=== devil is now known as Guest8411 [03:52] hi, I'm trying to preseed ubuntu on a raid1 and the preseed worked except that / and swap got flipped, so I got a huge swap and very small root partition [03:52] anybody seen that before? this is the relevant preseed snippet: https://paste.ubuntu.com/23546912/ [03:53] whcih seems about correct to me, shuold give me a swap no bigger than 2GB and a / of everything else (-1 max size) [03:53] a similar ex[ert recipe works just fine with non raid setup === Guest8411 is now known as devil_ === amoralej|off is now known as amoralej [09:31] HELLO CAN ANYONE HELP ME THROUGH ON ADDING A NEW NETWORK INTERFACE TO A UBUNTU SERVER 16.04 WITH ens33 === InfoTest1 is now known as InfoTest [11:05] coreycb, ddellav, zul: lescina was backed up on storage - freed up some space [11:06] the local apt repository had a morgue directory with the full history of deb updates in it... [11:14] mozart1893: no need fopr caps lock. are you still looking for assistence? if so, please describe the issue you're facing in more detail. === Guest21509 is now known as RickyB98 [11:43] I have install plain Linode server with apache and now want install my own mail server and need your help to do this. [11:44] use postfix or exim, read their manuals / getting started guides first of all. [11:44] poistfix has a pretty good manual / starting guide. [11:44] mailservers are a complex matter, be prepared to spend a lot of time. [11:45] ... or have someome else do it for you. [11:46] tomreyn, I want to do it on Linode server [11:46] cpaelzer: I looked at bug 1644595 but couldn't find the claimed violation. I commented. [11:46] bug 1644595 in krb5 (Ubuntu) "krb5-1.13.2+dfsg-5 source contains source subject to the aladdin license" [High,Incomplete] https://launchpad.net/bugs/1644595 [11:47] https://www.linode.com/docs/email/running-a-mail-server [11:48] Oh, I see what's going on. [11:49] rbasak: I found the file itself, but I'll read the bug - thanks for looking at it [11:49] cpaelzer: no, my mistake. [11:49] I'll comment with a correction. [11:50] rbasak: ah ok [11:51] rbasak: you might - as I did first - have stumbled over Al[l]adins number of "l" chars [11:52] It looks like it's typod in the boilerplate. [11:53] cpaelzer: so this should be reported to Debian as a policy violation. I can do that, or would you prefer to? [11:53] rbasak: If that is the way to go I can do that [11:53] rbasak: I was just a bit out of my comfort zone, that is why I pulled you in [11:54] cpaelzer: OK. Make it "Severity: serious" as it's a "severe violation of Debian policy". [11:54] I checked and it applies to current sid, too. === Valfor is now known as Guest24796 === disposable3 is now known as disposable2 === amoralej is now known as amoralej|off === amoralej|off is now known as amoralej|lunch [13:29] anyone thinks ZFS as a production fs is a bad idea..? [13:35] coreycb/ddellav/jamesage:The way that neutron loads its plugins have changed in ocata so im going through the neutron plugins in universe and make sure they are kosher [13:35] zul, got a pointer? [13:36] coreycb: yep [13:36] coreycb:gimme a sec [13:36] coreycb: for ex https://github.com/openstack/neutron-lbaas/commit/00568ab38e5e057aec9fd0231bf8fa3101e4ceac [13:39] zul, ok that's lbaas plugins. does that change how how plugins are configured? [13:40] coreycb:i dont think so because i was able to run the smoketests against ocata without a problem [13:41] zul, ok i'm still poking at horizon [13:41] coreycb: ok cool... [13:42] coreycb: i didnt upload horizon last week because i wanted to be double sure now im glad i didnt [13:42] zul, thanks [13:43] coreycb: ill be in plugin hell === amoralej|lunch is now known as amoralej === jgrimm-out is now known as jgrimm [14:13] please correct me if i am using ethernet pci passthrought in guest ,is that is full virtualization ? [14:16] samba35, nope, it's bare metal, that device cannot be shared with anyone else [14:17] ok [14:19] samba35, you typically use PCI passthrough only when speed, or some specific functionality is desired. Like giving a storage controller to a VM functioning as a virtual storage appliance is a good example. [14:19] samba35, it's also helpful for hacking kernel drivers in VM, it won't crash your main system [14:20] ok [14:20] thanks [14:21] you're welcome [14:21] i am confused with full and para virtualization [14:21] it is confusing :) [14:22] do you have better way to explain to lay man :) [14:25] In a nutshell: full virtualization is when you completely emulate a device, which can be expensive as the driver is doing "lots of stuff" as if it was still talking to the firmware on that device... but that device isn't there, the software just presents a device that quacks like it. [14:26] Thanks [14:27] para virtualization: is when you create a new class of device that performs the intended function e.g. network, storage. But actually knows it's in a virtual environment. So there's alot less overhead, which makes it faster, but also makes that VM's performanced biased towards that platform [14:27] virtio drivers are considered para-virtual [14:27] okk [14:28] pci passthrough is zero-virtualization, well everything else is virtualized but that device is not. So you get all the benfits of a VM... except the ability to move it, because the destination must have the same physical device available [14:29] can you please explain what do u mean by " completely emulate a device" [14:30] pick any device that shows up in lspci on your physical hardware, the VM presents a device that responds exactly as if the hardware was really there, down to touching it's registers [14:30] as a result, there are only a handful of fully virtualized devices available. [14:30] I can't even name one [14:30] i bought 1 motherboard to test pci-passthrouth with core2 quad cpu @40 us $ [14:31] does it have intel vt-d? [14:31] yes [14:32] then you're good, that supports "modern pci passthrough" you can attach and detach a device to a VM as much as you want. [14:32] can i do pco-passthrought without vt-d ? [14:32] you can... but you have to plan it. You literally have to tell the system kernel "don't touch this device", that's how Xen did it [14:33] no one does it that way anymore, if they can help it [14:33] ic [14:34] this info was new to me i never ever use xen [14:34] most people don't need to deal with pci passthrough, virtio is plenty fast [14:34] Thank you ! [14:34] :) [14:34] yw [14:34] have fun [14:35] i do use virtio with nic but what is best way to get gpu in guest ? [14:37] i have install palo alto networks firewall on ubuntu 12.04 as a 1st guest and again ubuntu as a 2nd guest as a dmz but i am not able to do some tasks [14:38] samba35, gpu as in cuda or brook gpu? or do you just want 3d accell? [14:38] as i am new to palo alto networks firewall [14:39] 3d accell? [14:39] 3 d accelerator ? [14:39] samba35, virtio-gpu and connect to it with spice [14:40] please correct me to use spice do i require special hardware ? [14:40] nope, it's just a viewer protocol [14:41] ic [14:41] the docs could be better: http://askubuntu.com/questions/250981/how-do-i-configure-spice-with-kvm [14:42] Thanks [14:42] follow the instructions for adding the virtio-gpu in virt-manager, then apt install spice-client-gtk python-spice-client-gtk [14:43] I'm surprised I can't find a native wiki on it... [14:43] jamespage, we're getting some xenial-ocata backport failures for packages that need debhelper > 10. but prior to those we had some that backported successfully and i think they were picking up debhelper 10.2.2 from xenial-backports. [14:43] coreycb, hmm [14:43] jamespage, does that make any sense? could something have changed? [14:44] coreycb, unlikely - backports is not enabled by default in either our build schroots or the ppa's [14:44] will give try later today [14:44] jamespage, ok. i did have debhelper 10.2.2 backported the the ocata uca for a bit until i realized it was in xenial-backports, and it seemed to be causing problems. maybe it's deb was lingering around after i deleted it. [14:45] coreycb, that might be possible [14:49] zul thanks for the heads up on the neutron plugins [14:51] ddellav: yep [14:52] jamespage, i think i'll try to backport it again after we test/promote b1 [14:52] coreycb, ack === devil is now known as Guest29760 [15:09] zul, horizon's uploaded. i have a charm change to go along with it. [15:09] ddellav, ^ [15:09] coreycb: sweet [15:12] coreycb excellent [15:14] zul, ddellav: do you have anything else outstanding for b1? [15:15] coreycb according to the list, just the debian syncs are remaining as far as i can tell [15:15] coreycb: we should be good to go...i was able to run a smoketest against the ci ppa [15:15] ddellav, ok yeah [15:15] zul, awesome, how many failures? [15:16] coreycb:we might as well upload newer versions of those "syncs" [15:16] a couple but those tests were not confgiured properly === BioKee is now known as Guest1712 === freyes__ is now known as freyes [16:44] powersj: based on your comment in bug 1643245, I think that should be Fix Released? [16:44] bug 1643245 in sssd (Ubuntu) "Fix related to skip invalid certificates" [High,Fix committed] https://launchpad.net/bugs/1643245 [16:45] rbasak: I only moved it to fix committed since zesty isn't released. [16:46] Should I have done something different? [16:46] powersj: we still treat it as Fix Released if it's in the zesty (ie. development) release pocket. [16:46] ok [16:46] then yes please, go ahead and change it to released. Thank you! [16:47] powersj: done, thanks. Consider it this way: if you uploaded a fix to Zesty with a bug reference, that would automatically change to Fix Released as soon as the package hit the zesty release pocket. [16:47] I appreciate that's a little unintuitive from a wider Ubuntu perspective. [16:48] rbasak: ah ok, that makes it clearer to me === Guest29760 is now known as devil_ === amoralej is now known as amoralej|off === JanC_ is now known as JanC [20:02] jamespage, zul, ddellav: xenial-ocata-staging looks good - http://paste.ubuntu.com/23550029/ [20:03] can't test zesty until some zesty charm updates land [20:07] coreycb: yeah waiting for them to land [20:07] coreycb: i think thats what i got as a result as well [20:10] zul, ok also there are 2 charm updates to land. beisner is picking up the swift charm-helper update to enable 2.11.0 in the zesty syncs, and there's a PR for horizon static files [20:11] coreycb: you mean this one? https://code.launchpad.net/~zulcss/charm-helpers/swift-ocata/+merge/311743 [20:12] zul, yup. sorry didn't know you had that out there for review. [20:12] coreycb:no worries [20:20] coreycb, I'll shove that lot into proposed then [20:20] jamespage, thanks [20:20] zul, coreycb: we should look to get libvirt into the UCA as well [20:21] anyone looking at that? [20:21] jamespage: i can probably [20:21] zul, I see qemu already went back [20:21] libvirt is a natural partner for that [20:21] there are some nice features for migrations [20:21] jamespage: qemu wasnt me ;) [20:21] oh [20:21] who was that? [20:22] jamespage, not me [20:24] coreycb, zul: might have been me [20:25] jamespage: just did a quick check we dont have to backport libxen this time ;)