[00:23] ahasenack: total guess, but bin/python/setup.py seems to be missing any statement of its dependnecies [00:25] Had a strange lvm issue here. LV mirror with 2 sides plus log. Made it inactive to "lvextend -L 60G Archive/SourceCode" and then "lvchange -aly Archive/SourceCode" fails with 'device or resource busy. Eventually noticed with "lvs -a -o +devices" the attributes of the mirror sides were marked I (inconsistent) and the mlog no longer had the 'l' flag - compared to other similar mirrored LVs. Then, after [00:25] having used the 'lvs' command, 'lvchange -aly' worked!' I'd already tried to force a resync and that hadn't helped. Any ideas what was going on? [01:05] do i just install squid and im on a proxy server ? or how does that work ? [01:10] b18c5: if you are connected to a server and install squid there then you're on a proxy server since squid is that. but i'm not sure that's what you're trying to do. [01:10] honestly im new on this o/s im used to windows, looking for a sense of security lol [01:12] installing squid will not magically provide security. in fact any extra listening service you install increases the attack surface. [01:12] what are you trying to achieve? [01:13] blocking my i.p adress [01:13] if you block your ip address from your server then you won't be able to connect to your server anymore. [01:14] sorry, i'm trying to block my i.p adress from other people [01:14] maybe you want to access the internet (or services on the internet) without disclosing the ip address of your router at home? [01:14] I'd suggest grabbing a book such as TCP/IP Network Adminstration before going any further [01:15] mainly on irc lol i feel like a target on here [01:16] if you wish to irc slightly anonymously then your best bet is Tor [01:16] also read up on "irc bouncer" [01:16] how would tor work ? isn't that for web browsing ? [01:16] tor works for any tcp protocol [01:17] irc is a tcp protocol, as is http(s) [01:17] when i whois myself though my information pops up [01:17] b18c5: if you've a registered nickname on Freenode you can ask in #freenode for a mask [01:17] i did lol [01:17] note that a mask is only a 'best effort'; if services is down when you connect it can't help. [01:18] i tried a vpn but that shit was confusing [01:21] besides that issue, how would i do a disc clean up on a linux based system ? [01:22] sudo dd if=/dev/zero of=/dev/whatever bs=131072 [01:22] that block size is just 128 kilobytes. there's nothing magical about that number. It's just big and a power of two. [01:23] some guy has put together 'dban', 'derik's boot and nuke', iirc, that does something similar for all connected hard drives. [01:25] the names' "darik", and that's not really ubuntu related, but it's a commonly used option for overwriting hdd's. [01:25] it's easy to do from initrd; just boot with "break=premount" and then use the shell and dd to do parallel dd's on each device [01:26] i think i ended up deleting and upgrading everything [01:26] tomreyn: d'oh! thanks :) [01:27] sarnold: and i mixed you up with b18c5, sorry. [02:06] Hello, I'm running into an issue where my Ubuntu server is not responding to port forwarded requests. [02:06] I have disabled ufw. Nginx for example works on the lan and localhost but not from wan [02:07] what error messages do you get onthe clients and the server? [02:07] nothing, it just hangs [02:07] I see the packets hit with tcpdump, but curl hangs [06:21] Hello, I asked this over in #ubuntu and they directed me here. I am trying to convert my /etc/network/interfaces to netplan, and am having trouble getting the routes to work the same. I was hoping someone would be able to help me make a new netplan yaml file based on my old config. Here is a picture of it (with netmask having a typo, woops) https://puu.sh/z9RAL.png [06:28] goot morning [06:28] taking a look HateNetPlan ... [06:28] Good to hear! [06:28] cpaelzer: thanks [06:29] get to bed nacc! [06:29] cpaelzer: yeah :) [06:32] HateNetPlan: FWIW, I don't think ifupdown is going away. Merely deprecated. [06:32] I plan to cling to it like grim death. [06:32] mason: +1 [06:33] i believe the default on fresh installs may have switched, or might switch, or something [06:33] Someone will save it, that much I am sure [06:33] but upgrades should keep ifupdown going forward [06:33] and you can always install it manally [06:33] Ubuntu 17.10 swapped already [06:33] yes [06:33] It's default netplan [06:33] it is no more there by default [06:33] ifup/down [06:33] but still exists [06:35] HateNetPlan: as you already found hooks no more directly exist in netplan (as some of the backends don't have them) [06:35] see https://wiki.ubuntu.com/Netplan#Frequently-asked_questions [06:35] so for your case the recomended way is to encode the former logic in systemd services afaik [06:35] like https://www.freedesktop.org/wiki/Software/systemd/NetworkTarget/ and such [06:36] it is a bad time of the day - afaik cyphermox works on converting some packages with hooks [06:36] I know they work on a how-to describing how to convert several old cases [06:36] but it doesn't exist yet (not that I'd know of) [06:37] HateNetPlan: I assume based on your eni-rule you wanted it to push/pop the routing rule every time it does up/down [06:37] Yes. [06:41] to my lack of good phrasing netplan is meant to provide one simple config so that you don't mind to have to understand networkd and/or networkmanager - so you become independent where you are for the most common cases [06:41] I think we can agree that your case is interesting, but maybe not the most common one [06:41] netplan doesn't prevent you to use any special backend features [06:41] so on a server it would render to systemd-networkd [06:42] I wonder if we could express it there somehow [06:42] but https://www.freedesktop.org/software/systemd/man/systemd.network.html is rather lengthy for a fast "ah there it is" :-) [06:42] Maybe I should have started around there haha, spent more time then I care to admit on this [06:48] Thanks by the way. Now that it can even connect to the internet I can grab ifupdown as a backup [06:48] HateNetPlan: like this maybe as a start https://serverfault.com/questions/667319/systemd-networkd-and-direct-routes [06:49] if you want to go on trying to convert it fully [06:49] Actually, based on that link [06:49] That explains how much lxc is connected correctly [06:50] Because that's how the .network file looks [06:51] Ok, well I need sleep. It's 1 am for another day in a row, and I am just happy someone was able to help me [06:51] I'll leave this open just to look back later to be sure [06:52] good luck on your way to a nic rename one day :-) [06:58] Good morning [06:58] hi lordievader, great Friday morning to you as well [07:01] Hey cpaelzer How are you doing? [07:03] good [07:03] after all - its Friday :-) [07:03] Indeed [08:25] Any plans on https://bugs.launchpad.net/ubuntu-release-notes/+bug/1531864 in the near feature? I think, people should decide themself, if they want to enable http2 support for there Apache2 from official packages... [08:25] Launchpad bug 1531864 in Release Notes for Ubuntu "HTTP/2 disabled in Apache httpd" [Undecided,Fix released] [10:26] rz_o1: yeah [10:26] rz_o1: that is done actually [10:27] well I need to read the details of the bug you linked, but I enabled http2 for apache in 18.04 [10:28] any plans for 16.04 too? :) but i am glad to here that for 18.04 [10:29] rz_o1: not that I'd know of any 16.04 plans for it [10:29] the security maintenance ack for the http2 lib that is used only covers the version in bionic onwards [10:30] nginx had it earlier since they use a different implementation which was ack'ed before [10:30] by getting nghttp2 not only apache but also curl got http2 now - so you get a test tool against http2 with it as well now [10:36] ok thansk for the info [10:36] yw [11:38] coreycb: I've done all of those dep updates apart from os-vif - which is being awkward with unit tests under py3 [11:38] working that atm [11:38] tobasco: hi expect to get to re-adding the py2 compat patckage for gnocchi next week when we do m3 [11:41] coreycb: re pxc-5.7 - I'm fairly happy with my branch now - its in https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3110 [12:06] nacc: indeed (wrt python3-ply), this fixed it: https://pastebin.ubuntu.com/26463772/ [12:31] jamespage: great, thanks. I'll take a look at 5.7 and try it out. i see a lot of b3's are out too so i'll plan on starting on those today. [13:08] jamespage: i'm getting started on b3's [13:43] rbasak: hi, could you please kick an import of sssd into git-ubuntu? 1.16.0-5 showed up in debian unstable [13:43] g-u still has 1.16.0-3 [13:43] ahasenack: running [13:43] thx [13:51] coreycb: bah minor upgrade issue from 5.6->5.7 - that split out of wsrep.cnf is going to break things so will drop it [13:51] jamespage: ok that's fine with me [13:51] jamespage: cool, thx for the update, i will make sure to work on the puppet side when its available [13:55] ahasenack: should be done [13:56] rbasak: hm, pkg/debian/sid is still at 1.16.0-3 instead of 1.16.0-5. LP could be lagging behind then? [13:56] ahasenack: yes: https://launchpad.net/debian/+source/sssd [13:57] ah, I see [14:01] coreycb: btw did you confirm that pxc 5.7 is broken with gcc-7 or was that a forward copy from pxc 5.6? [14:02] jamespage: i think i confirmed that but it's been a while so i'm not positive [14:04] jamespage: i feel like it had compile errors with gcc-7 === zhhuabj_ is now known as aabb [15:40] hi [15:40] i would like to know how we can custom package on a OS ISO ? [15:41] like that all require package are ready on the OS by defautl [16:54] rbasak: when you have a moment, could you chime in on my zstd branches that are up for review? [16:54] rbasak: cpaelzer already did, but you have way more context [16:54] rbasak: btw, cpaelzer's needs-fixing were addressed (I had forgotten to run update-maintainer) [16:55] rbasak: and one more thing for your queue, if you could: I'd like to know if having a build-dep on python3-distutils-extra (universe) for landscape-client (main) in bionic is ok: https://bugs.launchpad.net/landscape-client/+bug/1743562 [16:55] Launchpad bug 1743562 in Landscape Client "build-dependency on universe package: python3-distutils-extra" [Medium,New] [16:56] I think it's used just when building the package and no code from it will be used at runtime, and I think that's accepted, but I wanted to be sure [16:59] OK [17:03] ahasenack: yep, makes sense [17:03] ahasenack: seems like an upstream bug then? [17:04] cpaelzer: we had a separate discussion on this, and mayb ewill backport it to 16.04 [17:04] cpaelzer: but only if we backport nghttp2 as in bionic to it, as well [17:04] rbasak: around? [17:06] nacc: o/ [17:06] rbasak: HO? [17:06] We can sync? [17:06] Yep. Two minutes [17:06] rbasak: yeah [17:06] sure [17:06] Standup HO? [17:06] rbasak: yep [17:12] nacc: bind? Yes, I filed it upstream and with debian [17:12] ahasenack: yeah cool [17:13] nacc: right now, that is our only delta with debian :) [17:13] ahasenack: seems fien :) [17:21] rbasak: the importer picked up sssd 1.16.0-5 \o/ [17:21] 30 minutes ago [17:21] DSC file for 1.16.0-5 [17:21] nice [18:12] Is there any reason that loads of services still exist in /etc/init.d when a real systemd service would be much more suitable? [18:14] nothing specific, no [18:14] some packages even ship both [18:18] nacc: hmm, old/debian seems wrong in this output, no? https://pastebin.ubuntu.com/26465717/ [18:18] it should have been 1:9.10.3.dfsg.P4-12.6 [18:28] boxrick: not every packages are shipping systemd units. [18:28] nacc: ah, it's a difference between the current snap and the code from the MP with the branch ubuntu-package-merge-base [18:28] the branch gets it right [18:29] $ ~/git/projects/usd-importer/bin/git-ubuntu merge start --tag-only pkg/ubuntu/devel -f [18:29] 01/26/2018 16:28:20 - INFO:Created tag old/ubuntu for version 1:9.10.3.dfsg.P4-12.6ubuntu1 [18:29] 01/26/2018 16:28:20 - INFO:Created tag old/debian for version 1:9.10.3.dfsg.P4-12.6 [18:29] 01/26/2018 16:28:21 - INFO:Created tag new/debian for version 1:9.11.2.P1-1 [18:29] boxrick: if you know how to create systemd units, you could propose those to the package maintainer in Debian [18:31] nacc: but then the lint from that branch gets it wrong again [18:52] is there any known issues upgrading to a 4.4.x kernel on a stock trusty server? the upgrade is failing and I am getting stuck in initramfs after a restart [18:53] have you tried regenerating the initramfs? [18:53] did you also look for disk full issues, specially in /boot? [18:54] i ran update-grub, dont think ive ever needed to regenerate that before [18:54] (if you have that as a separate partition) [18:54] we do, but never had an apt-get install linux-image... error before [18:56] I had some weirdnesses lately with this. It tried to do Grub1 things when the machine is actually using GRUB2. Like, it wanted to do changes to menu.lst [19:01] lets see if this works, round 2. [19:01] im scared to reboot lol [19:03] nope failed [19:07] seeing if i can do it at preseed [19:13] Epx998: did you check the things I mentioned? [19:13] update-grub doesn't regenerate the initramfs file, fwiw [19:14] yeah on a reimage, I ran update initramfs -c -k linux-image-4.4.0-31-generic [19:15] then update-grub again, rebooted [19:15] Epx998: "update-initramfs -uk all" is what I use usually [19:15] it can't find the root device/fs? [19:15] yeah, -k all [19:16] to avoid silly mistakes typing the kernel version [19:16] I saw the all option, I wasnt sure about it [19:16] but it could break your older kernel/boot [19:16] i added a d-i option to my preseed to use the kernel i want, its restarting now [19:16] I'm assuming you are rebooting into that [19:16] or it's a new install? [19:16] the -k option wants only 4.4.0-31-generic (aka uname -r) [19:16] new install [19:16] ah [19:16] this 4.4.x is an hwe kernel? [19:17] no just out of the repo after a new install [19:17] you said it was a trusty server so a 4.4 outta be a HWE one [19:17] same result with a preseedf [19:18] funny thing is my installer image is using 4.4 heh [19:19] ill try again manually with the -uk all, if that fails ill hand off with the stock image and let dev deal with it [19:20] Epx998: with 14.04, at some point release the ISO got the HWE kernel added IIRC [19:20] sdeziel: this is a pxe/netboot install [19:21] Epx998: then you must be shipping a 4.4 vmlinuz ;) [19:22] running "file vmlinuz" on the pxe box should tell you [19:23] pxe box? you mean tftp server? [19:23] yes [19:24] yeah i remade my netboot images, trusty is using an updated 4.4 [19:24] i think even my precise images are 4.4 [19:24] wait no [19:24] but trusty is yeah [19:26] Epx998: I'd try https://askubuntu.com/questions/953430/using-preseed-how-do-i-select-the-hwe-kernel with sed 's/hwe-16.04/lts-xenial/' [19:28] hwe kernel is married to a specific version right? [19:29] Epx998: for Trusty, the HWE kernel you'll get will always be a 4.4 one [19:30] I think that earlier some other older versions were made available but they are now superseeded by the one backported from Xenial [19:31] https://wiki.ubuntu.com/Kernel/LTSEnablementStack#Kernel.2FSupport.A14.04.x_Ubuntu_Kernel_Support shows it well [19:32] ok reinstalled, about to do the kernel [19:32] so give linux-hwe-generic-trusty a try? [19:33] not sure such package exist [19:34] that was in my search for anything hwe [19:34] sorry, you are right it does exist [19:35] trying this 1 last time with "update-initramfs -uk all" [19:35] the HWE wiki page mentions "linux-generic-lts-trusty" though [19:35] yeah I saw that to [19:39] well fiddlesticks [19:39] the version in the name is the release version the kernel is based off of. for 14.04, to get the 4.4 kernel, you want linux-image-generic-lts-xenial. [19:40] sudo apt-get install --install-recommends linux-generic-lts-xenial [19:40] - this just aborts after I type 'y' heh [19:42] ok reboot with lts-xenial [19:44] i think that one worked [19:44] now to see if the ixgbe driver builds [19:45] ah no need, awesome. i think this is as close as ill get to what dev wants [19:55] How do you set up new users for FTP or SFTP to only have access to a certain folder? (ex: /var/www/html) [19:56] chroot is the safest way [20:00] akern07: openssh has a nice way to do chrooted SFTP, man sshd_config and search for "internal-sftp" [20:00] Thanks I'm going to try that out [20:07] ok thanks for the help, i think i got these 2 hosts where they need to be [20:08] great [20:47] ahasenack: please file bugs, i'm not context switched in [20:59] powersj: is there any equivalent to pytest's parameterize that unittest supports? Or do we nneed to use a framework (either pytest or nosetest) to do it? [21:02] nacc: I don't recall there being a direct way [21:02] powersj: found it (subTest() context manager) [21:02] it's new in 3.4 [21:02] ah [21:02] https://docs.python.org/3/library/unittest.html#distinguishing-test-iterations-using-subtests [21:03] means i can drop one import from my newe tests :) [21:40] i have to try for 4.4.0-31 again [22:01] powersj: what's the preferred layout for tests? if testing script.py, script_test.py? [22:03] nacc: I prefer tests in the same directory as code and name_test.py as you state [22:03] powersj: ack [22:12] my workmate tried 4.4.0-31, might be we need extras as well, which ive not seen that req before [22:30] yeah added the extras results in the happy path