[02:10] hi. I am going through a mini cd install. can someone point me to some documentation explaining the differences between all these selections in the kernel selection? [05:44] Regarding the LTS Hardware Enablement stack. (1) linux-image-generic-lts-quantal is PAE-only, and there's no "non-PAE" version available, right? [05:44] But (2) it's possible to install 12.04.2 and then install linux-image-generic, correct? [05:45] If so, my next question is, (3) what should one do to avoid xorg incompatibilities when he does that? [05:47] We've been shipping a customized version of the 12.04 and 12.04.1 ubuntu CDs with gnome-fallback preinstalled for 400+ Greek schools, and we're wondering how to handle the hardware enablement stack with 12.04.2... [05:59] Ouch, xserver-xorg-lts-quantal Replaces: xserver-xorg [06:00] So it's impossible to have both -pae and non-pae kernels with LTSP clients with the new stack, as they'll need different xorg versions [06:01] One would have to maintain 2 separate chroots for that... :-/ [06:10] * alkisg hopes that if some schools have to install the non-pae version, it'll will work with the new xorg stack... [06:13] alkisg: It will *probably* work; possibly with the exception of plymouth. [06:13] RAOF: as long as it doesn't halt the boot process, we don't mind at all [06:14] Thanks, /me installs the new stack to test all that... [06:14] We won't have tested new X stack + non-pae kernel, but upstream theoretically cares that it runs on arbitrary kernels. [06:15] That's nice; at least if we test + report incompatibilities, we'll stand a chance of getting them fixed [06:17] * alkisg reboots to check [06:32] Apart from virtualbox-dkms crashing "Error! Bad return status for module build on kernel: 3.5.0-23-generic", everything else seems to work fine... === work_alkisg is now known as alkisg [06:38] Ah, and our "purge old kernels" script gets confused and tries to remove "linux-image-3.2.0-37-generic" because it thinks "linux-image-3.5.0-23-generic" is a newer version of the same flavor, while it's internally -pae instead [07:17] moin === smb` is now known as smb [08:34] moin [08:35] ppisati, risen from the near-dead already? :) [08:35] smb: yep, i'm ok [08:35] ppisati, Thats what they all say :) [08:35] smb: i bit of sore throat but i can handle it :P === alkisg is now known as work_alkisg === henrix_ is now known as henrix [09:45] apw, (no) [10:36] apw: gbp-pq [10:36] in git-buildpackage [10:41] apw to commit a patch you use dpkg-source --commit ... now it is in the logs === work_alkisg is now known as alkisg [11:18] smb: what are we up to wrt the linux-crashdump metapackage change to kdump-tools ? [11:18] smb: both kdump-tools & makedumpfile 1.5.1 are now in Raring archives [11:19] caribou, I think in the end people wait for you to tell them all is main to apply the patch I sent (and by tell I mean on the thread in the kt-ml) [11:20] smb: looks like a chicken & egg thing : kdump-tools will not make it to main unless a package in main depends on it [11:22] caribou, Ok, well, if that is made clear... [11:23] apw, would you apply the change if that is made clear. Should be a quick resolution as soon as she complains [11:23] (or Tim (though he is not getting in today), or Leann) [11:34] if the resolution is that it is supportable in main (wahtever is moving), then i would happily the patch to make it happen [11:36] apw: kdump-tools is moving. It's part of the makedumpfile source package which is already in main [11:36] smb, remind me of the patch title and i'll poke it [11:37] apw, [raring-meta] Update linux-crashdump dependencies [11:37] smb, ta === alkisg is now known as work_alkisg [11:49] smb, is there a bug open for this issue? i have the deeling there is? [11:50] s/d/f [11:51] apw, don think specifically... a blueprint there is .... maybe caribou has more memory [11:51] don't need one, but if there is one i would reference it in this commit [11:51] apw: I don't think there's a bug open for it, only a blueprint as smb indicated [11:52] ack thanks [11:52] apw: I can create one if we want to track it [11:52] no if it is not going back for SRU then we don't care [11:52] apw: smb: I don't think we'll want to bring that into Precise [11:53] no [12:00] smb, you want this uploaded 'now' ? [12:01] apw, no strong preference... or with the next time we would do it... Though then we could be surprised by it getting stuck for a bit... [12:02] yeah i think it is safer to shove it in now when things are clean and see what happens [12:02] gives us a few days to fix it before leann is on her next rebase [12:05] * smb nods and walks back to the kitchen === yofel_ is now known as yofel [13:42] smb, ok your bits are in -release [13:42] * smb hopes not [13:55] infinity: hi! any news on the bugs verification? [14:00] smb, i would have thought you would have welcomed some -release [14:01] apw, not sure I really want that for my bits... [14:04] smb, blame britney she did it to you [14:05] ... [14:28] apw: i have a fague recollection that you possibly tried "ubiquity LVM install" and it "did nothing". I just checked raring-ubuntu-desktop and it does lvm install correctly. Was there something special about your lvm test? (e.g. were you testing a flavour?!) [14:28] s/fague/vague/ [14:33] fague is a nice word too though [14:33] :) [14:36] henrix: Not yet. Wrong timezone for people to get sagari happy in Boston. [14:37] infinity: ack [14:42] xnox, I think I sat in the same room when we tried. What we did was basically selecting lvm install but then trying to manually partition. And there we could not set partitions to be lvm pvs. Nor get any assemble VGs screen to show. [14:43] iirc that was normal desktop image [14:44] smb: true, manual partitions doesn't have LVM yet. I have some code, but it would be very d-i style LVM partitiong which can be confusing. [14:44] only automatic install can do lvm in the desktop image so far. [14:46] Yeah, so our usecase would still fail. :) Tried him to have his disk partitioned into two vgs... foolish of me. ;) [15:09] * ppisati disappears for ~20mins === zequence_ is now known as zequence [15:35] smb: *sigh* I do have lvm on top of full disk encryption, and then later a shrunk and split that to create an unencrypted lvm vg [15:36] (for vm/schroot/scratch partitions) [15:39] xnox, Heh, yeah. Well in the end we just went for a non-lvm root and swap partition and left another one unused for later. Beside its probably something that only people use that use the server install anyway. :) [15:40] * xnox totally uses lvm and multi-partition and stuff [15:40] no surprise I added ubiquity support =) [15:40] there are requests for better mdadm support from $enterprise as well [15:43] * ogasawara back in 20 [16:23] aloha - having a lot of trouble with https://launchpad.net/bugs/1011623 any idea of any work around, or an eta on it being fixed. [16:23] Ubuntu bug 1011623 in linux (Ubuntu) "8086:0091 Unstable wifi connection when connect to 802.11n AP with iwlwifi driver" [Medium,Confirmed] [16:24] sforshee, ^ this seems like a common bug for iwlwifi. i don't have any of that here right now [16:25] bjf, I've got some but nothing running the precise kernel. I can give it a try here in a little bit. [16:25] bjf, is this a regression? [16:26] sforshee, from the bug, it doesn't seem like it. seems like this has been a problem always [16:26] oh, I see the comment that it's still a problem in quantal/raring [16:27] though in raring there was a problem that has a fix, not sure if it's trickled down to us yet though [16:27] sforshee: ah good to know [16:27] I thought if I upgraded to raring it might be better [16:27] but just as bad as it was on quantal [16:27] bjf, czajkowski: so for a while we know that some people have had problems with 802.11n and iwlwifi, but it' [16:28] but it's very dependent on the particular machine and the environment [16:28] I've never been able to reproduce it with what I have, nor have the developers at Intel [16:29] sforshee: not sure it was happening on the cisco AP router but definately happens on UniFi AP [16:30] czajkowski, I've been sitting next to people having problems on their machines while on my machine it's fine [16:31] sforshee: yeah don't be telling me this :) [16:31] causes me no end of grief when I go over to my other halfs place and spend half an hour or more trying to get on the network to do work [16:32] bjf, czajkowski: the 3.8.0-6.11 raring kernel has the fix for the regression that was in 3.8 [16:32] oh that regression was in 3.7 too [16:33] ok, so the currently released raring kernel would have that fix [16:33] yeah, but that's fixing a regression from 3.7 so it doesn't help precise/quantal, and may not actually fix czajkowski's problems at all [16:34] it fixed problems I was seeing though [16:34] ack [16:34] we basically turn N off on precise [16:35] because it works poorly [16:35] I have an internal bug to 'figure out how to make it work, because N is awesome' [16:35] but its pretty low priority ;p [16:36] czajkowski: you say the quantal kernel fared no better in your case? [16:37] so in reality there's likely more than one issue that's causing people's problems with n [16:37] one aspect of the problem seems to be noisy environments though [16:37] 3.8.0-6-generic #13-Ubuntu SMP Thu Feb 14 17:22:41 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux [16:37] we get a lot more reports of iwlwifi problems from enterprises and people at confereneces [16:37] *conferences [16:37] is what I'm currently on [16:38] okay, that should have the fix for the 3.7 regression [16:38] great [16:38] but I'm still havng issues [16:38] so must be something else [16:38] czajkowski, so it's all fixed, have a good day [16:39] heh [16:39] the other alternative is to get the other half to turn off N but then I've to listen to him whinge he cant use it [16:39] *sigh* [16:40] czajkowski, are you in an environment with lot's of people using wifi or lots of APs around? [16:40] sforshee: well I mean people want to use N on the google bus, and let me tell you, the google bus network sucks, terribly [16:40] sforshee: nope - house, only me here. nobody else [16:41] okay, so maybe something related to your AP then [16:42] sforshee: will look into that, but no isues on iOS or fedora just ubuntu :/ [16:43] czajkowski, yeah, not saying the AP is at fault, just that maybe something it does exposes some iwlwifi bug or something like that [16:43] nods [16:44] is there any more info I could get that would help debug it ? [16:44] for you guys [16:44] czajkowski, so you get the same sorts of messages as the original reporter on LP#1011623? [16:44] bug #1011623 [16:44] Launchpad bug 1011623 in linux (Ubuntu) "8086:0091 Unstable wifi connection when connect to 802.11n AP with iwlwifi driver" [Medium,Confirmed] https://launchpad.net/bugs/1011623 [16:46] czajkowski, does N work for you if you rmmod/modprobe ? until next suspend or wifi off event ? [16:46] I've never seen an error message, it conects and hangs.. [16:47] apw: yup [16:47] exactly [16:47] once I learnt that my life was alot smoother [16:47] czajkowski, then your issue doesn't sound liek the one in that bug to me, that sounds like it breaks in the middle [16:47] :/ [16:48] czajkowski, then please file a new bug by running 'ubuntu-bug linux' after you've been experiencing those problems and assign it to me [16:49] sforshee: will do and thanks for the help folks [16:49] czajkowski, probably we should focus on raring to start because it's easiest to get upstream to pay attention to the latest kernel version [16:52] sforshee: lemmie go and reboot and not start wifi and see if it works [16:52] knowing my luck I'll have no issues [16:52] heh [16:54] perfect [16:54] typical === JanC_ is now known as JanC [17:23] * ppisati -> gym [17:51] apw: can an overlayfs be rebased from one base image onto another, with a goal of reducing amount of space used, since content on the new base matches content in overlayfs better. [17:52] xnox, rebased by like unmounting it and mounting it over something else ? [17:57] apw: well kind of, apart from I need to a de-dup. Eg. base-1: /blob1 base-2:/blob1.1, and overlayfs: rm /blob1; wget http://example.net/blob1.1 [17:58] apw: such that when overlay is moved from base-1 to base-2, it now should be empty and not store the copy of blob1.1 (identical and present on the new base) [17:58] or does it do this already? via lazy compare and discard [17:58] one may want to ignore modification/creation timestamps etc. === chuck_ is now known as zul === henrix is now known as henrix_ === henrix_ is now known as henrix [18:10] it does none of the things you are hoping for [18:11] xnox, ^^ [20:13] * ogasawara lunch === henrix is now known as henrix_