[01:59] where can I find information about running Ubuntu Server as a Dom0 (control domain) for the Xen hypervisor? Truth be told, I'm looking at various options for virtualization. I may be getting a work from home gig that will require me to have windows installed. I don't want to give Windows complete control of my machine but I want something more robust than Virtualbox. Any ideas? [01:59] **not looking to dual boot either. I absolutely detest dual-booting [01:59] using qemu via libvirt is probably the easier solution [02:00] chowder: https://help.ubuntu.com/lts/serverguide/libvirt.html [02:00] chowder: I suggest trying out the virt-manager and virt-viewer options on that page [02:01] I'm a programmer so I won't be doing anything extremely resource intensive like gaming but Windows takes up so many resources....I've tried Virtualbox and its nice but it slows down my main linux install while Windows is running. That's with 8GB of RAM and an i5-2400 at 3.1 GHZ. [02:02] sarnold, I'll check that out. thanks [02:02] that does sound pretty tight for serious use [02:03] there's just no two ways around it, you'll have two operating systems running and competing for resources at once [02:03] normally you can give the guest OS less RAM than you expect, because the host filesystem will cache the guest filesystem [02:03] so the guest won't need as much memoryaround for its own filesystem caching [02:04] but whatever the applications require to do their job you really better have available for them, for both host and guest [02:06] Maybe I should upgrade my RAM after all [02:06] or eventually buy a new machine with something like AMD's Ryzen [02:07] best case scenario is that my machine will run well and it can serve to cook my breakfast [02:13] heheh :) [02:23] I was honestly looking to virtualize several operating systems. I wanted to have an Ubuntu server host, Centos guest, windows guest, solaris guest, and netbsd guest. they're paying unix admins around 180k USD per year in my area so guess what I'm studying? [02:24] 180k for a solaris sysadmin. at that point why code? [02:24] wow, that's almost worth learning how to use those cranky old things again :) [02:31] I make that much without having to deal with Solaris. I'm good. [04:30] 8 GiB is definitely on the low side if you want to run VMs (especially if you want to run multiple) [07:09] Good morning === lukasa_ is now known as lukasa [12:01] hi, I'm running dpkg-reconfigure unattended-upgrades in ubuntu 16.04 and it's giving me a origins-pattern that points at debian, not ubuntu [12:01] is that correct? [12:01] specifically: [12:01] "origin=Debian,codename=${distro_codename},label=Debian-Security";_ [12:13] that doesn't seem right === 7ITABUPZE is now known as trobotham === jelly-home is now known as jelly === jancoow__ is now known as jancoow [15:22] ahasenack: it's the same in 17.10, so if you file a bug report: it's not only for that distro version... [16:04] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-bionic/bionic/amd64/o/open-iscsi/20171121_123741_f0d9a@/log.gz [16:05] anyone have a solution on that failure there ^ from http://autopkgtest.ubuntu.com/packages/o/open-iscsi/bionic/amd64 [16:05] the process runs [16:05] apt-get install -qy --only-upgrade libnss-myhostname=235-3ubuntu1 libnss-mymachines=235-3ubuntu1 ... [16:05] in an effort to only upgrade things (not install new ones) [16:06] smoser what's the testsuite test do? [16:06] but apparently that fails if there are other dependencies... [16:07] ddstreet: https://git.launchpad.net/~usd-import-team/ubuntu/+source/open-iscsi/tree/debian/tests/README-boot-test.md [16:07] it installs the packages that its being expected to test into a cloud image and then boots taht cloud image [16:08] but in this case that installation needed some dependencies, and that didn't work. [16:08] hmm [16:09] JanC: ok, thanks for checking [16:11] what i could do, buit had avoided and probably has similar issues... is go through the list of packages in the image (from dpkg-query --show) and filter it for the list of things of new things and then just 'apt-get install' those things. [16:11] smoser ah it looks like the -proposed systemd depends on a newer version of libgcrypt20 [16:12] Hello all! Office hours has officially started, feel free to ask whatever server related questions you have. [16:15] https://bugs.launchpad.net/curtin/+bug/1728742 [16:15] Launchpad bug 1728742 in curtin "curtin dname for bcache uses unstable devname instead of UUID" [Undecided,New] [16:15] dpb1: I think I have a cheap solution for this which doesn't require kernel work [16:15] https://bugs.launchpad.net/curtin/+bug/1728742/comments/4 <- this comment [16:17] * dpb1 looks [16:17] this is to make sure dname is set properly based on bcache superblock uuids. Currently the problem is that there is no superblock UUID in uevents coming from kernel on coldplug so udev rules provided by bcache-tools don't work and persistent device links provided by /dev/by-dname are wrong [16:18] you might want to highlight smoser and rharper on this Dmitrii-Sh [16:18] smoser: ^^ [16:18] true [16:21] smoser i think the problem is the test needs to pull the systemd deps from -proposed too, since it's getting systemd from -proposed [16:21] we normally rely on dname because as [16:21] ceph charm config <-> /dev/by-dname/ [16:21] or [16:21] Juju storage <- /dev/by-dname/ <- MAAS API [16:24] hi server guys, i have a q that may not be relevant to the server team but i'll ask anyway, i have a large update to the ubuntu-dev-tools package i'm trying to get merged but nobody seems to 'own' the package, most of the top contributors have not wanted to review my changes [16:25] so i'm planning on taking over (co-)maintainer of the package, anyone have any thoughts on that? [16:25] i'm going to talk to foundations people as it's probably more relevant to them [16:25] (talk to them also i mean) [16:26] ubuntu-dev-tools is maintained in Debian. [16:26] Which is a pain, IMHO, but anyway. [16:28] Also, IMHO, if patches are being sent to Debian and they aren't going in, we should just fork in Ubuntu (and/or offer to co-maintain in Debian if a DM or DD). [16:28] ahasenack: https://code.launchpad.net/~racb/usd-importer/+git/usd-importer/+merge/334033 and https://code.launchpad.net/~racb/usd-importer/+git/usd-importer/+merge/333888 please. The second is a pre-requisite of the first. [16:28] ok [16:29] ddstreet: are you by accident and D* in Debian already? [16:30] rbasak ubuntu-dev-tools is maintained in launchpad, owned by Ubuntu Development Team...unless i'm missing something? [16:32] e.g. https://packages.debian.org/source/sid/ubuntu-dev-tools lists Ubuntu Developers (and Benjamin Drung and Stefano Rivera) as maintainers, but hosted in LP owned by UDT team [16:35] cpaelzer i'm not debian devel [16:47] JanC: https://bugs.launchpad.net/ubuntu/+source/unattended-upgrades/+bug/1577215 somebody filed it already in 2016 [16:47] Launchpad bug 1577215 in unattended-upgrades (Ubuntu) "Origin pattern is unexpected on dpkg-reconfigure" [Undecided,Confirmed] [16:49] ddstreet: it is in debian as well [16:49] ddstreet: and a sync for us [16:50] cpaelzer where is the debian repo for it? seems they just use the lp.net/ubuntu-dev-tools git repo [16:50] ddstreet: you are at the right repo [16:50] ddstreet: but the upload has to go to debian [16:50] and synced from there [16:50] which will be an auto sync [16:50] so the target to upload to with dput&co is debian [16:51] right, i'm trying to change the upstream repo tho [16:51] which will go to debian of course [16:51] anyway i think we're all saying the same thing [16:51] ack [16:51] thnx! [17:03] ddstreet, cpaelzer: I think so. VCS maintained by ~ubuntu-dev, but uploads via Debian by Stefano and Benjamin only. [17:04] I'm a DM, so perhaps I should ask to be added. [17:08] rbasak: with https://code.launchpad.net/~racb/usd-importer/+git/usd-importer/+merge/334033, pytest-3 gitubuntu/*.py no longer hangs, but I get test errors that I don't get with tox [17:09] http://pastebin.ubuntu.com/26013402/ [17:11] ahasenack: you have to run the tests on artful I think. [17:11] that is artful [17:11] Otherwise your version of dpkg-parsechangelog won't do the right thing. [17:11] Hmm. [17:11] tox works [17:11] hm, I don't have dpkg-parsechangelog [17:12] rbasak: fixed [17:12] \o/ [17:12] we need a script to install all the dependencies [17:12] I might do that [17:13] ahasenack: here's what I'm doing: http://paste.ubuntu.com/26013419/ [17:13] nice [17:19] OK, thanks all for the questions. office hours is officially over. [17:19] * dpb1 shuts door, only to open it back up for regular business [17:19] rbasak: just checking, I'm getting this in master as well, not just your branch: http://pastebin.ubuntu.com/26013467/ [17:20] rbasak: I was trying a code path that would need a launchpad login, to see if your change affected that [17:20] rbasak so re: ubuntu-dev-tools, stefano hasn't uploaded to it in years (according to changelog) and there have been lots of uploads be people other than Benjamin, who last uploaded May this year [17:21] Perhaps they're (effectively) NMUing? [17:22] I'm not sure about that error. [17:22] could be [17:22] lemme see if i can talk to benjamin [17:22] Technically in Debian a team-maintained package can be uploaded without it being an NMU. [17:22] And I suppose the maintainer team is defined as "Ubuntu Developers". [17:22] So perhaps that's what they're doing: a team upload from bzr every tinme. [17:23] In that case, we should just land stuff into bzr. [17:23] And when needed do an upload into Debian. [17:23] they moved it to git earlier this year, but yeah [17:23] Which any DD can do, or I can do if someone sets my DM upload bit against the package in Debian. [17:27] rbasak i see you also have a MR open for pull-debian-source, you may want to review my MR as it consolidates all the pull-*-source and rewrites most of it [17:27] https://code.launchpad.net/~ddstreet/ubuntu-dev-tools/+git/ubuntu-dev-tools/+merge/322863 [17:27] cpaelzer: did you manage to use git-ubuntu submit without hitting this backtrace? http://pastebin.ubuntu.com/26013467/ maybe because you are logged in already? [17:32] ddstreet: you're right. the apt-get does not have the proposed entry there. [17:38] rbasak: still around? [17:43] ahasenack: yes but otp [17:52] rbasak: ok, I'll comment in the mp [18:05] ahasenack: back now but I'd like to EOD. Do you still need anything? [18:05] nope [18:05] OK. Thanks for reviewing! [18:05] np [19:03] Has anyone had experience installing Ubuntu Server on a Poweredge T30? It's telling me it can't boot off of my RAID array and I'm wondering if it's not liking the one I set up via Dell's utility or if I have to set one up in Ubuntu manually. [19:04] I've got 4 1TB disks set up in RAID5 [19:05] trekkie1701c: at least -someone- was able to get the thing booted https://certification.ubuntu.com/hardware/201702-25369/ [19:06] Yeah, that's why this is frustrating. [19:19] trekkie1701c: booting the installer, or booting the installed OS? [19:19] Installed OS [19:20] I can get the installer to boot off of a USB thumbdrive, but when I try to boot the installed system it tells me no boot device found. [19:21] trekkie1701c: right, because of the RAID-5 array? how does the T30 firmware select that device for booting? Is it UEFI or BIOS boot ? [19:22] trekkie1701c: the boot-method of the installer controls which boot-method it is installed for. E.g. installer boots in BIOS/legacy mode, it installs that mode. [19:22] BIOS mode, and the array shows up as a single volume that's selected for boot. [19:23] Overriding the boot settings and just going for that volume brings up an error about failing to boot off of the selected device. [19:49] trekkie1701c: at a guess I'd think GRUB didn't install to it correctly. Can Linux see the underlying disks as well as the logical RAID-5 array as well? if so, that can cause problems [20:20] trekkie1701c: I've had that kind of problem with a supermicro machine trying to boot off of nvme... booting the installer from pxe was fine, but when I tried to boot the system it would not [20:20] in my case it was a bios issue, ie the bios version I had could not boot from nvme [20:20] Hmm. [20:21] I'm wondering if in bios mode it has sufficient drivers to boot from that raid device or if for example you'd need UEFI with additional drivers [20:21] so in your case I may try to install using UEFI instead of legacy mode [20:21] Alright. [20:21] but it's just my WAG [20:22] Better than what I had to go on. [20:27] Hmm, UEFI won't even let me set the disk in the boot order. [20:31] trekkie1701c: how do you mean? what disks? usb one to install? [20:32] The hard disk. Shows up in SATA but not in the boot order. [20:32] I'm going to try installing anyways with UEFI and seeing if maybe it picks up once GRUB is reinstalled [20:32] I would not necessarily worry about that right now and just try to install, it may not do that because it doesn't see it as installed or something, even tho a bios shouldn't care [20:32] If not I guess I'll just disable the RAID array with Dell's software and use the Ubuntu installer to set it up [20:32] yeah [20:33] is this supposed to be a good hw raid with battery backups and stuff? [20:33] because otherwise ime it's not worth it [20:33] Software RAID provided by Dell's firmware, but I want to experiment with a RAID setup. [20:34] Technically speaking this server isn't worth it. I don't need it, but I want to play around on it and doing virtualbox off of a laptop is getting annoying. [20:34] if it's a sw raid then do it with linux's sw raid... mdadm [20:34] (or zfs :D ) [20:34] Alright. [22:39] Sweet. EFI worked. [22:41] woot