[00:01] rbasak: powersj: fyi, i'm testing (https://jenkins.ubuntu.com/server/job/git-ubuntu-ci/357/) some further fixes to our snap that may unwedge CI [00:01] thanks to kyrofa's help [00:01] i'll need to run it a few times ( powersj if there's any way you can clone that same job 3 or 4 times easily, that'd be great), as that's what failed before (about 1/4 of times run in parallel) [00:02] oh i'll just rebuild it a bunch [00:06] rbasak: hrm, our reimport tagging needs to apply to applied/ too, doesn't it? [00:06] rbasak: does it look werid to have applied/reimport/0 ? [00:50] powersj: rbasak: it does look like it fixes it [00:50] rbasak: so i'll need to rebase my branches so we get clean CI runs, if that's ok with you? [00:50] contentfully the same for you [01:13] powersj: rbasak: pushed to master, but i'll hold off on rebasing just for rbasak's sake [02:57] rbasak: actually, i lied, i'm going to push over the top to retrigger CI [03:01] rbasak: branches pushed, should pass CI in all 4 cases === led_ir23 is now known as led_ir22 [06:06] Good morning [08:14] What is the kernelversion in 18.04? [08:18] OpenTokix: 4.15.0.13 atm [08:22] ok, nice === ackkk is now known as ackk === beatzz_ is now known as beatzz [10:40] is there any other method for complete unattended install for ubuntu ? cause MAAS is failing really bad [10:44] gunix: preseed [10:45] OpenTokix: yea, and pxe boot [10:45] OpenTokix: when I try to PXE boot from the ubuntu install ISO i get an error because the format is not correct [10:46] OpenTokix: is there any blog on how to prepare iso images for pxe boot so that they work? [10:46] no, preseed with iso is broken [10:46] why would you ever do that? [10:46] OpenTokix: i didn't add preseed to the iso [10:46] Dont boot the iso from pxe [10:46] OpenTokix: i just tried to boot the iso via PXE and it failed [10:46] You just boot the kernel + initrd, and fire of a netinstall [10:46] than how do i boot ? [10:46] hmm. [10:47] OpenTokix: do you have a tutorial for that ? [10:49] gunix: hold on, I am preparing a pastebin for you [10:54] OpenTokix: you are kind. thank you. [10:57] gunix: https://pastebin.com/S2BBwKuH [10:57] gunix: this is what I use fro 16.04 [10:57] This will give you swedish keyboard layout, for us, change all sv/se to us and you will be fine. - I keept it because to get correct layout was non-trivial D: [10:57] :D [10:58] This is both location of installer binaries, pxe config and preseed file in the same paste. - Also username and password is: ubuntu for this install. [10:58] OpenTokix: where did you add that PXE boot section ? [11:04] gunix: in the configuration my tftp-server. [11:04] https://debian-administration.org/article/478/Setting_up_a_server_for_PXE_network_booting [11:04] It's the same for all debian and derivates. [11:05] OpenTokix: does it work with http server or do you really need tftp ? [11:06] gunix: yes, pxeboot support only tftp [11:06] but you pull the preseed from http. [11:08] OpenTokix: does it default to the LABEL 2 when you boot? [11:08] ipxe supports http(s) network booting [11:08] i mean, from what i understand, that will be a boot option [11:08] but you'd need to get ipxe installed ony our NIC / system firmware beforehand [11:08] tomreyn: yes, that would help, since we already have nginx in the network ... [11:09] tomreyn: the system is proliant gen9 [11:09] just one system? [11:09] tomreyn: about 6 or 7 [11:10] tomreyn: i will get about 50 more incoming but not all gen9, some gen7 and gen 8 [11:10] so i don't know if the old tech can do ipxe [11:10] uuh those are eol [11:10] tomreyn: yes. [11:10] poc infra. [11:11] i have no first hand experience with ipxe on proliants. there's this https://community.hpe.com/t5/ProLiant-Servers-ML-DL-SL/UEFI-iPXE-and-Service-Pack-for-ProLiant-SPP/td-p/6933534 [11:11] (result of a quick web search) [11:11] weechat and links ... [11:13] http://ipxe.org/howto/romburning [11:13] tomreyn: i think i just figured out something [11:14] tomreyn: i don't need to set this up. ilo can boot the iso from the network and i just add the preseed to the iso [11:14] the only problem is, i need a centralised way to change boot order on all servers [11:15] hp ilo booting off an iso requires manual interaction, though, i guess [11:15] if the iso has preseed than it shouldn't need it [11:15] ok, good luck :) [11:15] :D [11:16] there are hp utilities / services / servers to manage ilo configurations centrally. i forgot the name, though, and i think you also need an extra license for that. [11:17] tomreyn: IPMI can change boot order too, but i tried that and it didn't yield the results i was expecting [11:17] and AFAIK ilo has some sort of API [11:17] i need to inspect [11:28] yes it does, it's ugly, though [11:29] ipmi doesn't seems like a hp concept ;-) [11:37] iirc iLO supports IPMI [11:40] gunix: no, you have to configure your pxe-server. [12:17] rbasak: hi, do you think you could review https://code.launchpad.net/~ahasenack/ubuntu/+source/ubuntu-advantage-tools/+git/ubuntu-advantage-tools/+merge/342208 today? [12:17] I'm off tomorrow (holiday), and monday is the final freeze [12:17] nacc: hi, do you think you could look at https://bugs.launchpad.net/ubuntu/+bug/1752378 today? I'm off tomorrow (holiday) [12:17] Launchpad bug 1752378 in Ubuntu "Please add Userspace Packages for NVDIMM support" [Medium,In progress] [12:18] ahasenack: I'll take a look [12:19] ahasenack: do you happen to have any bileto knowledge? [12:19] no [12:19] just that it exists [12:19] but since I couldn't use it, I didn't read about it [12:21] OK [12:26] ahasenack: stopping for lunch. I'll continue when I'm back. [12:26] sure, thx === miguel is now known as Guest14023 [12:48] i am getting error while installing wordpress.. need argent help [12:49] Reload : postfix [OK] Reload : nginx [OK] Restart : php5.6-fpm[OK] Reload : mysql [OK] Running pre-update checks, please wait... Setting up NGINX configuration [Done] Setting up webroot [Done] Downloading WordPress [Done] Setting up database Oops Something went wrong !! Calling cleanup actions ... 'ee_mysql_grant_host' [12:55] kabar: check the logs - looks like the mysql grants failed [12:56] Royk: how i check log [12:56] not sure where they may be - I've never installed wordpress from packages [12:58] ok brother [13:05] well, start in /var/log [13:05] if that output was from a php script displayed in your browser, check the apache logs in /var/log/apache [13:10] dont get it [13:16] kabar: perhaps /var/log/dpkg.log [13:17] ahasenack: that php script was probably run directly, not through apache, as installers mostly do [13:17] kabar: you may want to consider installing wordpress from source - there's a *ton* of updates that never reaches the repos [13:23] ahasenack: I don't see any tags "upstream". Which commit on Github matches the current state of Bionic please, and which is the one from which you're uploading? [13:23] rbasak: I'm releasing on github after ubuntu [13:23] rbasak: I'm treating ubuntu as the upstream [13:27] log is so big [13:29] ahasenack: I'm just after the broken down commits for the change please. Where can I find those? [13:31] there are no broken down commits in the lp branch, since there are no changes in debian/ and it's a new upstream [13:31] I see where you are getting at, though. If I'm treating ubuntu as the upstream, you will want to see every individual commit there? [13:32] there are commits/merges that do not warrant an entry in d/changelog [13:32] otherwise it would be a git log [13:43] ahasenack: I just want to review the diff, and it would be easier if broken down. [13:43] rbasak: would you do that if this were not a native package? Review each change in, say, a new apache tarball? [13:44] not do, but require [13:44] sorry [13:45] I can break it down but one commit would be like "cleanups". The others would match the d/changelog entries [13:45] ahasenack: I'm only interested in the packaging implications. For apache, that's pretty evident from the debian/ directory. For this package, though, it pretty much is all packaging so the "upstream" changes are relevant. [13:46] for example, a change like this: [13:46] -install_package_if_missing_file() { [13:46] +apt_install_package_if_missing_file() { [13:46] that's just previxing apt functions with apt_ [13:46] I wouldn't mention this in the changelog [13:46] d/changelog, that is [13:46] Yeah so if that were in its own commit, I could see that and ignore it. [13:46] Making the diff that I do need to review smaller. [13:46] I see [13:47] so I can have changes, small ones let's say, in the git commit log, that are not mentioned in d/changelog, as long as the commit log is self explanatory, is that ok? [13:53] ahasenack: I would probably say something like "Other trivial miscellaneous changes; see git log for details" in the changelog or something. [13:53] ahasenack: but then Vcs-Git really needs to point to something that has the changes. [14:02] rbasak: the vcs-git url is correct, you just won't find matching commits between that and launchpad, not even commit messages [14:02] lp is not a mirror [14:04] rbasak: ok.. so imporer.. [14:04] importer [14:05] ssh-import-id https://git.launchpad.net/ubuntu/+source/ssh-import-id/tree/ [14:05] has ssh_import_id.egg-info/ dir that i dont know where they came from. [14:05] they're not in ssh-import-id_5.5.orig.tar.gz [14:05] $ tar tvzf ../ssh-import-id_5.5.orig.tar.gz | grep egg [14:05] $ tar tvzf ../ssh-import-id_5.5.orig.tar.gz | grep egg [14:06] $ tar tvzf ../ssh-import-id_5.5.orig.tar.gz | grep egg || echo "no eggs" [14:06] no eggs [14:07] oh. i see they're in ssh-import-id_5.5-0ubuntu1.diff.gz [14:08] Oh. It's a version 1 package? [14:08] i guess, yeah :) [14:09] so how would you do this SRU ? [14:09] smoser: you don't need the z in that tar command when extracting - it's automatic [14:12] RoyK: yeah, thanks. but my fingers dont remember that. [14:12] hehe [15:13] rharper, dpb1: https://www.reddit.com/r/Ubuntu/comments/880spj/systemd_ifdown_hook_services_for_individual/ [15:29] rbasak: networkd-dispatcher is a possiblity [15:30] rharper: yeah that's why I pinged you :) [16:03] rbasak: fyi, i'll keep adding tests to my last branch [16:03] nacc: it looks like there is a race in the vm_setup script when you launched multiple of them at the same time. [16:03] nacc: thanks. It seems unlikely I'll get to reviewing it any time soon now though :( [16:04] One job was still creating a VM and then when a 2nd tried to run uvt sync it throws the libvirterror [16:04] rbasak: understood; how did you want to proceed with that, given the time crunch(ish) [16:04] i can try and get 100% coverage of all added code [16:05] (i'm hoping i can get to 100% of importer.py, tbh) [16:05] powersj: :) [16:05] nacc: I don't mind you continuing to work on it provided that it doesn't make my reviews unwieldy and big. If that's going to happen, I think you should hang back. [16:06] nacc: too much stuff pending review makes things exponentially difficult :( [16:06] And IMHO, the only reasonable way to handle that is to reduce velocity. [16:06] rbasak: yeah, just a bit frustrating in that stuff is getting pushed further and further back that is ready now [16:06] :) [16:07] rbasak: 'unwieldly' is relative. the last commit that just adds test is up to about 600 lines [16:07] but it's all tests [21:01] is there any way to export the preseed config after install, like on centos? [21:05] gunix, /var/log/installer/cdebconf/questions.dat also answers.dat... not sitting at debian now [21:06] gunix: debconf-get-selections is the same as above [21:06] but not everything is preseedable [21:06] ChmEarl: how do i transform that into a preseed config? [21:06] nacc +1 [21:07] gunix: you ... generally can't, iirc [21:07] gunix: these days, people make image snapshots [21:07] gunix: it's not typically worth the cost to re-install multiple times [21:08] nacc: what if you want to run tests on an environment and want to reinstall a bare metal server 5 times per day? [21:08] gunix: you installed by hand once? preseed from the get-go [21:09] nacc: how do i get the preseed from the get-go ? [21:09] gunix: you start with the default preseed [21:09] customize it from there [21:09] nacc: on centos you get a file called "anakonda" after the install, which can be used to kickstart all new installs [21:09] and it provides a fully automated feeling to the install [21:10] gunix: yes, i know. [21:10] gunix: ubuntu and debian do not. [21:11] :D [21:11] gunix: a lot of people do what i suggest and are able to get to the preseed they want [21:11] then they just use that, basically, forever, IMO [21:12] nacc: ok, so i should install get-go and generate a preseed with it ? [21:12] gunix: 'get-go' was a turn of phrase sorry [21:12] gunix: start with a preseed, then work from there [21:12] oh [21:14] gunix: http://blog.dustinkirkland.com/2011/03/ubuntu-server-quick-install-no.html e.g. [21:15] ok, i will try to tweak a preseed. i have multiple examples [21:15] but was trying to avoid that [21:15] but it seems to be the easiest way to do this [21:15] gunix: yeah, that's my recommendation [21:16] gunix: get one that works, first, then tweak from there [21:41] nacc: ok, thank you! [21:42] gunix: np, good luck! [21:49] Has anyone ever resized a partition in ubuntu server? [22:13] yes [22:16] akern07: gparted is quite easy to use [22:22] I was wondering how to resize a partition in ubuntu server after extending the size of my virtual machine in VM. [22:24] what filesystem are you using? [22:26] akern07: you could use 'parted resizepart ...' [22:32] I think it's the primary volume [22:59] nacc: how do i encrypt password for d-i ? [23:05] you first need to resize the partition (e.g. with 'parted'), then resize the filesystem (e.g. with 'resize2fs') [23:41] does fcgi handles sessions differently from mod_php? [23:42] reposting the response i provided in #ubuntu, for convenience: [23:42] the only thing i can think of is if nginx is loading a different php.ini file, some session related configuration variables might be different