[00:05] Does anyone have experience installing on Lenovo's machines with Software RAID? The install keeps failing during grub-install. [00:06] TDJACR: how large is your RAID in size? [00:07] xnox: It's 6TB, software raid. I wish I could disable it (and use mdadm instead), but it won't boot unless it is used. [00:08] TDJACR: MBR and bios grub has limit of 2TB [00:08] TDJACR: make sure your /boot is on a separate partition, which ends << first 2TB [00:08] TDJACR: or make sure you are using UEFI boot. [00:09] xnox: I tried with EFI too, but the system wouldn't add the EFI boot option... [00:09] On both CentOS and Ubuntu [00:09] =/ [00:09] TDJACR: right, i've had that with a few raids. [00:09] TDJACR: so yeah, have you been using separate /boot ? [00:11] xnox: Nope, but I will try that and legacy boot again [00:12] TDJACR: make sure you have "biosgrub" partition first, then "/boot" (e.g. 500 MB) and then whatever else, e.g. "/" for all remaining portion of the raid array. [00:14] xnox: How big should the biosgrub partition be, and how should it be specified in the install? [00:15] TDJACR: in the manual partitioning, just create partitions as needed. biosgrub is just for the overflow of things that don't fit in the MBR, ie. kBs, make it 10MB [00:15] xnox: So I'm leaving it unallocated? [00:16] TDJACR: no. [00:16] TDJACR: in the partitioner one marks it as of special type for biosgrub..... although let me check that. Maybe it's done behind the scenes with no UI. [00:17] Okay, that makes sense. [00:17] it's like "swap" but different =) [00:18] Alright, I'll give that a try [00:21] Trying now, thanks. [00:28] TDJACR: hm, biosgrub type of partition is not offered to me in the installer, that's weird. But this is mbr without gpt, maybe with 6TB it will be using gpt and then offer it. If not do leave out space before first partition then. [00:29] TDJACR: good night and good luck [00:30] Thanks xnox [00:43] I need someone to run an nmap command for me, the command would be this : nmap -F -T5 -Pn -sS -v x.x.x.196-222. And pm me the results. I can't do it because all my computers are on 'trusted' networks. [00:43] oops [00:43] nmap -F -Pn -sS -v x.x.x.196-222 [00:44] I'd PM you the IP. [00:44] and PM me the results :) [00:45] i'll do it. [00:45] there are websties that do this :) [00:47] jrwren_, I PMed my IP. [00:50] I tried a few, but they don't seem to work === arrrghhh is now known as arrrghhhAWAY === Guest85592 is now known as Locke2002 === arrrghhhAWAY is now known as arrrghhh === arrrghhh is now known as arrrghhhAWAY === xtriz is now known as fifo === CripperZ- is now known as cripperz === cripperz is now known as N0DE` === N0DE` is now known as CripperZ === kaitanya_ is now known as chaitanya === a1berto_ is now known as a1brto [08:31] <`ajven> Hi, got problem with my eth0 card, i dont know why its not running on system start. Any idea what can be wrong ? [08:50] <`ajven> how to make sh script autostart on systemboot ? [09:02] `ajven: Your /etc/network/interfaces file is broken. [09:14] hello. I am doing a dhcp server on my ubuntu computer using dnsmasq. For this I have disabled dnsmasq in networkmanager and everything works as expected. The only problem is that i cannot make dnsmasq start at bootup. How can I debug this problem? [09:15] man updaterc.d [09:15] Sorry, update-rc.d even. [09:16] updaterc.d dnsmasq defaults [09:18] the links already exist but dnsmasq reports an error that the interfaces do not yet exist [09:22] Hmm, it sounds like the job to start dnsmasq is being done before the job to start network interfaces [09:23] i supposed so, but i dont have the knowledge to debug this problem [09:24] I assumed that the dnsmasq package does the update-rc.d stuff so that this issue is taken into account [09:24] and I recall that in my previous works dnsmasq worked out of the box [09:24] but I dont know why it does not now [09:24] Ok, can you try using `update-rc.d dnsmasq disable` then `update-rc.d dnsmasq enable 5` to start dnsmasq on runlevel 5 only. [09:24] there are no logfiles in /var/log/upstart/ for dnsmasq [09:25] Which version of Ubuntu are you using? [09:25] 14.04 [09:25] and its desktop - please dont bash me -> the desktop people sent me here despite that fact [09:26] I'll try and reboot now with that enable 5 thing [09:26] brb [09:27] Make sure to disable first, please. [09:28] didn't work [09:29] Ok, did you disable first? [09:29] yes [09:29] i issued those two commands and then restarted [09:29] Where do you see the error message? Can you paste the error message in full? [09:29] !paste [09:29] For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [09:32] ill try to get the error message again [09:33] the problem is taht i don't remeber where I found it -> I tried fixing the problem some time ago and just came back to it [09:37] ok. I rebooted and checked syslog [09:38] these two lines are relevant: http://paste.ubuntu.com/7893428/ [09:39] I have got three network interfaces - eth0 (pci network card) eth1 (motherboard card) and wlan0 (usb adapter) [09:39] i command dnsmasq to work on the eth1 interface [09:43] heres the log from the whole boot. The error is at 1151 [09:43] http://paste.ubuntu.com/7893466/ [09:44] it kind of starts in parallel with networkmanager and is always a little bit earlier? Below theres some networkmanager eth1 related stuff [09:56] smoser: please could you triage 1341710 for cloud-init? Looks valid to me. [10:04] hi [10:07] I upgraded my server to 14.04. After the upgrade finished the server asked to restart but never came up again. How do I get it up again? (I have no physical access and each hour of a virtual keyboard at my hoster costs money) [10:07] !server | hasdf better ask in other channel perhaps [10:07] hasdf better ask in other channel perhaps: Ubuntu Server Edition is a release of Ubuntu designed especially for server environments, including a server specific !kernel and no !GUI. The install CD contains many server applications. Current !LTS version is !Trusty (Trusty Tahr 14.04) - More info: http://www.ubuntu.com/products/whatisubuntu/serveredition - Guide: https://help.ubuntu.com/12.04/serverguide/C/ - Support in #ubuntu-server [10:09] cfhowlett, I am in #ubuntu-server? Or does my irc-client lie to me?= [10:09] hasdf doh! never mind. too many tabs. [10:11] hasdf: that depends on why it's broken. I'm not sure anybody on IRC can figure that out without knowing what it says on the screen - sorry! [10:11] my question is more: what would be the next steps? [10:12] to find out what is wrong [10:12] because I don't want to pay my hoster for a virtual keyboard and then not know what to do [10:13] hasdf: it might actually minimise your costs to pay somebody for support to help you with what to do, if that leads to a faster resolution. [10:13] hasdf: (since you hardly get an SLA on IRC) [10:14] Or switch to a hosting company that doesn't charge for remote access to the console. [10:14] ok, thanks [10:15] If it doesn't boot up it's most likely a problem with grub isn't it? [10:15] or sshd doesn't start (but why shouldn't it) [10:18] Could be, yes. [10:18] How did it get installed in the first place? [10:19] IIRC, there is a class of grub bugs that occurs on upgrade, if it wasn't installed correctly originally. [11:04] I am running VBoxHeadless --startvm "Ubuntu server 12.04LTS #1" --vrdp=off in an ubuntu server to start my vm but it reserves a terminal window. IS there ay way to let it run in the background ? [11:10] morning === baffle_ is now known as baffle [12:12] hi does ubuntu 14.01 update has new kernel version? [12:21] zartoosh: there is no 14.01, but 14.04.1 only. And most likely it ships with a new version, not a new release. [12:24] zartoosh: bugfixes only. You'll see a new kernel with 14.04.2, which will have an HWE kernel backported from Utopic. [12:30] thanks [13:01] hi I would like to know if there is an application that can list the name of the packages installed on my system? thx [13:02] zartoosh: dpkg -l [13:06] bekks, thx [13:24] just wondering if there has been any progress on the release for a fix of the 3.13 Ubuntu kernel issue .. ? [13:24] What issue? [13:25] the issue in 14.04 which makes me have to disable KSM on my KVM hosts [13:26] Ah. Check in the bug, maybe? [13:26] Else you can ask in the weekly server team meeting. [13:26] I am not sure which is "the" bug ... [13:27] there is this one: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1341195 [13:27] Launchpad bug 1341195 in linux "KVM guests getting slow by time" [High,Incomplete] [13:27] but maybe this is not "the" one!? [13:29] zetheroo: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1346917 maybe? [13:30] Launchpad bug 1346917 in linux "Using KSM on NUMA capable machines can cause KVM guest performance and stability issues" [High,Fix committed] [13:34] thanks [14:07] i have ubuntu server 14.04.. I cloned a project into the '/var/www/` directory. But, when I goto the browser, I cannot access the 'index.php' page [14:15] jeffreylevesque: the default document root has changed to /var/www/html. See https://wiki.ubuntu.com/TrustyTahr/ReleaseNotes#Apache_2.4 [14:16] Looks like there'a broken link there [14:16] apache2.NEWS as shipped in /usr/share/doc/apache2/ anyway [14:42] a lot of dovecot's sub-executables do not have apparmor profiles, afaik these are available and IMHO it does not really make sense to ship haf-baked/partial profiles for a programme, even if some customization is usually needed anyway [14:44] blaaa: dovecot profiles are in the apparmor-profiles package. those aren't shipped by default and are considered example profiles [14:46] jdstrand: OK, clear, I'll remove that package then and roll my own profiles. Mabye the description of apparmor-profiles should be adapted a bit, to make it clearer not to expect too much [14:47] which is fine by the way., creating own profiles should not be too much of an issue, at least when the aa-tools are working, which is a bit of a poblem on 14.04 right now [14:49] yes, that is known. fixes will be coming, but profiling by hand will always work [14:49] !info aide [14:49] aide (source: aide): Advanced Intrusion Detection Environment - static binary. In component main, is optional. Version 0.16~a2.git20130520-2 (trusty), package size 499 kB, installed size 1356 kB [14:50] true, just a bit of a chore sometimes. But fortunately only need do it onxce === CripperZ is now known as cripperz [15:24] hey guys, what's the difference between using 'sudo stop mysql' and '/etc/init.d/mysql stop' ? [15:24] I noticed that when using the latter it stopped the mysql socket from listening but kept all mysql processes running [15:25] Good afternoon. [15:25] bitfury: upstart vs sysvinit [15:25] I had to use 'sudo stop mysql' [15:28] pmatulis: got it, thank you. using init.d/mysql stop should've yielded the same results though right? [15:28] bitfury: not necessarily. what Ubuntu release are you using? [15:30] bitfury: https://bugs.launchpad.net/ubuntu/+source/mysql-5.5/+bug/1273462 [15:30] Launchpad bug 1273462 in lsb "Users can mistakenly run init.d scripts and cause problems if an equivalent upstart job already exists" [High,Confirmed] [15:30] bitfury: in other words, it's safest to *never* call /etc/init.d/X [15:30] Always use the service wrapper. [15:30] "sudo service mysql stop" [15:30] Unless you're absolutely sure which init system is in use. [15:31] And the correct thing to do for that particular release. [15:32] got it, thank you rbasak [15:50] stgraber, hallyn fwiw, i did: lxc-userns-setup at https://gist.github.com/smoser/49444542158f2e5f88f1 [15:55] if I use ngrep to monitor a network card, should I see packets destined for other IPs on the same switch ? [15:55] probably not [15:55] likely the switch would not send you traffic not for your mac. [15:56] a "hub" would do that. but i dont know that you can even buy one of those. [15:56] toyotapie: Not as long as you dont use promiscious mode. [15:56] I'm working on configuring a netboot installer and am trying to understand the process better. I've got the dchp server configured with next-server, and have tftpd-hpa set up. My system is getting as far as looking for the installation media. Is 'mini.iso' the installer? [15:56] smoser: you don't want to shut down the contaienr when done? [15:57] hallyn, the first thing is 'lcx-chroot'. i was pointing at second. (lxc-userns-setup) [15:57] regarding shutting down the container, it does that in cleanup (on trap EXIT) [15:58] ah [15:58] the lxc-userns-setup just basically adds the .config/lxc/default.conf and adds user to /etc/lxc/lxc-usernet if not htere. [15:59] toyotapie: you need to use a managed switch and put that port into monitor mode. also, "other ips on the same switch" is nonsensical. switches operate at layer2. [15:59] smoser: is lxc-userns-setup new? [15:59] ok, will i don't feel like working on th eother two windows i have open, so lemme try a patch to give more helpful error on no-userns [16:00] mgw, new in that i just put it together. not new in that it is nothing not described at https://www.stgraber.org/2014/01/17/lxc-1-0-unprivileged-containers/ [16:00] smoser: you could simply copy /etc/lxc/default.conf to get the system-defined default networking... [16:00] the point of it is that i just want to run one program and be able to then run 'lxc-create -t ....' === arrrghhhAWAY is now known as arrrghhh [16:00] smoser: thanks, yeah... I've used stgraber's directions already [16:00] I was just curious as I did not see it on my system :-) [16:01] hallyn, yeah. getting the bridge from there woudl be good. [16:01] hallyn: or use "lxc.include = /etc/lxc/default.conf" [16:01] i was just goign to ask that. [16:01] if i could #include [16:01] cool [16:04] toyotapie: another option is arp spoofing ;] [16:14] I'm trying to get a minimal netinstall working... what are the files from http://archive.ubuntu.com/ubuntu/dists/trusty/main/installer-amd64/current/images/netboot/ that I absolutely need for an unattended install? I have a custom pxelinux.cfg that points to a preseed on an http server. [16:15] this is my cfg https://gist.github.com/mgwilliams/e56a131bcdffce384a7b, but I'm not sure whether the ks= is needed. On my old setup that ks file points to a full install media location [16:16] but I want to do the installs via apt (I have a caching proxy) === psivaa is now known as psivaa-bbl [16:19] mgw: For a kickstart file, you need the ks= parameter. For a preseed file you need a preseed/url or preseed/file parameter. [16:20] bekks -- so do i not need ks= at all if I have preseed? [16:20] mgw: And their is no such thing as a "kickstart full install location". The kickstart/preseed file determines what is going to be installed and what not. [16:21] ok, so I'm not sure what I used to have :-) it works, but I think it's rather convoluted [16:21] Just configure your kickstart file to install what you want. [16:21] and I can avoid http altogether by using preseed/file, correct? [16:21] and serve the preseed over tftp [16:24] I'm getting the bekks: something like this? https://gist.github.com/mgwilliams/e56a131bcdffce384a7b [16:24] with that file in by tftpboot dir === matsubara is now known as matsubara-lunch [16:32] hi folks! [16:32] i have a new issue [16:32] o/ [16:33] !ask | paco1 [16:33] paco1: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [16:34] when i boot my server DELL poweredge 720, with 4 broadcom netxtrem BCM5720 (driver tg3), ubuntu 14.04 lts don't up the eth0 and eth1 interfaces [16:42] bekks: I'm still getting stuck on "searching drives for an installer ISO image" [16:42] Does this mean it's not getting my prseed? [16:44] paco1: pastebin ifconfig -a [16:48] mgw: It doesnt find the ISO to be booted. [16:49] bekks: i'm trying to figure out where it's looking for that iso [16:49] i dropped into busybox and it looks like there's no networking set up [16:54] Nivex, so no real response to the launchpad bug yet... is there nothing else we can do, just wait? [16:55] (I hate waiting too :P) === jrwren_ is now known as jrwren [16:59] paco1: ? [17:00] paco1, heh? you actually have eth0 and eth1 interfaces? [17:00] patdk-wk: why do you think I asked him for the ifconfig -a output? ;) [17:00] didn't see that [17:01] was way too many lines away :) [17:01] hehe [17:02] arrrghhh: I suppose we could ask at the Ubuntu Engineering Live! hangout tomorrow at 1400x [17:02] *1400z [17:06] Nivex, haha yes let's troll them :P [17:08] I'm not a fan of trolling. I'm interested to know what challenges remain in making that upgrade path possible. [17:10] Agreed, I wasn't serious about trolling them... [17:11] although I always do feel a bit trollish asking for ETA's on anything [17:11] they just usually stick to the release schedule pretty closely... [17:12] where does the pxe linux look for the installer iso? [17:14] arrrghhh: it made it on to askubuntu too http://askubuntu.com/questions/502886/ubuntu-12-04-4-lts-to-14-04-1-lts-upgrade-no-new-release-found [17:15] mgw, I think you can define it in the dhcpd conf... I had a dream of setting up a pxe server but never got around to finishing. [17:15] here's an article I found which was going to be the setup I wanted - https://help.ubuntu.com/community/PXEInstallMultiDistro === Ursinha-afk is now known as Ursinha [17:17] Nivex, lol the responses... I gotta quell the thought that it could be "random" [17:17] er, can't comment. oh well [17:19] I think my immediate problem is it's not grabbing the preseed b/c networking is not up, and I can't figure out why. [17:19] Ya I think at some point I will resume trying to do a pxe server... but at the moment my need for that is not so great. [17:20] ugh, maybe I don't want to upgrade: http://sebastian.marsching.com/blog/archives/134-Linux-KVM-Problems-with-Ubuntu-14.04-LTS-Kernel-3.13.0-30.html [17:23] normally, pxe doesn't look for the iso anywhere [17:24] arrrghhh, this issue: [17:24] do-release-upgrade -c [17:24] Checking for a new Ubuntu release [17:24] No new release found [17:25] patdk-wk, yes [17:25] what's -c? [17:25] http://manpages.ubuntu.com/manpages/precise/man8/do-release-upgrade.8.html [17:26] check for upgrade [17:26] ah manpage isn't updated I spose [17:26] -c, --check-dist-upgrade-only [17:26] Check only if a new distribution release is available [17:26] and report the result via the exit code [17:26] jeez that pasted like crap sorry [17:27] arrrghhh: interesting - can't find that flag in the 14.04.1 manpage [17:27] RoyK, do-release-upgrade -h has it. I pasted the result, although badly. [17:27] dunno, it is in the 12.04 manpage === Ursinha is now known as Ursinha-afk [17:27] file a bug. man page is wrong. [17:27] patdk-wk, it's not, I just linked it. [17:27] but no biggie lol === Ursinha-afk is now known as Ursinha [17:30] well, there is the issue [17:30] http://changelogs.ubuntu.com/meta-release-lts [17:30] it's not listed [17:30] yes, we know. https://bugs.launchpad.net/ubuntu/+source/update-notifier/+bug/1344762 [17:30] Launchpad bug 1344762 in update-notifier "update-notifier tells me to upgrade from 12.04.4 LTS to 14.04 LTS (because of HWE), but that release is not found" [Undecided,Confirmed] [17:30] Noone is saying why that hasn't been updated yet. [17:30] ittakes so long to read bug reports :) [17:30] patdk-wk, that's what we're wondering - why hasn't that been updated [17:30] only that it will happen "soon" [17:31] odd, I had assumed it would happen when 14.04.1 was released, like the other lts versions [17:38] patdk-wk, indeed... that's why we're thinking some bug is preventing it from happening - but we're not sure what. [17:39] well, there was a delayed lbuntu release [17:39] butthat is long past [17:39] most all of my systems I upgraded months ago though [17:40] I've been waiting for the .1 release on my production rig... === arrrghhh is now known as arrrghhhAWAY [17:47] Here is my interfaces file for my kvm server : http://pastebin.com/bDtW415Y [17:47] br0 is used by kvm to link vms to the network card. eth0 is my primary link, eth1 is my backup link. [17:47] Will this configuration allow traffic to pass between eth0 and eth1? I think it shouldn't allow this, but I am not sure. [17:48] gnuoy, matsubara-lunch: sorry, for some reason I said that gnuoy will be chair for the next meeting, but actually it'll be matsubara. [17:48] I think I just forgot after getting confused. Apologies. No point trying to add corrections now I think. === arrrghhhAWAY is now known as arrrghhh [17:58] has anyone here ever used a bond interface with a bridge interface ? [18:01] toyotapie: it shouldn't bridge eth0 and eth1 as that is not how its configured. [18:01] Ok. Thanks :). There is a loop on my network, and I wanted to make sure it wasn't the source. I was reasonably certain it wasn't, but a second opinion is always nice [18:01] thanks === matsubara-lunch is now known as matsubara === cps0_ is now known as cps0 [18:27] Any idea how to solve this upgrade from ubuntu 12.04.4 LTS to 14.04.1 LTS [18:27] $ do-release-upgrade [18:27] Checking for a new Ubuntu release [18:28] I do not what to have "$ do-release-upgrade -d" [18:28] Shutterstrom: That has been a topic of discussion in here for a few days. Noone knows when they are going to allow the updates other than "soon" [18:29] Shutterstrom: I believe there was a kitten-killer upgrade bug that was annoying enough that the upgrade isn't yet being offered.. [18:29] Nivex: Alright, so the upgrade is delayed? [18:29] I read something about Grub issues. [18:30] we have no idea what the hold up is. just that it hasn't happened [18:30] sarnold: Thanks! I just wondered if I was doing something wrong :) [18:30] Nivex: Enough info for me. Thanks! :) [18:31] Thanks for the info and have a great day! [18:38] heh, I posted the grub bug many months ago === Ursinha-afk is now known as Ursinha === psivaa-bbl is now known as psivaa [18:58] at what point in the boot process does the netinstall try to access the preseed file? And at what point is networking supposed to be up? [18:59] stgraber: ok so for the cgmanager/cgproxy systemd jobs, I'll just (a) assume systemd only runs on newer kernels, so never cgmanager+cgproxy; then (b) use ConditionVirtualization = lxc | libvirt-lxc | systemd-nspawn in cgproxy (and inverse for cgmanager) [18:59] mgw: Networking must be up and running. [19:00] bekks: of course... but at what point in the boot process is it supposed to bring up networking? [19:00] The installer is flaking on me with no interfaces configured [19:00] I've scoured /var/log/syslog from busybox ash prompt [19:01] mgw: So configure the interfaces - the exact point in time when the installer requires that is irrelevant. :) [19:02] bekks: the whole point is it's supposed to configure it unattended [19:02] this is for an unattended network installation (PXE) [19:02] mgw: you cant use preseed to configure the network when you need the network to actually boot and get the preseed config. [19:03] hello, i have a problem with this simple traffic shaping command: tc qdisc add dev eth0 root tbf rate 2048kbit latency 20ms burst 1540 [19:03] I've tried stuff like interface=auto in the kernel options, but that does not help [19:03] mgw: "interface=auto" is a wrong syntax. [19:03] bekks: that's obvious... I'm trying to figure out what is going wrong though that it can't get the preseed [19:04] bekks: there seems to be a lot of contradictory information online about how those kernel params are supposed to be formulated [19:04] mgw: which "kernel parameters"? [19:04] The default route from my cisco router goes through an Ubuntu-Server. On the Ubuntu-Server on eth0 i added that script, but i wonder why my Cisco only has 2Mbit of Upload and still 4mbit Download [19:04] can someone explain? [19:04] preseed/url, interfaces, auto, etc [19:04] that go on the append line of the pxelinux cfg [19:05] mgw: interfaces and auto arent kernel parameters, preseed/url=http://... is [19:05] I mean the default route goes through an gre tunnel which is my ubuntu server on the other end [19:06] bekks: it is my understanding that interfaces is an alias for netcfg/choose_interface [19:06] is that wrong? [19:07] mgw: you are mixing up preseed configuration paramaters and kernel parameters. [19:07] mgw: Using PXE, you cannot configure the network using preseed later on. [19:08] bekks: I'm basing my command line off this assumption: " If a preconfiguration file cannot be used to preseed some steps, the install can still be fully automated, since you can pass preseed values on the command line when booting the installer. " [19:09] mgw: thats a wrong assumption :) [19:09] ok, so ubuntu is different than debian in this regard? [19:09] mgw: That assumption is wrong on debian as well ;) [19:09] ok, do you suggest somewhere other than official docs to get info? [19:10] Nope, thats just my experience. [19:11] bekks: what boot params do you use to get network to auto configure? [19:11] I am using a plain dhcp server and preseed/url=http://... - nothing else. [19:12] And for the sake of convience, since preseed is pretty painful, I am using kickstart. [19:14] Has anyone here successfully used the boot parameters to preseed, as described here: https://help.ubuntu.com/14.04/installation-guide/amd64/apbs02.html#preseed-bootparms [19:14] What would happen if someone did? [19:14] bekks: I'm looking for help, so maybe such a person could help me? [19:15] mgw: then it would be more targetting to show your configuration and describe the error as detailed as possible. [19:15] bekks: I've done that already, remember? [19:15] mgw: what error are you seeing? [19:16] Did you pastebin both the pxe config and the pressed file? Along with the full, exact error, not what you think the error is? :) [19:16] *preseed [19:17] pmatulis: the installer is booting successfully over tftp (using various permutations of this: https://gist.github.com/mgwilliams/e56a131bcdffce384a7b) [19:17] but rather than pulling the preseed, it errors out trying to find install media on local drives [19:17] when I drop into a shell and run 'ip addr', there is no interface configured [19:23] pmatulis: sorry, that particular permutation is definitely bad [19:23] I'm now using preseed/url=http://x.x.x.x/server-14.04.presed [19:24] but it's never even trying to hit that url [19:24] mgw: To make that clear - you are using that parameter in the kernel command line of your PXE booted ubuntu? [19:25] yes, just as in the gist i posted [19:26] but I"m using preseed/url rather than preseed/file -- I"ll update the gist, on second [19:26] preseed/file uses a local file on the install media. preseed/url uses an URL. [19:26] bekks: right [19:27] mgw: that must be why it's not working [19:28] pmatulis: this is what I"m currently using: https://gist.github.com/mgwilliams/da96bc073452db7b7f4d [19:28] it's still looking for local install media rather than pulling the preseed [19:28] And that file is reachable with a browser? [19:28] yes [19:29] but I can't ping from the shell on busybox, of course [19:29] Do you have a working DHCP in that setup? [19:29] since ip addr shows no ips [19:29] bekks: yes, that's how it gets the files from tftp to start with [19:30] let me clarify: the machine that's running the installer is successfully getting an IP and PXE booting into the installer [19:30] I can watch the tftp traffic onthe tfpt server [19:30] I can see my kernel command line in syslog on the system that's running the installer [19:30] Which network interface make and model do you use? [19:30] but the system is not getting its networking configured or looking for the preseed [19:34] The box is an HP ProLiant with Broadcom networking.... [19:35] mgw: i'm having difficulty parsing this: "I can see my kernel command line in syslog...". you are viewing your logs on the system being installed and you see what? [19:36] pmatuils: when I drop to a shell in the installer, I can cat /var/log/syslog [19:36] yes, and? [19:36] mgw: which make and model _exactly_? :) [19:36] pmatulis: there I see a syslog entry including the kernel params [19:36] one sec, i'll get a screen shot [19:37] mgw: ah [19:37] mgw: where did you get your kernel from? [19:37] archive.ubuntu.com [19:37] i'll grab the exact url in a sec [19:37] mgw: specifically, thanks [19:38] pmautils: http://archive.ubuntu.com/ubuntu/dists/trusty/main/installer-amd64/current/images/netboot/ubuntu-installer/amd64/ [19:39] mgw: Did you check wether that kernel supports your networking mae and model? [19:40] bekks: are there known issues with 14.04 install media and fairly modern HP hardware? [19:40] yeah, you might want to test on a virtual machine [19:41] mgw: Would you please start answering my questions? :) [19:41] a lot faster to boot as well during the maddening preseed phase [19:42] 5 seconds vs 2 minutes prolly [19:43] Broadcom network cards frequently require loading extra firmware or other fiddling that can complicate things [19:43] mgw: there are known issues with the *linux kernel* with any fairly modern hardware ;) [19:44] If you want NICs that "just work" with linux, your best bet is 2+ year old intel chips [19:45] Good old Intel MT Server :) [19:45] PT :) [19:45] so many flavors [19:46] pmatulis: FWIW, I previously used the precise netinstall just fine on here [19:46] i350 should be great :) [19:46] so it would be a regression [19:46] mgw: Which network interface make and model do you use _exactly_? [19:46] I wouldn't use mt's though [19:46] bekks: the model is BCM5719 (rev 01) [19:46] hallyn: sounds reasonable [19:46] make is Broadcom [19:46] bnxe sounds like [19:46] mgw: are you saying your current preseed config works on 12.04 with exactly the same h/w? [19:48] pmatulis: yes, the same preseed [19:48] but the preseed is clearly not an issue here (yet) -- the preseed is never accessed from the http server [19:48] mgw: with the same h/w?? [19:48] pmatulis: yes, the exact same system [19:49] bloody hell why did you say so an hour ago? [19:49] it currently has precise installed, and it was done from precise's netboot installer [19:50] pmatulis: only so much data can be provided at once, sorry [19:50] i suggest re-confirming. exact same h/w with exact same preseed. minus the different pointers to precise bits instead of trusty ones [19:50] I mean, I *could* compile a complete history, including photographs of the hardware, before mentioning the issue ;-) [19:51] I doubt that networking hardware will work out of the box. [19:51] but it would with Precise you think? [19:52] bekks: ^ [19:52] It wouldnt. [19:52] hmm, well, it did [19:52] i set up eight of them on precise [19:52] Why dont you use the ubuntu server image? [19:52] with no special drivers [19:53] bekks: remotely? [19:53] Sure. Using PXE. [19:53] Or attach the ISO to the ILOM and boot it. [19:53] Even easier than PXE. [19:54] if you know how, sure [19:55] is the kernel on ubuntu-server iso different than the one on archive.ubuntu.com? [19:56] No, but the kernel on the netinstall thingy is much smaller than on the isos, for the sake of its size. And quite a lot of module are stripped off. [19:57] Looks like that one tuns the tg3 driver [19:58] bekks: thanks, that's very useful information [20:00] i had trouble with the Tigon stuff in the past === mikal_ is now known as mikal [20:04] does ubuntu.com or some mirror have an unpacked version of the server iso, where I can grab just the installer files? [20:06] you want PXE files or a mini ISO? [20:06]