[06:50] <Noskcaj> Logan_, has my packaging quality improved enough for a testimonial? If not, i've got at least 10 reviews waiting, plus another 10 in debian
[06:51] <Logan_> Noskcaj: I honestly haven't done much of your stuff recently
[06:51] <Noskcaj> I'm just going around asking everyone possible for testimonials so i can apply by email, since irc isn;t really an option for me at the times available
[14:19] <maxiaojun> any reason trusty always show "The update information is outdated ..." ?
[17:40] <cjwatson> doko: Can you figure out this python3.4 autopkgtest failure?  It's complaining about not being able to import runpy, but when I run the test manually the traceback ends inside Lib/runpy.py, so I'm pretty sure the exception is lying
[17:40] <cjwatson> doko: http://paste.ubuntu.com/6646639/
[17:41] <doko> cjwatson, yes, it fails during the build too. did you look at the others too?
[17:42] <cjwatson> doko: no, I was misled by the "failures=1".  But several of the others seem to be broadly the same pattern
[17:43] <doko> now generating locales for the test too
[18:04] <maxiaojun> anyone take care of software-center?
[18:33] <darklight_> is there a chance for this https://bugs.launchpad.net/ubuntu/+source/unity/+bug/775434 to be fixed? honestly it's a moronic design for shortcuts to be hardcoded and nobody seems to fix it release after release
[19:17] <vlad_starkov> Question: Installing Ubuntu Server 12.04 LTS 64bit. While installing it asked to choose the kernel. What should I choose? See sccreenshot http://cl.ly/image/1p2c0w043V3V
[19:17] <TJ-> vlad_starkov: please don't ask in multiple channels simultaneously
[19:18] <vlad_starkov> I never seen this message before
[19:18] <zyga> vlad_starkov: I suspect that is the hardware enablement thing
[19:19] <vlad_starkov> TJ-: I just thought that different channels have difference audience
[19:19] <zyga> vlad_starkov: you can choose a kernel from the more recent releases, to get hardware enablement for current hardware
[19:19] <zyga> vlad_starkov: aka 12.04.{123}
[19:19] <vlad_starkov> zyga: I'm installing Ubuntu Server on server (2xXeon, VT-x)
[19:19] <zyga> well, with commas inside { }
[19:20] <zyga> vlad_starkov: pick any that works with your hardware if you have no special needs (like you need a particular kernel feature only available in future kernels)
[19:20] <vlad_starkov> zyga: I'm pretty newbie to Linux and don't completely understand the kernel stuff
[19:20] <zyga> vlad_starkov: then just pick the -lts-raring kernel
[19:21] <vlad_starkov> zyga: I'm using 2xSSD + 4xHDD on server
[19:21] <vlad_starkov> zyga: which one of "-lts-raring" there are 4 of them
[19:22] <zyga> vlad_starkov: if you have haswell CPU pick more recent kernel
[19:22] <zyga> vlad_starkov: hmm
[19:22] <zyga> vlad_starkov: linux-generic-lsts-raring or -signed- if you are using secure boot
[19:22] <vlad_starkov> I have Intel Xeon E5345
[19:22] <vlad_starkov> zyga: I'm using RAID + encryption + LVM
[19:23] <zyga> vlad_starkov: the ones with explicit version numbers will be pulled in implicitly by the meta package that you are currently installing (if it makes no sense to you then just pick what I said)
[19:23] <zyga> vlad_starkov: also, this is a development channel you may want to go to #ubuntu for support
[19:23] <zyga> vlad_starkov: happy holidays :)
[19:23] <vlad_starkov> zyga: hey, thank you for your help
[19:24] <vlad_starkov> zyga: just last question, what does "secure boot" mean?
[19:24] <zyga> vlad_starkov: http://en.wikipedia.org/wiki/Unified_Extensible_Firmware_Interface#Secure_boot_2
[19:25] <vlad_starkov> zyga: I'm posting here because in #ubuntu nobody can give an answer for such tricky question. But I was sure that there is someone in dev channel who can
[19:26] <zyga> vlad_starkov: this is not a support channel, you may have better luck next year, many people are on hoildays
[19:26] <zyga> vlad_starkov: if you need you can alwas purchase paid support
[19:30] <vlad_starkov> zyga: I have nothing against paid support, I wish there is a per-question paid support :)
[19:31] <vlad_starkov> zyga: thanks for helping!
[19:31] <vlad_starkov> zyga: happy holidays!
[19:32] <zyga> vlad_starkov: there is support like that
[19:32] <zyga> vlad_starkov: http://www.canonical.com/enterprise-services/ubuntu-advantage/support
[19:32] <zyga> http://shop.canonical.com/index.php?cPath=41_39
[19:33] <zyga> vlad_starkov: you may also want to compare offers: http://www.canonical.com/enterprise-services/ubuntu-advantage/server
[19:33] <vlad_starkov> zyga: hm this is per year subscription
[19:37] <vlad_starkov> zyga: I'll consider authorized support next year
[19:37] <vlad_starkov> thanks for tip
[21:47] <vlad_starkov> Question: After install Ubuntu Server 12.04 64bit, got messages like "BUG: soft lockup – CPU#7 stuck for 22s". I can't even boot the system. Is it a known bug?
[21:49] <vlad_starkov> screenshot http://cl.ly/image/221D100o3W3S/o
[21:49] <infinity> vlad_starkov: Does it do this on every boot, to the point where you can't even get to userspace and login?
[21:50] <vlad_starkov> infinity: Yes, on every boot
[21:50] <infinity> vlad_starkov: That's not necessarily a known bug (soft lockups can happen for any number of reasons), but filing a bug on the kernel with enough info to debug would be helpful.  Being able to log in would be more helpful.
[21:50] <infinity> vlad_starkov: Which installation media did you use?
[21:51] <vlad_starkov> I have read somewhere that if I make "apt-get install intel-microcode microcode.ctl" it will solve the problem. But I can't install anything as I even don't have booted system.
[21:51] <vlad_starkov> infinity: Flash 8GB
[21:51] <infinity> vlad_starkov: Well, that advice may or may not be true, depending on the CPU and kernel version.  Like I said, soft lockups can happen for any number of reasons.
[21:52] <infinity> vlad_starkov: Sorry, I didn't actually mean "what media", but "what did you write to it"?  12.04, 12.04.3, desktop, server, d-i mini.iso, etc.
[21:52] <vlad_starkov> infinity: How to make sure that it is now HW problem?
[21:52] <infinity> vlad_starkov: Basically, trying to sort out if you installed a 3.2, 3.5, or 3.8 kernel, and what exact version. :P
[21:52] <vlad_starkov> infinity: ubuntu-12.04.3-server-amd64
[21:53] <vlad_starkov> infinity: the kernel was as selected on this screenshot http://cl.ly/image/1p2c0w043V3V
[21:53] <vlad_starkov> infinity: how can I check it?
[21:54] <infinity> vlad_starkov: Right, so, to get you out of your jam, I'd recommend installing from http://old-releases.ubuntu.com/releases/precise/ubuntu-12.04.1-server-amd64.iso
[21:54] <infinity> vlad_starkov: Which will give you a 3.2 kernel instead.
[21:54] <vlad_starkov> hmm ok
[21:54] <infinity> vlad_starkov: And then if you confirm that works, it might be nice for you to work with the folks in #ubuntu-kernel (perhaps after the new year, when everyone's back from holiday) to debug why 3.8 hates your computer.
[21:55] <vlad_starkov> but then if I make apt-get update will it install the newer kernel version and the system will fail again?
[21:55] <infinity> vlad_starkov: But that bit's up to you.  If you just need a machine running today and can't really sacrifice it to debugging, such is life.
[21:55] <infinity> vlad_starkov: No, update/upgrade will keep you on the 3.2 kernel if you install with it.
[21:55] <brainwash> my 13.10 system wasn't bootable either after I've upgraded the kernel from 3.12 to 3.13, same cpu soft lockup message
[21:56] <infinity> vlad_starkov: 12.04 and 12.04.1 use "linux-generic (3.2)", 12.04.2 uses "linux-generic-lts-quantal (3.5)", and 12.04.3 uses "linux-generic-lts-saucy (3.8)"
[21:56] <infinity> brainwash: To be fair, while the message was the same, the bug probably wasn't.
[21:56] <vlad_starkov> oh, I see
[21:57] <infinity> brainwash: But please do file a bug, or annoy the kernel people on IRC.
[21:57] <infinity> (or both)
[21:57] <brainwash> infinity: right, thanks to your conversation I got reminded to actually report this bug :)
[21:57] <vlad_starkov> infinity: when I reinstall Ubuntu should I repartition everything again? (I have encrypted disks)
[21:57] <infinity> vlad_starkov: That's between you and the installer. :)
[21:58] <vlad_starkov> infinity: I mean, the installer doesn't see encrypted raid disks. Is it possible somehow to unlock it, so installer will see everything?
[21:59] <infinity> vlad_starkov: I'm not sure, to be honest.  Probably a question better aimed at xnox.
[21:59] <infinity> xnox: ^
[21:59] <vlad_starkov> infinity: I mean I can go to console and make cryptsetup luksOpen on every md device. But not sure that installer then will see it.
[22:00] <infinity> vlad_starkov: If you create/mount partitions at a shell and then return to d-i, partman will see them as valid installation targets, yes.
[22:00] <vlad_starkov> Ok, I'll try
[22:01] <infinity> vlad_starkov: But you may need to do some manual setup in /target to generate an initrd that will unlock tihngs properly on first boot, I suspect the installer only "knows" about crypt setup for encrypted things it setup itself.
[22:01] <infinity> vlad_starkov: But I don't do crypted stuff terribly often, so that's just based on my knowledge of d-i and some educated guesses, not actual experience.
[22:02] <TJ-> In the target chroot, it requires an "apt-get install cryptsetup",  "update-initramfs ..." to install the cryptsetup and cryptab pre-requisites
[22:03] <vlad_starkov> infinity: by the way, on 13.10 I got the same messages
[22:04] <TJ-> vlad_starkov: It sounds like a potential IRQ issue, maybe needs kernel command-line options such as "nolapic"
[22:04] <infinity> vlad_starkov: Kay.  And I'd expect you would in 13.04 as well (since that's where your precise kernel was backported from).
[22:04] <infinity> If it happens with a 3.2 kernel as well, I'd agree that the next step it trying fun kernel commandline options.
[22:05] <vlad_starkov> TJ-: I tried nolapic, noapic and acpi=off. No effect
[22:05] <TJ-> vlad_starkov: I had a similar issue with a Dell PowerEdge box
[22:05] <vlad_starkov> TJ-: so how did you solve it?
[22:05] <infinity> Probably also possible to boot it with 'nosmp', at least long enough to get debugging info out of it.
[22:06] <TJ-> vlad_starkov: Have you tried booting it to a LiveCD image? Sometimes that can work when an installed system won't, which gives you a base to work from in solving the issue.
[22:06] <infinity> The important thing here is to realise that how other people solved it, while all interesting things for you to try, are not necessarily the solution for you. :P
[22:06] <infinity> "BUG: soft lockup" can be caused by any number of actual bugs.
[22:06] <TJ-> vlad_starkov: I seem to recall I installed the most recent kernel from main-line
[22:06] <infinity> TJ-: Well, he can boot the installer so, yes.  He can boot the kernel without a full userspace.
[22:07] <TJ-> infinity: Yeah... but with the liveCD it's easier to work with additional tools than from the server install ISO, I find anyhow
[22:08] <infinity> TJ-: Given how quickly his machine explodes after running init, I'd give 20-to-1 odds that the livecd (with the same kernel, anyway) would fail just as hard.
[22:08] <infinity> Unless the added latency from running from slow media avoided whatever spinlock races he's hitting by sheer force of luck. :P
[22:09] <vlad_starkov> TJ-: I tried ubuntu-12-1.04.2-desktop-i386 and it is the same
[22:09] <TJ-> It looks like something on the PCI-X bridge
[22:10] <vlad_starkov> infinity: Probably it makes sense to try ubuntu-12.04.1-desktop-i386 at first?
[22:10] <TJ-> the two devices locked up are chipset FSB registers, and the PCI-express to PCI-X bridge
[22:10] <vlad_starkov> TJ-: I have PCI-X there
[22:10] <infinity> vlad_starkov: Sure.
[22:11] <vlad_starkov> TJ-: probably I need to adjust something in BIOS? The machine is Supermicro 6015B-TB
[22:12] <TJ-> vlad_starkov: I've had success allowing the PCI subsystem to remap bridges where their windows are too small for devices within them: "pci=realloc" added to kernel command line
[22:12] <vlad_starkov> infinity: downloading ubuntu-12.04.1-desktop-amd64.iso
[22:27] <vlad_starkov> infinity: TJ-: param nosmp is not supported, param pci=realloc does nothing
[22:28] <TJ-> vlad_starkov: Shame, worth a shot. If you have physical access to the server it could be worth removing the PCI-X devices to determine if thats where the lockups are coming from.
[22:30] <vlad_starkov> TJ-: I'm in datacenter in front of server. There is no PCI-X devices
[22:34] <TJ-> vlad_starkov: OK, I've been there done that! If there are no PCI-X devices, and the hangs are in modprobe when loading drivers for the PCI-X bridge and system chipset, that suggests a shared IRQ imay be the root issue
[22:38] <vlad_starkov> ubuntu-12.04.1-desktop-amd64 boots so slow...
[22:38] <vlad_starkov> Almost done. "* Starting configure network device security [OK]"
[22:43] <brainwash> any reason why only trusty is stuck on firefox 25? firefox 26 is available via proposed
[22:43] <brainwash> https://launchpad.net/ubuntu/+source/firefox
[22:47] <maxiaojun> +
[22:52] <vlad_starkov> TJ-: infinity: Given I boot from installer flash drive. Then I go to console. Then I mount my /boot and / partitions to /media/sda1 and /media/sda2. Then I need to install microcode.ctl tool with intel-microcode packages to sda2. Is possible to implement it from installer console?
[22:57] <TJ-> vlad_starkov: You'd be better off mounting /boot/ inside / ... as in "mkdir /mnt/target && mount /dev/sda2 /mnt/target && mkdir -p /mnt/target/boot && mount /dev/sda1 /mnt/target/boot". That'll allow you to use chroot if necessary.
[23:00] <vlad_starkov> TJ-: will chroot help me to install software to my system on ssd?
[23:00] <vlad_starkov> TJ-: I'm not familiar with chroot
[23:01] <TJ-> vlad_starkov: Depends on whether you need to configure the target system from within the installer or liveCD environment. You may need it to edit and rebuild configurations if the server fails to boot
[23:02] <vlad_starkov> TJ-: I can't boot with liveCD. Installer is the only option.
[23:02] <TJ-> vlad_starkov: Another option for the kernel's command-line you might find useful is "debug" whilst ensuring that any "quiet splash" are removed, to ensure maximum kernel debug messages
[23:03] <vlad_starkov> TJ-: "debug" should be boot param?
[23:04] <TJ-> vlad_starkov: OK, so you're hitting an init issue. You might be able to capture more information using a combination of "debug" and "break=..." where ... is one of the break-points in the initrd init script, prior to the upstart init daemon being started
[23:04] <TJ-> vlad_starkov: correct
[23:04] <TJ-> vlad_starkov: It enables maximum kernel messages
[23:04] <vlad_starkov> TJ-: hmm, ok fine. I'll try to add debug in boot params.
[23:05] <TJ-> vlad_starkov: I'm just grabbing the list of break-points for you
[23:05] <vlad_starkov> TJ-: could you guide me with break-points (I understand the point but don't know how to implement it)
[23:05] <vlad_starkov> thnx
[23:10] <vlad_starkov> TJ-: grub says that "debug" is not valid option
[23:12] <vlad_starkov> infinity: TJ-: I have 3.8.0-29-generic kernel
[23:12] <TJ-> vlad_starkov: I believe you may have added the option as a separate line, rather than as part of the "linux ..." line
[23:13] <vlad_starkov> TJ-: yes, shouldn't I?
[23:14] <TJ-> vlad_starkov: For the "break=..." option to drop to a shell during initrd, these are the options: "top modules premount mount mountroot bottom init". They are in order from earliest to latest... I'd suggest trying the latest "init" first and if you get that far, you know the problem is caused when the real /sbin/init upstart daemon is started and the real root file-system is mounted
[23:14] <TJ-> vlad_starkov: The kernel command-line options must go on the "linux ..." line ... which often wraps around in the grub editor, which causes confusion
[23:14] <TJ-> vlad_starkov: All other lines are for grub itself
[23:15] <vlad_starkov> TJ-: funny))
[23:16] <vlad_starkov> TJ-: then I'll try all the params again, as I was mistakenly added them with the new lines
[23:16] <TJ-> vlad_starkov: Ahhhh! sorry, I assumed you already knew how to do this :)
[23:17] <vlad_starkov> )
[23:20] <vlad_starkov> TJ-: with "acpi=off noapic nolapic" it behaves a bit different. Is stops on "e1000e ... (unregistered net_device): Failed to initialize MSI interrupts. Falling back to legacy interrupts."
[23:21] <vlad_starkov> TJ-: and then "/sbin/modprobe -bv pci:........"
[23:21] <TJ-> vlad_starkov: Yes, most systems these days can't function well without ACPI ... I'd try "nolapic" on its own, not with "noapic", as another permutation
[23:22] <vlad_starkov> TJ-: ok, so should I remove "acpi=off" as well?
[23:22] <TJ-> vlad_starkov: I would, yes.
[23:23] <vlad_starkov> TJ-: trying...
[23:25] <vlad_starkov> TJ-: "nolapic" does nothing. Got soft lockup messages again
[23:26] <TJ-> vlad_starkov: OK ... are you also including "debug" at this time? I'd use that now and hope the kernel outputs more info which you can screen-capture
[23:27] <vlad_starkov> TJ-: nope. Have to try again with "nolapic debug"
[23:28] <TJ-> vlad_starkov: If you're not getting better results with "nolapic" I'd not use that anymore since it isn't helping
[23:29] <vlad_starkov> TJ-: hmm it seems debug doesn't output any additional messages
[23:30] <TJ-> vlad_starkov: shame, but when dealing with these issues its always good to have it on the kernel command-line in case it reveals more information. Did you also remember to remove any "quiet" or "splash" entries? Those will cancel out what "debug" does
[23:32] <vlad_starkov> TJ-: there is only "ro" param by default. Now try to add "debug maxcpus=1 "
[23:33] <vlad_starkov> Nothing
[23:33] <vlad_starkov> TJ-: OK, probably it's time to try chroot to "apt-get install intel-microcode microcode.ctl" on systems ssd
[23:35] <vlad_starkov> TJ-: As https://www.kernel.org/doc/Documentation/kernel-parameters.txt implies, I had to do "debug ignore_loglevel" instead of just "debug"
[23:40] <vlad_starkov> TJ-: Yay, now I have debug messages
[23:40] <vlad_starkov> "intel_rng: FWH not detected"
[23:41] <TJ-> vlad_starkov: That's what I love about linux, being able to see precisely what it is up to
[23:42] <TJ-> vlad_starkov: intel_rng report is a minor one, not related to the soft lockups.
[23:44] <vlad_starkov> Hmm, modules linked in: intel_rng(F-) floppy(F+).....
[23:44] <vlad_starkov> TJ-: what does all these F- and F+ means?
[23:44] <vlad_starkov> TJ-: "... e1000e: .... Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode"
[23:45] <vlad_starkov> TJ-: is it fine that my hard drives is in AHCI mode?
[23:47] <TJ-> vlad_starkov: AHCI is the better mode, usually
[23:48] <vlad_starkov> TJ-: should check jampers on motherboard?
[23:49] <TJ-> vlad_starkov: If the kernel is using AHCI mode for the drives that support it, that is fine.
[23:49] <vlad_starkov> TJ-: this problems makes the server overheat
[23:50] <TJ-> overheating sounds like a symptom of the CPUs not being idled
[23:51] <vlad_starkov> TJ-: yepp
[23:53] <TJ-> vlad_starkov: The (F-) are module flags; "-" means "being unloaded", "+" means "being loaded". "F" means 'forced'
[23:55] <vlad_starkov> TJ-: Now I tried "debug ignore_loglevel acpi=off". The systems boots fast until "microcode: CPU0 sig=0x6f7, pf=0x40, revision=0x69". After that it freezes, and then something new appeared: "Watchdog detected hard LOCKUP on cpu 4"
[23:55] <TJ-> vlad_starkov: makes sense, since without ACPI some core sub-systems won't be initialised correctly
[23:55] <vlad_starkov> TJ-: Ok
[23:56] <vlad_starkov> TJ-: so, what next? :)
[23:56] <TJ-> vlad_starkov: ACPI == Advanced Configuration and Power Interface, and it contains definitions of all the key hardware sub-systems and embedded controller(s)
[23:56] <vlad_starkov> TJ-: is it possible to disable all the CPU cores but one
[23:57] <TJ-> vlad_starkov: I'd try adding "break=init" to what you have already and seeing if you can get to the initrd shell prompt without seeing the CPU lockups
[23:58] <vlad_starkov> OK, I'll try it now
[23:59] <vlad_starkov> Strange thing.... 5 years ago this server used to boot normal with ESX