc" cmdline option to avoid the fsck on boot too [12:16] right, i was about to say that :) [12:18] so I should add "fixrtc" into ubuntu.cmd file ?? [12:18] k [13:59] amitk: ping [14:01] sebjan: pong [14:02] amitk: Hi, I would like to take on of our 'standard' defconfig file and change it to anú'ubuntu' one [14:02] amitk: do you have directions for this, or a tool to help? === ericm|ubuntu is now known as ericm-Zzz [14:04] sebjan: right, I'll answer your email now. Been a little busy with testing our release going out today. [14:05] amitk: ok, thanks! === XorA is now known as XorA|gone [14:48] damn, I missed him === bjf[afk] is now known as bjf [15:19] * ogra really enjoys the silence in here compared to #ubuntu-release-party [15:21] :-) [15:24] ogra: reading the release-party channel is an exercise is speed reading.. [15:25] *in [15:25] like booting without quiet and splash in lucid :) [15:25] heh, yeah a linux bootup [15:25] ogra: try ubuntu-mobile if you really wanna relax [15:25] lol, yeah === jmcgee|gone is now known as jmcgee === bjf is now known as bjf[afk] === bjf[afk] is now known as bjf === JaMa is now known as JaMa|bbl === hrw is now known as hrw|gone [17:19] JamieBennett: hi there, do you know when the london launch party starts? [17:19] dmart, i'd bet once we launched :) [17:19] dmart: no idea :) [17:19] approximately? [17:20] dmart: I'll find out [17:20] thanks! [17:20] dmart: http://loco.ubuntu.com/events/team/121/detail/ [17:21] 8pm then [17:21] * dmart gets confused by all this UTC [17:22] dmart: just remember we are +1 atm ;) [17:22] heh [17:22] Indeed. Remembering is the hard part :) [17:22] :) [17:22] Especially when the times look nearly right... if we were +10 is would be harder to forget! [17:23] anyway, thanks [17:23] you should just sell greenwich to australia or so :) [17:24] wow, over 1200 ppl in #ubuntu-release-party ... thats definately a new record [18:03] NCommander: hey there, I found something on the OOo build issue... [18:03] dmart: ack, OOO! [18:03] * NCommander runs through a wall in fear [18:03] dmart: (whatcha find?) [18:04] Ramana couldn't see why -fno-omit-frame-pointer was being used... I rebuild the segfaulting object without it and that made the segfault go away [18:05] I'm doing a full rebuild to see what happens [18:05] iirc thats a defualt [18:06] It is, but we're not sure why. I grepped for problem asm poking the "fp" or "r11" registers, but found nothing interesting [18:06] The cpp_uno bridge's makefile explicitly adds -fno-omit-frame-pointer back, but it might be genuinely needed there. I haven't attempted to touch that. [18:07] Anyway, I'll let you guys know if anything interesting happens... [18:08] dmart: if it works, we'll look at an SRU for OOo *shiver* [18:08] dmart: thanks for putting engineering resources on this [18:11] NCommander: haven't really spent much time, but it at least seemed worth reproducing the segfault and trying to see how it falls over. Dunno if the changed compiler options will really fix it, but we'll see [18:11] wow, 1400 people in #ubuntu-release-party [18:11] thats definately a new record [18:11] dmart: we can put our heads together at UDS [18:11] sure [18:11] Ramana will be there too [18:11] ogra: 1401 now ;-) [18:11] yeah, saw you :) [18:11] dmart: oh good, I owe him a beer [18:11] ogra: I joined to make up the numbers, but it's all in a foreign language ;) [18:11] its scrolls so fast, easy to miss :) [18:12] dmart, lol [18:12] Full of guys joining to ask how many people there are in the channel ;) [18:12] dmart: ahahah [18:12] dmart: we're getting populary, what can we say [18:13] yeah, or to ask "is it out yet" :) [18:13] Someone suggested a bot for that [18:13] there is a bot for a response to that but i think it doesnt cope anymore and they took it down [18:14] I notice there are armel images already on releases.ubuntu.com/lucid :) [18:14] * ogra loves this moment where you can actually *see* the community every release [18:15] ogra: indeed. [18:15] dmart: \o/ [18:15] means we should be releasing immidately [18:15] :D [18:15] I should find a release party in NY [18:15] * cwillu_at_work kicks dmart from the channel for posting links :p [18:15] bad? [18:15] dmart, try it in #ubuntu-release-party [18:15] you'll last about 3 seconds :) [18:15] nah [18:16] NCommander, nope, thats only the main mirror, release will happen if images were populated over to the others [18:16] Does it jinx the final announcement, or do people post bogus links? [18:16] slangasek does the announcement [18:16] people post all sorts of things [18:16] ogra: I can't see them doing a respin this late [18:17] and he will do so in a muted #ubuntu-release-party channel once its ethere [18:17] NCommander, no respin but mirror population === JaMa|bbl is now known as JaMa [18:17] * dmart pulls the imx51 image just in case [18:17] the announcement will point to the mirrors, so they need to have the images first [18:18] ogra: wow, its really getting people at a staggering rate [18:23] there we go [18:24] poor partybot didn't get the word :p [18:25] heh [18:30] * dmart upgraded days ago to beat the rush :P [18:31] bye guys [20:22] saeed: you around?