=== Monthrect is now known as Piper-Off === wolsen_ is now known as wolsen [06:40] Hey everyone, sorry to bother y'all. I recently bought and built a new computer. However it has the new "Secure Boot" from Asus. No matter the install method I do I can't see to get ubuntu server working properly. The closest I got was just a few minutes ago when everything seemed to go fine. Then when it reboots I see the first few lines of code p [06:40] rocessing then a blank screen. Any suggestions? [06:41] logixmorg: Some details like version of Ubuntu and model of motherboard may help [06:42] Ubuntu Server 14.04 LTS version [06:42] Asus Z170k motherboard [06:43] logixmorg: Z170 is new hardware which 14.04 point release [06:43] it might not recognize stuff as that is skylake [06:44] Ahh well that is unfortunate, so the i5 I have on the motherboard is considered skyline? [06:46] I might try 15.10 [06:47] Thank you, I'm downloading it now and will report back (Hopefully with success) [06:59] logixmorg: well 14.04.3 should have the same HW support as 15.10 [06:59] iirc [07:00] That's really unfortunate [07:00] !hwe [07:00] The Ubuntu LTS enablement stacks provide newer kernel and X support for existing LTS releases, see https://wiki.ubuntu.com/Kernel/LTSEnablementStack [07:00] hateball: I think 14.04.3 is same as 15.04 [07:01] as there has not been a lts point release since 15.10 came out [07:01] I think that is like feburary [07:01] So as far as the hardware I have now, support may not be released till February? [07:01] ianorlin: ah yes [07:02] logixmorg: you can install 15.10, but you will need to get 4.3 kernel from the mainline ppa for full support [07:04] hateball: what is in 4.3 that is not in 4.2 [07:05] Forgive my ignorance but how would I combine the 4.3 kernel with the 15.10? Would I open it up using my windows computer and simply delete the kernel then had the 4.3 kernel in? [07:07] ianorlin: support for skylake gpu iirc [07:09] OH so the on board graphics is the problem here? Would buying a compatible graphics card that is not skylake be a work around for this problem? [07:09] you can boot with nomodeset and then add mainline kernel ppa, all according to google [07:12] well I think you could nomodeset and install server [07:18] Good news!! A simple install of 15.10 did the trick [07:18] Thank you for the help and suggestions :) === InfoTest1 is now known as InfoTest [09:28] what is the size of a trusty mirror atm ? [09:33] I'm here now at around 370GB but that is a lot of you ask me and I think sources are included also [10:04] smoser, I had another run at https://code.launchpad.net/~james-page/simplestreams/multihypervisor/+merge/278127 [10:04] optionally enabled and also sets hypervisor_type in resulting streams data so that external tools can query by [11:38] smb, hey - remember that problem we kept hitting on utopic where umounted block devices got help onto by the kernel? [11:39] smb, http://paste.ubuntu.com/14003285/ [11:39] jamespage, erm... I vaguely remember some umount problems somewhen in the past but I have troubles parsing your line there... [11:40] Ah ok.. the case where something managed to get into a state where the journal would not go away [11:44] smb, yah - that's the one [11:45] jamespage, So did you find a reliable way to get into the state? iirc we could never isolate the reasons [11:46] smb, I see that on a trusty daily image [11:47] jamespage, That is not exactly an answer to my question... ;) [11:48] smb, we'll I've spun up 10 machines today all of them have the same problem; just re-trying with the release stream image to try to bisect [11:48] jamespage, OK, spun up means using maas or uvtool or something else? [11:49] smb, nope these are all ontop of an openstack cloud [11:49] so using juju [11:49] Best Ubuntu VPN server software for Ubuntu server? [11:52] T3DDY: anyone you like [11:53] people have different perceptions of "best" === cpaelzer is now known as cpaelzer_afk [11:59] jamespage, Which adds a ton of things that get done. Just near impossible to give any advice. I think it was doing the install via curtin and that I think is using a trusty kernel environment but I am not sure about anything there === cpaelzer_afk is now known as cpaelzer [12:20] smb, release image does the same thing - how do I start to debug this? === Lcawte|Away is now known as Lcawte [12:52] jamespage, first thing would be to figure out what things get mounted and unmounted over time... then the question is whether it is possible either to stop in between or even better if it is possible to recreate the problem with simple manual steps that follow that sequence [14:04] jamespage, nice. [14:04] smoser, you like? I was toying with adding another config option to allow a user to specify the mappings [14:05] its not clear what i would pass to --hypervisor-mapping? [14:05] oh. thats what you were playing with i suppose. [14:06] i like the config to enable that or not. that is good. [14:13] smoser, so something like --hypervisor-mappings="root.tar.xz=lxc,disk.img=qemu,disk1.img=qemu" [14:13] or that could be a pointer to a yaml or suchlike [14:13] mappings: [14:14] - lxc: [14:14] [ "root.tar.xz", "root.tar.gz" ] [14:14] and so on [14:28] jamespage, i think thats fine. you considered the other way? [14:28] yeah. i like your way [14:28] (i was thining the other way as 'lxc=root.tar.xz'...) [14:40] magicalChicken, o/ [14:40] smoser: o/ === CiPi is now known as cipi [15:33] jamespage, fyi, you are chair on server meeting this week [15:33] smoser, okay.... [15:33] better make and effort to show up them [15:33] n [15:34] Hello I have installed postfix on my ubuntu 14.04 LTS I have tried to send emails because I am currently setting up a Firefox Sync service. I just need to send emails. But when I try to send an email manually via connecting to port 25 via netcat I just get that the "recipient was rejected: " [15:34] someone an Idea how I can fix this? I use normally centOS and I didn't had this problem before... [15:36] it's a fresh postfix install [15:44] rbasak: ping, if you're around, could use someone to bounce an opinion off of [15:44] teward: o/ [15:44] (regarding nginx) [15:45] rbasak: so, working on the merge, lotsa changes. including the modules Debian ships with nginx-full, do we update nginx-core to have the non-third-party module set that nginx-full now has, or no? [15:45] (already had to alter the description of the package and build rules) [15:45] when we created -core we gave it the same set of modules as -full, minus third party [15:46] if we diverge i'll have to edit the package descrip. further [15:46] just need a second opinion :) [15:46] (i'm for making the change to match -full, OR to match what nginx.org upstream does) [15:46] teward: so there's a change in upstream or Debian that adds additional modules, but these are official upstream modules and not from third parties? [15:46] rbasak: correct [15:47] teward: are they coded from scratch by upstream or adopted from previous third party modules? [15:47] they're not the third party modules at all, far from it actually [15:47] urgh i hate my system one sec [15:47] graphics glitch [15:47] * teward beats xorg with a stick [15:47] I just want to confirm that it isn't that upstream have said "that's a handy third party module, let's adopt it!" [15:49] If the provenance of the new modules aren't any of the modules that the security team explicitly nacked, then I'd say stick with keeping the Ubuntu delta from Debian down. The definition of core was "Debian's -full less third party modules" and we can retain that definition. [15:49] That's easiest for you too, I think? [15:50] It's fine for new upstream code to end up in main without a security review by default. We rely on upstream diligence in that regard for all packages in main. [15:50] rbasak: indeed. "OPTIONAL HTTP MODULES" was any optional modules that shipped with the nginx source code but were not third-party [15:50] and we have several changes. [15:51] enable http substitution support, enable 'stream' module, enable gunzip module, enable Thread Pool functions, all from nginx's upstream tarball [15:51] that's the only changes [15:51] though we *could* pull right from nginx.org's main repository for their open source version and base nginx-core on *that* instead [15:51] since they made that open-source and available recently [15:52] but i think it's easier to just clone nginx-full minus third-party modules [15:52] and none of these appear to be NACK'd by the security team [15:53] teward: +1 === cpaelzer is now known as cpaelzer_afk [16:16] Ugh I hate this fucking distro [16:19] nat0: tried anything better? ;) === cpaelzer_afk is now known as cpaelzer [16:19] Yeah, almost anything else. [16:19] then why do you use it? [16:19] Clients are always right, even when they're idiots. [16:20] RoyK: maybe a good question would be: "come here to complain will make things better?" heheh [16:20] * RoyK is managing some 100+ rhel/centos machines and feel like he's in the stone age when doing so [16:20] But seriously, how is it I need 500+GB of local mirrors to preseed deployment onto a new machine? [16:20] icezimm: perhaps better, yes [16:20] hehehe [16:21] icezimm: but then, I rarely use ubuntu on servers - turned my head back to debian [16:21] Why does deploying trusty require a full wily and wily-update path? Why doesn't disabling upgrades and PGP sig integrity in the preseed file prevent anna from 404ing upon not finding a wily-updates path for a trusty install. [16:21] nat0: sounds like you've done something funny [16:21] I would agree, except it's been a week of combing over this server and finding no reason for it. === cpaelzer is now known as cpaelzer_afk [16:51] nat0: that doesn't sound right to me. [16:52] nat0: I don't think any of the things you mention are mandatory. [16:52] nat0: deploying one particular release certainly doesn't require any other. === cipi is now known as CiPi === devil is now known as Guest3575 === Piper-Off is now known as Monthrect === cpaelzer_afk is now known as cpaelzer === cpaelzer is now known as cpaelzer_afk [20:16] zul, can you sponsor an upload to xenial for python-os-win? https://code.launchpad.net/~corey.bryant/ubuntu/+source/python-os-win/+git/python-os-win/+ref/debian/mitaka [20:30] HI there [20:31] I need help reconfiguring a degraded RAID10 array [20:32] it seems like I have no swap anymore, the partition type is unknown and all my disks have 1MB of unused space [20:37] sounds normal [20:43] patdk-wk: Normal ? no swap ? cat /proc/mdstat showing only three devices and saying md0 (swap partition) is inactive ? Disk manager saying the new disk has unknown partition while it's MBR for the three others ? and the date partition on this disk is of type "oxfd" while it's "RAID auto Linux (bootable)" for the three other disks ? There are a lot of things which don't seem normal at all. I turn to this IRC as a last solution bec [20:43] ause Google is not halping [20:44] heping [20:44] helping ! [20:44] so you checked everything, except your raid [20:45] if it says the md0 (swap) is inactive (bad, not working, missing, ...) [20:45] you need to make it active [20:45] first yo uhave to see why it's inactive (missing disks, corrupted, ....) [20:45] your problem is no different than, someone took the drive out of my computer, and now my drive missing, why [20:46] so first step is to ask mdadm what is wrong [20:47] padtk-wk: ok it may be inactive because I removed one faulty disk and replaced it by a new one. I did rebuilt the main array (md1) but could not do the same for the swap space (md0) So I guess it's why it's inactive, but how do I rebuild a swap space ? [20:48] I dunno, you haven't stated why md0 is inactive yet [20:48] I cannot guess [20:48] inactive != has a failed disk [20:48] inactive == offline, or all disks failed that are needed [20:49] patdk-wk: How can mdadm could tell me what is wrong with md0 ? [20:49] mdadm --detail /dev/md0 [20:50] mdadm is stating the obvious: "md device /dev/md0 does not appear to be active" lol [20:50] hmm, different command [20:51] patdk-wk: which different command ? [20:55] patdk-wk: Perhaps I could add /dev/sdb1 (dedicated space for swap on new disk sdb) to the array, and it would be active then ? but how do I proceed ? [21:03] patdk-wk: anyway, thank you for paying attention [21:16] zul: hey, how can I test mitaka-1 ? is https://launchpad.net/~ubuntu-cloud-archive/+archive/ubuntu/mitaka-staging tested & stable ? [21:17] DannyButterman, maybe, mdadm --examine --scan [21:17] EmilienM, it's not available yet [21:17] coreycb: hey, ok good to know [21:17] coreycb: do you have any timeframe? that's for puppet ci [21:17] EmilienM, hopefully in the next week we'll have mitaka-1 in -proposed [21:17] coreycb: awesome! [21:17] coreycb: can you ping me? [21:17] EmilienM, sure [21:17] I'll work on the bump [21:18] thanks coreycb :) [21:21] ddellav_, I just pushed python-oslo.messaging 3.0.0-1ubuntu2, so once that is built you should be back in business with ceilometer [21:22] s/pushed/uploaded [21:32] patdk-wk: it displays five arrays, first being /dev/md/0, second metadata third and fourth containers and last /dev/md/1 [21:32] DannyButterman: hahahahaha [21:32] DannyButterman: thanks for the wonderful nickname :) [21:32] not much we can do, without seeing it [21:33] sarnold: congratulations, I've used it for years and you are the first to notice :D [21:36] patdk-wk:pastebin.com/QhzyTU6Z [21:36] hmm [21:37] DannyButterman: looks like there won't be a shortage of chunky monkey for the next month! [21:37] what if you do a, mdadm -A /dev/md/0 [21:37] sarnold ^^ [21:40] patdk-wk: no output for this command, no error either. It changes nothing regarding the mdadm examine scan output [21:40] it won't [21:40] but what does /proc/mdstat say? [21:40] no ouput should be, it worked [21:42] patdk-wk:pastebin.com/pW0cZsBu [21:43] odd [21:43] it marked all your disks as not active, but as spares [21:45] patdk-wk:what if I started over ? Removing the disk and again trying to rebuild the swap array and the data array ? [21:46] that would be my recommendation [21:46] I can only guess it was a raid0 [21:46] I would not bother doing raid at all [21:46] just mount 4 seperate swap volumes [21:49] patdk-wk: it was a raid10, and still is for the /dev/md1. let's say okI do the four swap volumes. But how do I make the "new" disk MBR ? And the data partition on it, bootable ? [21:50] heh? [21:50] what does that have to do with swap or raid? [21:50] you don't boot to raid (normally) [21:50] and you do not boot to swap [21:50] if your asking about mbr now, it's too late [21:51] you have to set that up long before you touch thedisk [21:52] patdk-wk: That's waht the Disk manegr tells me for the three other disks. I find it odd that the new disk has a different partition type and its data volume has no bootable flag like the three others on the other disks ('RAID auto Linux (bootable)') [21:53] bootable flag is useless [21:53] you are installing grub [21:53] grub could care less about *bootable* [21:54] and yes, mbr cannot boot to a raid volume, it won't work [21:54] unless the raid is raid1 ONLY [21:54] so to set it bootable, is pointless, unless the mbr bootloader unstatands it, and it doesn't [21:55] patdk-wk: the how to's instructed me to do so for the bootable flag. About the mbr, I would expect the raid mecanism to replicate it also [21:55] if the mbr is replaced with grub (like normal), it will work, and grub doesn't use the bootable flag [21:55] raid mecanism? [21:55] the bios doesn't support raid [21:55] and the mbr is too small to handle raid (500 bytes) [21:55] or less, 450bytes? [21:56] software RAID logic I you prefer [21:56] yes but you need software to do software raid [21:56] I just said and explained, nothing that boots, understands it [21:56] so it won't work [21:56] grub supports raid, grub doesn't use the bootable flag, so pointless [21:57] if you had a raid1 /boot, then it could make sense, but you would still use grub, so still no need [21:58] when you boot your computer with ubuntu, it is pretty simple [21:58] bios -> mbr -> grub -> uuid of /boot [21:58] grub must be installed in a raid1 fasion on all *bootable* disks [21:59] so a mirrored copy on all of them [21:59] patdk-wk: most useful 1 line of the month, kudos [21:59] what one? :) [22:00] the boot process on raid [22:00] pretty much %99 of raid questions need that before they speak [22:02] patdk-wk: Following the RAID configuration in ubuntu 14.04 server install, with the help of howto's on RAID configuration ended up on what I explain. I would assume the ubuntu server install would not do stupid and useless things :s [22:02] stupid? no, useless, yes [22:02] it's not stupid, as it is a *human hint* [22:02] but it is rather useless, from a boot point of view [22:02] the raid install works just fine [22:03] you just have to understand the rules of raid [22:03] and the documents are not created to be sliplined and efficient [22:03] they are designed to be generic and mainstream [22:04] they also expect a level of knowledge [22:15] patdk-wk: ok but the install process did not ask me how I would like grub to be managed in the raid context, I don't even know how to tell it do do otherwise ("grub must be installed in a raid1 fashion on all *bootable* disks") [22:17] that is cause it's not an option, and the installer does it for you [22:17] it installed grub on all bootable disks [22:17] being every raid disk that holds /boot [22:17] as far as marking the partition bootable, that is likely just code leftover from old/before grub [22:18] it's not needed, it doesn't hurt anything [22:18] you can set it bootable if you want, no problems with that, but doing so won't *fix* anything [22:18] more important is, make sure you get grub reinstalled onto that disk [22:19] patdk-wk: OK, but the MBR partition type ? I did not ask it to do so. I only specified I wanted the volumes to be "primary" not "logical" [22:20] and? [22:20] is there something wrong with the type? [22:20] just change it, no issue there [22:20] grub won't care about it [22:20] mdadm I think will care some [22:20] but I do not think it will even care [22:21] most things use uuid these days [22:21] and the mbr infomation is just for human identification === TheSilverSentine is now known as TSS [22:22] And the three other disks are MBR partition type. The new one is not. I can't change it, nor I can't add the bootloader flag (I know it doesn't matter but it my "monk" side lol) [22:23] you mean mbr vs gpt? [22:23] that depends on your bios [22:23] gpt is better, supports disks > 2tb [22:23] has more *partitions* [22:23] though i find 1 or 2, more than enough [22:26] patdk-wk: I don't know if it's related to "MBR vs GPT" , it's just what the disk manager says, and I don't like when severeal disks that are supposed to have the same configuration, do not have the same configuration. And this MBR partition type listed in disk manager is one of these differences [22:26] well, change it [22:27] and set bootable if yo uwant [22:27] it's very simple to do [22:27] I have no idea how you created the partitions on the new disk [22:27] but that all has to be done manually by you [22:29] patdk-wk: the partitions were created by "sfdisk -d /dev/sda | sfdisk /dev/sdb" something like that [22:30] well, that will make an exact copy [22:30] assuming they did not use GPT anywhere [22:30] if the disks where gpt, instead of mbr, it would not work correctly, and have kindof random results [22:30] hmm, none of my disks have the boot flag set at all [22:31] patdk-wk: as far as disk manager can tell, it's not an exact copy alas [22:32] run gdisk -l on them and see what it says [22:32] Partition table scan: [22:32] MBR: MBR only [22:32] BSD: not present [22:32] APM: not present [22:32] GPT: not present [22:32] if any say gpt, than fdisk/sfdisk should not be used [22:32] patdk-wk: about the boot flag on the data volumes, it's something from the howto I used. [22:33] it's not from the server RAID install process [22:35] coreycb, ok cool, let me know [22:37] patdk-wk:they're all MBR only [22:47] ddellav_, it's available in proposed now [22:53] Anyway, thank you so much for your time patdk-wk, bye === Lcawte is now known as Lcawte|Away [22:58] coreycb, fantastic, thanks === Monthrect is now known as Piper-Off === CiPi is now known as cipi