[00:37] nacc: looks like it's stuck on thunderbird. Not sure how long for [00:38] It's on 38.3.0+build1 currently [00:38] (doesn't appear hung, just chugging through it) [00:39] Possibly since the 17th. [02:04] rbasak: yeah, i just blacklisted thunderbird [02:04] rbasak: fyi, i've also been removed from usd-import-team -- you and dpb1 can decide if i should be readded [02:08] nacc: fixed [02:08] hi there [02:11] nacc: I wonder if thunderbird will actually finish [02:11] So I've been hesitant to kill it [02:12] rbasak: sure, it will be an interesting test of a huge repo [02:12] orig tarballs are massive [02:12] dpb1: thanks [02:13] It's on ~#140 now, of 792, working backwards. [02:13] So it's most of the way there. [02:14] rbasak: cool [02:14] rbasak: if it does finish, we should look at final repo size to determine if we should remove from blacklist, i think? [02:16] If it finishes, it will push, and after that it'll grow only incrementally. So will it matter? [02:16] rbasak: every import has to clone it [02:16] rbasak: which implies pulling down GBs of data for one package [02:16] It's 31G currently [02:16] Good point [02:16] interesting [02:17] 45G remaining [02:17] and we wanted to sync with cjwatson the impact of these huge packages [02:17] (of free space) [02:53] rbasak: yeah that seems about right [05:21] good morning [10:27] nacc: would you like to continue working with me on landing these outstanding MPs, or would you prefer to hand them over to me to get them landed? I'm happy either way, just want to be sure that my expectations match your intentions. === unreal_ is now known as unreal [11:47] dosaboy: jamespage: keepalived 1:1.3.9-1build1~cloud0 is in queens-proposed now [11:59] hi [11:59] I can send messages from my mailserver to gmail.com and accept [11:59] using roundcub [11:59] it seems dovecot works [12:00] how can I put my server to thunderbird? [12:00] no there each linux user from /etc/passwd accosiated with mail user [12:08] error? http://pix.toile-libre.org/?img=1524571581.png [12:08] What data there should be? [12:08] what is my IMAP, what is my smtp ? [12:17] who can test my settings? [12:18] telnet mail.kselax.ru 25 [12:18] show nohitng and [12:18] telnet mail.kselax.ru 143 [12:18] works [12:18] is it right? [12:45] Hi all. I'm looking at trying to build custom cloud images (i.e., the same user experience, but a larger set of pre-installed packages). Does anybody know if the information on this page (https://help.ubuntu.com/community/UEC/Images) is still up-to-date? [12:46] Is this how the images on cloud-images.ubuntu.com are built? [13:08] In 16.04 how can I set-up two network interfaces so that one uses dhcp and the other one has a static IP? (This is a guest in virtualbox). [13:09] lorenzosu: Fairly easy. [13:09] /etc/network/interfaces, just add a second iface stanza appropriately. [13:10] hyperlumic: meaning both interfaces need to be set to static? [13:10] lorenzosu: No, why would they need to be [13:12] hyperlumic: https://paste.ubuntu.com/p/y2BvY2mB2Z/ [13:12] hyperlumic: with that enp0s8 works correctly but enp0s3 (which is actually NAT in virtualbox) can't connect to the internet... [13:13] I will never understand why people ever specify the broadcast address... [13:14] So, DHCP provides a gateway, typically. I'd check your vbox configuration to make sure that it's not trying to hand out a gateway if you're expecting to route internet traffic over the other interface. [13:14] And I'd also check that vbox is actually using .1 as its gateway "interface" address. [13:15] hyperlumic: I don't understand what you're saying sorry [13:15] Then I'm not sure how to help you further. [13:15] entries under iface (address, netmask, etc...) need to be indented iirc [13:15] hyperlumic: if I set both to dhcp everything works as expected, apart the fact that I want enp0s8 to have a static IP [13:15] blackflow: Don't think it matters. [13:15] blackflow: I could be wrong. [13:16] blackflow: so that would mean I'd have to explicitly set the other interface to static and specify parameters for that as well? [13:16] I could be wrong too, but I remember something about that being important [13:16] lorenzosu: no, what I meant is hit a tab before each address, netmask, gateway and you don't need broadcast. [13:16] You don't even need netmask. [13:17] 192.168.x.x/24 [13:17] Use CIDR format in the address. [13:17] yeah, should be sufficient. [13:17] blackflow: you mean indenting the addresses makes them 'local' to only that interface? [13:18] What? [13:18] yeah the manpage says indentation is not required, so I was wrong. [13:18] mhh [13:19] Was switching into Python mindset, but that sounded weird lol [13:19] (I'm a python dev, so probably that's why I remembered something about that being important for interfaces :) ) [13:19] lol [13:20] so anyway, are you sure that's the correct subnet for enp0s8 ? [13:21] wait, you said enp0s3 didn't work, the dhcp one? [13:21] blackflow: yes, the problem isn't enp0s8 (I have ssh'd into it etc.) [13:21] blackflow: correct [13:22] blackflow: I think enp0s3 is trying to use the gateway set for enp0s8 [13:22] or other parameters. [13:22] :| [13:23] well what does netstat -rn show? [13:24] Don't use netstat. [13:24] Long deprecated. See also ifconfig, etc. [13:24] alright then ip route show [13:24] 'ss' is the correct command. [13:24] Oh, sorry, routing. [13:24] ss is sock stat, not routing [13:24] ip route, yeah. [13:25] Not enough coffee yet... [13:26] and here's the thing, the two NIC setup can't work with two (default) gateways. dhcp sets one, as hyperlumic said, so it should be dropped from the static config. dhcp + static config is usually for a default/public/WAN + LAN interfaces, so depending on which is which, the default one should have the gateway [13:27] blackflow: https://paste.ubuntu.com/p/ZFzbtQKq2Z/ [13:28] from what I understand both interfaces seem to try and use 192.168.56.1 [13:28] I think [13:29] of course. that's the meaning of "default" gateway. Meaning, the kernel will route packets for 10.0.2.* via enp0s3, and packets for 192.168.56.* via enp0s3, and all other packets via gateway IP and NIC which belongs to the subnet of it. [13:30] in your case, enp0s8. [13:30] so note that it's not the NIC using the gatway, it's the kernel when routing packets. [13:31] mhh ok so... any idea on how to amend this? [13:31] :) [13:34] lorenzosu: what exactly do you want to achieve? [13:35] blackflow: idea was to have enp0s3 all automatic with dhcp, and enp0s8 with a static IP. This is a virtual machine, so it makes sense to try and keep the host-only network interface on the same IP so as to access to the database, server, ssh, etc. [13:41] hyperlumic: or just "ip r" ;) [13:41] RoyK: s'what I do. [13:42] lorenzosu: in that case don't specify the gateway for enp0s8 and let dhcp set it, so basically just address 192.168.56.102/24 for the iface stanza [13:42] does the bot have any "network basics" links? seems they are needed all the time in here [13:48] blackflow: that worked, thanks! [13:48] Essentially now I have this: https://paste.ubuntu.com/p/cNmkvBg5qz/ [13:49] And this: https://paste.ubuntu.com/p/WCsgTDxcqy/ [13:56] rbasak: i'm happy to work with you, pending the CLA, i guess -- so if there are small changes, maybe you can just do them? [13:56] rbasak: also not sure on my time availability [13:56] OK, thanks [13:58] lorenzosu: yup, that's it. [13:58] blackflow: thanks [14:06] who know what the error? wher auth? https://paste.ubuntu.com/p/h4XW3YP2c3/ [14:07] Neo4: possibly only available after STARTTLS? [14:07] Though I don't know if ESMTP/STARTTLS allows that. [14:07] Just speculating. [14:25] rbasak: see Ive got new error http://pix.toile-libre.org/?img=1524579856.png [14:26] now servers are accessible and pasword or login not accessible, when I do authentication to roundcube I use only login [14:26] neo and password [14:26] now maybe need something change for I use mail instead login? [14:31] Neo4: you might find "swaks" helpful to debug. [14:32] Neo4: I'd focus on getting swaks working first, and then move on to other frontends. [14:35] I moved my mdadm raid1 to a new system, but it looks like it is in a degraded state with sdb missing: https://pastebin.ca/4018147 [14:35] if I do fdisk -l, sdb does inded show up [14:36] should I just mark it as failed, wipe sdb, and re-add it to the array? I'm currently booting from it. [14:59] see this https://paste.ubuntu.com/p/86KFtf8ZrV/ [14:59] how to make authentication? [15:00] there test@kselax.ru and test pasword user [15:00] base64? [15:10] Ive got this https://paste.ubuntu.com/p/CVfMXtwNGT/ [15:11] hash was made from row '\000test\000test' [15:11] What does ti means? It means my SASL not use domain and I got error? [15:33] who can add this mail to your mail clietn [15:33] mail: test@ksela.xru [15:33] pass: test [15:33] and say me what is wrong with it? [15:35] smtp: mail.kselax.ru port 587 [15:35] imap mail.kselax.ru port 143 [15:35] rbasak: i can cleanup the merge-base branch and land it. do you want to work on the tmpdir fixes? [15:37] dpb1: ahasenack: --^ fyi, i think that means ubuntu merge will work soonish [15:41] is MAAS something like Ansible or Puppet? [15:44] dpb1: although i probably need to wait for an answer on the CLA [15:58] nacc: I can, but I was intending to look at the importer test/fix branches again first. What order do you think would be best? [15:58] rbasak: i think that's fine, the importer cleanup side is mostly annoying when a big package fails (e.g., it's in our current failure list). I think that should be relatively minimal if you update your importer snap to the latest (which has thunderbird blacklisted) [15:59] OK [15:59] rbasak: otherwise, those failures will fill up your disk quickly, if they happen (or if one does, then it affects others, which also fill up the disk, etc.) [15:59] Good to know, thanks [15:59] The thunderbird import is still running. [16:00] It's had maybe an hour doing "git gc" after the patches unapplied import. At least that's where I think it's up to. [16:00] 31G. [16:00] After that it'll be patches applied and a push presumably. [16:00] rbasak: alternatively, if you do happen to be watching and see failures, you can kill the large imports and manually clean up the tempdirs [16:00] rbasak: the `git gc` is manually or git auto-performed? if you can tell [16:00] Looks manual to me. [16:01] rbasak: ok, it could be done then [16:01] we only gc at the end of the import [16:01] At the end of patches unapplied you mean? [16:01] rbasak: the end of the full import [16:02] importer.py:378 in master [16:02] I was watching the ps output, and saw each dpkg-source -x (since that was taking a while) followed by the git add -A (-f?) which I think is the dsc to tree hash function [16:02] rbasak: both unapplied and applied use that now [16:02] The dpkg-source was giving away the dsc filename [16:02] Oh, OK [16:02] rbasak: so it depends on if you see the --skip-patches or not [16:03] I don't recall. [16:03] But now I look, applied tags are present. [16:03] if you do, unapplied, if you don't, applied [16:03] so yeah, it's presumably done [16:03] and is GCing a large repo :) [16:03] That's good news, thanks. [16:40] nacc: it's pushing :) [16:40] rbasak: nice! [16:44] hi nacc, good news [16:44] rbasak: so (presuming i find some time today), you're ok with me landing the merge fix + your cleanups in master? [16:45] rbasak: also, from a release manager perspective, i think i'm assuming you'll cut an 0.8 and force push it to all branches (master will be FF) after we land the importer changes [16:59] nacc: +1 to fix the docstring and land the merge fix in master [16:59] Yeah, agreed on the 0.8. [17:01] rbasak: great, thanks [17:34] rbasak: done (merge fix) [17:45] Thanks! [18:06] rbasak: lol, thinko on our part [18:06] rbasak: we've got people.canonical.com/~nacc in our snapcraft yaml [18:06] rbasak: can you regen those tarballs, put them in your people.c.c space and update the yaml in master? [18:07] rbasak: i believe each tarballs instructions are in the yaml explicitly [18:41] what is difference between: [18:41] $ hostname [18:41] and [18:41] $ hostname -f [18:41] ??? [18:41] it show diffrerence output [18:42] check the manpage :) [18:42] hostname -f shows [18:42] yes I guessed [18:42] -f FQDM [18:43] I don't know how there that certificate appeared https://ssl-tools.net/mailservers/kselax.ru [18:43] I didn't generate it [18:44] oh that's a cool site, thanks [19:21] nacc: ack [19:33] rbasak: thanks, i think CI and store builds will be broken w/o that [21:20] anyone familar with the current state of ubuntu jeos? [21:21] granttrec: uh, it doesn't exist? [21:22] lol, rip, did something replace it? [21:22] dunno what was it? [21:22] "just enough operating system" [21:22] for a few months ten years ago that was the name given to netboot images [21:22] stripped down ubuntu server for virutlization [21:22] maybe it was replaced by ubuntu core, which then got replaced by a different ubuntu core which then got replaced by snappy? :) [21:23] granttrec: maybe one of the generic cloud images would suffice [21:23] sarnold: can I get a link? new to this stuff [21:25] there's also minimal now [21:25] iirc [21:25] nacc: now == 18.04 ? [21:27] sorry, went off to talk with furnace repair guy .. [21:27] granttrec: images are published here http://cloud-images.ubuntu.com/ [21:28] sarnold: not sure, but i think so [21:28] granttrec: there's a "simplestreams" tool of some sort that knows how to parse and verify json data that describes them [21:28] smoser: --^ perhaps you know better [21:28] sarnold: thanks [21:33] granttrec: .. and back again, hah [21:34] granttrec: there's a uvtool or something similar from the server team that import the images into libvirt and boot the things without much effort; a little bit of time with cloud-init to customize the installs might be worth while too [21:36] sarnold: thanks for the advice so far, my goals are more along the lines of setting up a sandboxed dev enviroment for myself that I can nuke whenever I feel like it, a vm seems like a bit of over kill, I just want to create something simple and easy to set up [21:37] granttrec: ah! then perhaps just perhaps lxd would be a better fit? :) [21:37] amen to that ^ [21:39] sarnold: I was thinking that, but not sure how that would go since I might want multiple projects in one place, and it seems containers are best suited for seperate specific projects, but I'm looking into it, this is for hobby devel btw [21:40] hobby, job, job, hobby ... pretty thin line in the open source world :) hehe [21:43] sarnold: lol :) this idea started since I want a little peace of mind when following setup steps `curl | sh` and easy un-install when i'm no longer interested in a project. [21:44] heh, I'm way too paranoid to ever blindly do the curl | sh thing even on a VM. there's just too many VM escapes and terminal escapes and ssh channels terrify me and and and .. [21:44] I wind up reading so many shell scripts .. [21:46] haha aint that the truth [21:46] to be fair even apt-get has a veeeery long chain of trust that you implicitly trust [21:46] irwiss: knowing half the people involved with that chain of trust helps immensely :D [21:47] I realize that doesn't scale real well.. [21:47] you need just 1 spoon of poop to spoil a barrel of honey [21:48] ewwwwwwwww [21:53] granttrec: hey [21:53] dpb1: sup [21:53] granttrec: the stripped down ubuntu is called 'minimal server image' [21:55] i'm aware but ubuntu wiki indicated jeos purpose was for virutlization [21:56] currently reading up on lxd to get what I want [21:58] if you want to try with lxd, you can: [21:58] lxc remote add ubuntu-minimal-daily https://cloud-images.ubuntu.com/minimal/daily --protocol=simplestreams [21:59] then lxc launch ubuntu-minimal-daily:bionic [21:59] we haven't published a lot of instructions yet since bionic isn't released yet, but will be doing so in the coming days/weeks [22:01] dpb1: thanks for the advice === devil is now known as Guest92578 [22:05] also, one last thing, since you are interested in virt, and not just containers. https://community.ubuntu.com/t/beta-release-multipass/2696 That's intended to make it dead simple to launch vms with updated images. [22:27] apt-get trust: one key thing is that there's a single respository as a sync point. You're in the same boat as everyone else, and others care enough to examine that boat. [22:28] And a rather large set of Debian and Ubuntu developers will fix the boat if a serious problem is reported to them, as opposed to what might be a much smaller upstream or a more focused upstream team who have other priorities. === led_ir23 is now known as led_ir22