=== ksx4system_ is now known as ksx4system === Amgine_ is now known as Amgine [07:38] Does ubuntu server need as much swap as there are ram ? (it is the default!) Thinking of resizing the swap from 32 GB (on a 40 GB SSD) because I'm running out of disk space on root [07:39] zeta: 1 or 2 gigs would be best [07:40] if you need more swap than that then there's something wrong with the amount of hardware you're using for the job :) [07:43] or just use swap files instead [07:44] also fine choice [07:44] the defaults are terrible though, no two ways around that [07:45] the defaults are so that you can suspend, I guess, but that's probably less useful for most servers... [07:47] wanted to double check because of the default [07:48] there might be use cases for hibernation, but I'm not using it === miczac is now known as miczac\away [09:57] if I do file -i mytextfile I get charset=iso-8859-1 but inside the file the characters are UTF-8 encoded. How is that possible? [10:15] hi, I have a Cloud Image based server, where I configured a bridge (br-mgmt) and after rebooting I haven't found a single error in the logs, but br-mgmt is down - `systemctl restart networking` solves the Problem, anyone with an idea what could be wrong? [11:07] coreycb: jamespage: could you take a look at https://bugs.launchpad.net/ubuntu/+source/murano/+bug/1659270 please? zigo has fixed it upstream, should only need a sync [11:07] Launchpad bug 1659270 in murano (Ubuntu) "Startup of murano-cfapi is broken" [Undecided,New] [11:44] cpaelzer, whoop whoop =) new qemu [11:56] yep xnox [11:56] it was actually good for 2 weeks, but one of the dependent dep8 failed due to OOM [11:56] thanks to apw this got a test on a bigger machine and migrated now [11:57] apw: btw - is there any flag that coreycb / zul could add to d/t/* mark this test to need more ressources ? [11:58] cpaelzer, nope, this is a per package hint server side [12:00] apw: is the hint you added persistent now ? [12:04] cpaelzer, yes [12:06] frickler:its been fixed in zesty [12:26] zul: looks like it at least would need another sync into ocata UCA, then. getting a fix for Newton, which we are currently deploying, would be even better [12:27] frickler: yeah i might add it to the next SRU cycle thats probably going to happen this week [12:45] cpaelzer, apw: thanks for the hint update === miczac\away is now known as miczac [13:22] hello, i'm trying to fix my own ipv6 router/tunnel/gateway [13:23] i have a server which has an ipv6 subnet [13:24] on that server i created a inet tunnel mode sit [13:24] i created a v4tunnel on the client side [13:24] that all seems to go well [13:24] i can connect [13:25] i can even ping a local ipv6 on the server and get response [13:25] but if i want to ping to google on ipv6 for instance it stops [13:25] i allowed forwarding in the v6 iptables, if i add logging i see the request is getting there [13:26] also tcpdump shows this [13:26] but it is not passing the request, like it is not routering it or something [13:26] i also have enabled the net.ipv6.conf.all.forwarding [13:26] so what am i missing? [13:44] rbasak: can your permissions clean the non Ubuntu AppArmor, libvirt, and systemd tasks from bug 1594902 ? [13:44] bug 1594902 in libvirt (Ubuntu) "Failed to upgrade to libvirt-bin 1.3.1-1ubuntu10.1 on Ubuntu 16.04 64-bit" [Undecided,Confirmed] https://launchpad.net/bugs/1594902 [13:47] cpaelzer: I can set the upstream AppArmor task to Invalid. But I can't do anything with the other two. [13:47] cpaelzer: a workaround might be to change the project for those tasks to Ubuntu. Then you might be able to delete them :-) [13:49] rbasak: that was evil but it worked [13:50] the log on this bug looks messy now === madwizar1 is now known as madwizard [13:52] it all quickly was qemus fault, but then not :-) === JanC_ is now known as JanC [14:50] good morning [14:50] I am curious if there is a "fully updated" iso of older versions of ubuntu server [14:50] I am looking for ubuntu 8 LTS updated as much as it could be and am curious if such things exist? [14:55] upon further looking up it seems that Ubuntu 8.04.4 LTS is the final version :) === zerick_ is now known as zerick [14:58] BrianBlaze420: it is, also 8.04 went fully EOL in 2013. [14:59] yeah I am aware just trying to update a server it 8 LTS so I need to make sure I can test :) [14:59] it is 8 LTS* [15:00] coreycb, where are we with ocata b3/rc's? I'd like to flush things through before we to much closer to release daty [15:02] I would not update from something that old [15:02] I would clean install [15:02] coreycb: FYI for bug 1641532 the Xenial and Yakkety fixes are now in the unapproved queue for SRU [15:02] bug 1641532 in Ubuntu Cloud Archive liberty "machine-types trusty and utopic are not unique (depend on the qemu version)" [Critical,In progress] https://launchpad.net/bugs/1641532 [15:02] coreycb: just FYI so you can coordinate UCA release of the change [15:03] jamespage, they're not completely tested, but i agree [15:03] BrianBlaze420, there is no fully updated server iso, you * could * do the old-releases trick, just for testing [15:04] coreycb, I'm sure mwhahaha would rather have visibility sooner than later [15:04] jamespage, let me put some focus on that and come back to the charm stuff [15:04] BrianBlaze420, see https://help.ubuntu.com/community/EOLUpgrades .. but i guess you waste your time [15:04] * jamespage looks at the reports [15:04] yeah I know what I said didn't make sense but I could only find the original iso and not the later realeaes of it [15:05] I found the last release and will work with that :) [15:05] the idea is to update to the latest LTS :) [15:05] coreycb, lest at least push to -proposed [15:06] BrianBlaze420, 8.04 > 10.04 > 12.04 > 14.04 > 16.04 ... [15:06] BrianBlaze420: I've tried things like that before and only had trouble, like a lot of it, to the point that I don't do it anymore, I just look for ways to reinstall, swap something in or whatnot [15:07] I am stubborn and so will try myself and see where it leads me. I appreciate the input :) [15:07] sure thing, np === miczac\away is now known as miczac [15:07] BrianBlaze420, have fun :-D [15:12] cpaelzer, ok sounds good. mind pinging me when they're in proposed and then i'll get the patch into liberty-proposed? [15:13] coreycb: you should be subscribed to the bug, but I'm fine pinging [15:13] coreycb: I'll run my full test matrix anyway then [15:13] cpaelzer, yeah i think i am [15:13] hope for good SRU review on that - it is complex enough to drive anybody mad === jj-_ is now known as jj- [15:28] coreycb: so manila barfs with sqlite because its trying to rename a colum, fyi [15:29] zul, is that an upstream bug? [15:29] coreycb: yeah [15:29] zul, ok got a bug #? [15:29] coreycb: i switched the dep8 test to use mysql [15:29] coreycb:not yet [15:30] zul, ok [15:31] coreycb: https://bugs.launchpad.net/manila/+bug/1664279 [15:32] Launchpad bug 1664279 in Manila "Traceback when configuring manila with sqlite." [Undecided,New] === Agent is now known as Guest87159 [16:05] another problem I'm getting with preseeding is if I'm trying to reinstall a box that had disks with raid on it [16:05] it seems like mdadm will detect the raid and set the device up, at which point the installation fails because "resource is busy" [16:06] I have to get onto another console, dd the MBR/kill the raid, and then reboot/re pxe and it works [16:07] googling around it seems people have had a similar problem and are basically doing the same with a d-i early command, but that seems kinda nasty... [16:07] does anybody know fo a better solution? [16:08] this is basically a similar problem to the one I was asking the other day with the installer failing because there was no "root" partition defined, which is just a symtpom of not wiping the disk and repartition but trying to reuse what's there [16:09] you can reuse what is there [16:09] but you still have to label to the installer what partition is should be installed with what [16:10] assign root, boot, swap, maybe home, var, usr, ... [16:15] patdk-wk: that's the thing, I don't want, these disks were from another system. And in pretty much all the cases I have I don't care about reusing [16:15] quite the opposite, I wish there was just a "nuke" option [16:15] that stopped asking so many questions and just simply wiped and reinstalled everything completely disregarding what's there [16:15] there is [16:15] I knwo what you're talking about, doesn't work for me it seems, maybe I'm doing something wrong [16:16] I think I know* [16:17] d-i partman/confirm , partman/confirm_nooverwrite [16:17] is that what you had in mind? [16:19] also to make matters worse with preseeding it seems the ubiquity installer doesn't honor a whole bunch :(, but I guess that's not a ubuntu-server problem [16:37] rbasak: https://bugs.launchpad.net/ubuntu/+source/aodh/+bug/1645772 [16:37] Launchpad bug 1645772 in ironic (Ubuntu Yakkety) "[SRU] newton stable releases" [Undecided,Incomplete] [17:44] coreycb: https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1664306 [17:44] Launchpad bug 1664306 in nova (Ubuntu Yakkety) "newton stable SRU releases" [Undecided,New] === glomerowisk is now known as Mr_Anderson === Mr_Anderson is now known as MrAnderson === MrAnderson is now known as Mr_4nd3r50n === Mr_4nd3r50n is now known as MrAnderson === MrAnderson is now known as Mr_4nd3r50n [18:39] powersj: ohai [18:39] teward: hi [18:40] powersj: thanks again for your triage work on NGINX bugs, but when the error is 'nginx.conf missing' it's able to be considered Invalid, because that's not a package issue, usually a system-specific issue but not a bug and easy to work around [18:41] teward: ah that makes sense. Sorry I have a habit of moving to incomplete to let the user try to fix that and get back to us. [18:41] powersj: no problem :) [18:41] Did you already fix that up? [18:42] powersj: there's also a few other cases as well to point out [18:42] "address in use" as "cannot bind to :80" is also an Invalid [18:42] and I think a few others that *should* already be caught by the apport hooks I wrote [18:42] probably gonna have to update them after the merge [18:42] powersj: and yep I did [18:42] and provided the workaround [18:42] ok, thanks for that list I'll add those to my notes [18:43] powersj: no problem, hope you don't mind me poking [18:43] :) [18:43] teward: please do :) appreciate it [18:43] I should probably document NGINX-specific bug triage things on the wiki [18:43] 'cause there's a few cases that are known to show up occasionally but aren't package issues [18:43] (and especially show up on VPSes built from a premade KVM or OVZ image) [18:44] powersj: far be it for me to tell the Canonical Server Team what to do, but... :P [18:44] haha [18:44] i do appreciate the triage work though, it helps me quickly weed out the evil ones [18:44] and then focus on actual bugs [18:44] current focus: nginx merge from Debian [18:45] now that the several hours of evil PIE-failure debugging is fixed heh [18:45] but first, my midterm studying needs done. [18:45] :) good luck [18:46] meh not really that hard [18:46] just have to 'start over' like we did for 14.04 and build the delta from scratch again [18:46] given the radical changes to the packaging it's easier to rebuild than merge [18:46] the core issue was the fPIE/fPIC compile evils, but now that that's been fixed in Debian, yay less delta for me :P [18:47] (but never an issue except for the merge heh) [18:47] as for the midterm, blah [18:47] it's art history, as long as I get a B I will be happy [18:49] ruebens liked fat people, bosch was insane. lapis lazuli. and, uh, caravaggio [18:49] how'd I do? [18:49] *throws an Ubuntu 17.04 USB stick at sarnold* [18:50] sarnold: of relevance, pending SRU, because HTTP/2 stuff/fixes, just so you're aware. https://bugs.launchpad.net/ubuntu/+source/nginx/+bug/1663937 [18:50] Launchpad bug 1663937 in nginx (Ubuntu Yakkety) "[SRU] Please update nginx in Xenial and Yakkety to 1.10.3" [Wishlist,Triaged] [18:53] teward: nice [19:06] and i'm still working on the merge. [19:06] after my midterm :p [19:29] jgrimm: i think my blueprint stuff is all in -proposed or -release now, fyi -- just working on getting it into -release as appropriate [19:30] nacc, nice! indeed the bp is quite green for the most part excepting for testcase failures (that are not related to the specific upload) [19:31] jgrimm: yep === elmo_ is now known as elmo [19:57] jamespage, ocata-staging is ready to promote [19:57] jamespage, one more charm update needed for cinder [20:37] server team members: is there anywhere specific I can drop a Wiki page with certain special-cases for nginx triage that diverge from standard triage procedures? [20:37] nginx bug triage* [20:37] I know i told powersj here direct those two special cases, but... [20:37] teward: probably worth doing somewhere under the ServerTeam pages, but don't konw specifically where [20:37] jgrimm: powersj: --^ [20:38] I'm booting an Ubuntu AMI with 2 x 2Tb disks. Looks like cloud-init only formats them to 1Tb??? [20:38] hm so I do have a page that I keep for my use of responses and tips [20:39] teward: I added your notes to it, but gotta remember to look at it :P https://wiki.ubuntu.com/DebuggingServer [20:39] ideally we update the apport stuff to auto invalidate those === Guest87159 is now known as Agent [20:42] nacc, powersj, teward: additionally feel free to add a nginx as a "Project of Interest" with its own page if you feel that's useful too. [20:43] jgrimm: good point! [20:43] its pretty free form as long as we don't go crazy with clutter / desire to keep updated & relevant [20:49] jgrimm: where would i put that? just under the Server Team? Until we get a buttload of NGINX issues that deserve a whole page, I think shoving it into DebuggingServer will work. So I'll go poke there. Probably deserves its own page somewhere in the triage guide, or a link to that page heh [20:50] teward, off the main Ubuntu Server page there is a section for "Packages of Interest" to link their own pages. Just offering that up if there's enough useful info to put in its own page. [20:50] I'm fine with DebuggingServer [20:51] teward: and you can do a link with ancchro on the triage page, myabe? [20:51] I'm booting an Ubuntu AMI with 2 x 2Tb disks. Looks like cloud-init only formats them to 1Tb??? [20:54] fwiw I like the separate page idea as well + a link === mikal_ is now known as mikal [20:56] jgrimm: I will probably write a page, at some point. [20:56] nacc: I'm not awake, ancchro? [20:56] teward, sounds good [20:56] s/awake/fully awake/ [20:56] powersj: i tweaked said page a little, added log excerpts as examples of such cases. [20:57] teward: ok thx [20:57] There's another case I know that is also Invalid, but I forget what it is at the moment [20:57] I think it's "Please add foobarbaz to nginx" where foobarbaz is any given third party module [20:57] teward: typos! anchors :) [20:59] yay wiki lag >.< [21:08] powersj: even better, now, canned comments to use for those two cases. :p [21:08] teward: oh see now you are marking me really happy :) [21:08] at some point this is going to become infinitely long, so I'll have to move it to its own 'Package of Interest' page heh [21:08] keeping that my cut-and-paste page :) [21:09] powersj: i give two separate canned comments, depending on which case we see, feel free to adjust as you wish (same for all other Server Team people). [21:09] but it will at least indicate of course the status to set, why we set it, and how someone should respond if they still think it's a bug. [21:09] And since i'm subbed to ***every*** nginx bug for all notices, even if the bugs are closed as INvalid or such, I'll see the alerts xD [21:09] even if the rest of the Server Team doesn't :) [21:10] ooo I see LXC is on the POI page... [21:11] I got tired of having to manually adjust all my LXC containers, I wrote a fairly in depth Python script that can create a container *and* modify the container, sort of 'bootstrapping the container to specified needs' at the time of creation. [21:11] it's on github now xD [21:11] in case, you know, anyone is interested :P [21:11] (it's not LXD, but that's because dnsmasq wouldn't behave with LXD on my system oh well) [21:42] coreycb, ok promoting to -proposed now [21:43] jamespage, thanks [21:59] I'm booting an Ubuntu EC2 instance with 2x2TB disks attached. They are mounted as 1Tb disks instead. I think cloud-init is doing this. Is this a known bug in cloud-init? [21:59] dgarstang: you could ask in #cloud-init, but I don't think so. They probably will want to see your cloud-config if you are doing any customizing [22:00] powersj: Zero customization [22:01] I don't think it's a filesystem issue. Fdisk reports the disks as 2Tb each [22:11] I have an instance with 2 x 2Tb disks. When I run "mount -t ext4 /dev/xvdj" I get a 1Tb volume. Why? [22:19] dgarstang: look at how they are partitioned? what format, etc. === tanuki_ is now known as tanuki [22:59] jgrimm: what should go in the "Packages of Interest" page, if I decide to make one? [23:00] teward, its really free form, if you feel important items to know about nginx in ubuntu [23:02] cool. It will contain a brief blurb about what it is, maintaining it, what to not ask about, and a small guide to the versions in the Repositories. Since they're all different right now :P [23:06] teward, sounds good. thank you! [23:08] jgrimm: I made a note on the whiteboard for the server work blueprint [23:08] namely about http/2 readiness/stability in Apache, because IIRC the Security team NACK'd it [23:08] not sure if that's changed [23:08] jamespage, coreycb: did you guys propose m3 or rc1? there's a bunch of stuff that got broken/fixed for nova between m3 and rc1 [23:09] (it's why NGINX has HTTP/2 but not Apache currently) [23:09] teward, I saw that (and yes I was aware), nacc just wanted a reminder to look at it again in the future [23:09] OK [23:09] still leaving my comment if there's no objection :) [23:09] no objection at all! :) [23:09] cool. :) [23:11] mwhahaha, rc1 was promoted to -proposed a few hours agao so may still be in progress [23:11] coreycb: http://logs.openstack.org/66/433266/1/check/gate-puppet-openstack-integration-4-scenario002-tempest-ubuntu-xenial/e043e14/console.html#_2017-02-13_21_16_13_600854 so you're missing the create_cell command [23:12] which i thought was in rc1 [23:12] maybe we don't have it yet in that test run [23:12] ah nm still b2, 2:15.0.0~b2-0ubuntu4~cloud0. i'll check on it tomorrow [23:13] mwhahaha, ok cool, create_cell will be in rc1. i've been using it the past week while testing staging.