[01:23] Hey guys [01:24] http://i.imgur.com/pm4ClLI.png [01:24] ^ that has to be a bug [01:26] There is no way thats right. Especially since lets encrypt has been out for a while and im pretty sure that doesnt happen on debian since my domain uses lets encrypt and i've never had that issue before. [01:45] Nope im wrong [01:45] people are incorrectly configuring lets encrypt on their sites [01:45] https://github.com/certbot/certbot/issues/3159 === stooj_ is now known as stooj [05:04] I noticed zul, I wanted to say that the Debian package info still points to https://github.com/lxd/nova-lxd as homepage. I can't find the Debian package files in the new repo === athairus is now known as afkthairus === zerick_ is now known as zerick [06:38] lol why is like 90% of the ubuntu server minimal install selecting locale stuff [06:38] xD [08:07] I get wsrep errors when starting up mariadb up with mariadb-server + galera packages, is there some other package required to enable galera function [08:27] anyone using kvm/qemu ? [08:31] rbasak, hey - are you around? I'm still struggling to wrap my head around this nested directory rename issue in ceph if you have time to help me think it through [08:55] does mariadb-server have galera extentions built-in ? [09:06] jamespage: o/ [09:06] rbasak, good morning! [09:08] jamespage: I would look through the implementation of dpkg-maintscript-helper. [09:09] jamespage: then try to find something that is simple but still works for your case. [09:09] It will be a hack but I think that's the best we can do. [09:16] rbasak, yah that's the best I could think of as well [09:48] hello, http://kopy.io/JyG6e is this an incorrect configuration [09:48] ? [09:52] Test it using apachectl configtest [09:58] rbasak, http://anonscm.debian.org/cgit/pkg-ceph/ceph.git/commit/?h=ubuntu-yakkety&id=b65513d8bcc4a63e91f72df1139693b36f849f04 quick second pair of eyes before I upload that to a PPA? I did somet esting with a smaller test-ceph fake package to avoid the 2 hr build lags... [10:00] * rbasak looks [10:00] bekks: how can i test that, I am lill new to apache2 serve [10:01] Place that config somewhere it gets evaluated by apache, then run the command given. [10:03] what i want to ask is do we require to say explicitly using the Listen statement at the top [10:03] Sure. [10:04] is it required ? [10:04] Yes. [10:05] You need to tell your webserver which port you want to use. [10:09] I said that in bekks [10:10] IF you want your Virtualhost to be listening on port 3000, you need that. [10:11] rbasak, its bascially the dpkg-maintscripts-helper thing adapted to deal with the directory nesting... [10:11] jamespage: what if I had modified /etc/default/ceph/ceph before to source /etc/default/ceph/foo? [10:12] jamespage: I don't see any way of fixing that, but I would want to preserve /etc/default/ceph/foo, perhaps in /etc/default/ceph.dpkg-remove/foo. [10:12] rbasak, hmm so check for anything other than ceph in the dir and abort the removal if found? [10:12] (or dpkg-old or something) [10:13] jamespage: I was a little surprised that you were renaming to /etc/default/ceph.dpkg-remove. Let me ponder this. [10:13] rbasak, tbh the likely hood of that is minimal - this only exists since xenial release, and all of the users of /etc/default/ceph source that file, not the nested one... [10:14] rbasak, so I wonder how far I really need to take dealing with the unknown here [10:14] I suspect that most users may have created their own /etc/default/ceph [10:14] hmm - now I probably should with that or maybe dpkg might do that anyway [10:14] lemme check [10:15] hmm but they can't because of the dir... [10:15] ok [10:16] jamespage: how about in the preinst you rename only /etc/default/ceph/ceph to /etc/default/ceph/ceph.dpkg-remove if the md5 matches. Then in the postinst, rename /etc/default/ceph to /etc/default/ceph.dpkg-old, rename /etc/default/ceph.dpkg-old/ceph to /etc/default/ceph if it exists, remove /etc/default/ceph/ceph.dpkg-remove if it exists, then remove /etc/default/ceph/ceph.dpkg-old if it is empty? [10:16] that's slightly neater [10:17] jamespage: I agree about dealing with the unknown - I'm just keen on preserving a customised file rather than deleting them even if a user still needs to fix up. [10:17] bekks: http://kopy.io/nLW9Y i get this error when i restart the server [10:17] after adding the line listen 3000 [10:17] jamespage: with my logic above you'll also need to fix up postrm to reverse the dpkg-remove in the subdirectory only. [10:17] ShekharReddy: So did you do what the message tells you? [10:18] ShekharReddy: se the error log [10:18] s/se/see/ [10:18] jamespage: perhaps also file a bug against dpkg-maintscript-helper asking for support of your subdirectory use case :) [10:19] jamespage: it might be worth explaining in a commoent or the changelog or something why you aren't just using dpkg-maintscript-helper. Also, do you need any consideration of version numbers for Debian? [10:19] rbasak, no [10:19] (debian version numbers) [10:19] yes - comments [10:22] RoyK: http://kopy.io/Lirlh what does this mean ??? [10:23] rbasak, reworked - http://paste.ubuntu.com/17287347/ [10:23] * rbasak looks [10:23] ShekharReddy: (98)Address already in use: AH00072: make_sock: could not bind to address [::]:3000 [10:24] (98) ?? what does this mean [10:24] ShekharReddy: something else is listening on that port - check netstat -lnop --tcp [10:24] oh wait thats not quite right... [10:24] rbasak, no that does not work because of the new incoming /etc/default/ceph file... [10:25] rbasak, I think I'll just preserve it in a different way [10:25] jamespage: ah. [10:26] jamespage: in that case, maybe rename /etc/default/ceph to dpkg-old in the preinst, then rename it back in the postrm rollback if necessary? [10:26] I guess that's what you were doing before, sorry. [10:27] http://kopy.io/phuie i don see anything listening on port 3000 RoyK [10:30] rbasak, ok take three - http://paste.ubuntu.com/17287431/ [10:31] bekks: hello, can u also look into that issue i mentioned above [10:33] ShekharReddy: Did you define another VirtualHost for port 3000? [10:33] No [10:34] I added the line Listen 3000 to already existing on e [10:34] i.e the one which i have mentioned earlier [10:35] Show us your exact config please. [10:35] rbasak, that approach will preserve foo in /etc/default/ceph during an upgrade [10:35] And show us the output of "apachectl configtest" too. [10:35] just tested that [10:37] rbasak, so close [10:37] cannot remove old backup config file '/etc/default/ceph.dpkg-old' (of '/etc/default/ceph'): Is a directory [10:37] dpkg-old confuses dpkg... [10:37] bekks: http://kopy.io/vjGyw conf file [10:37] as it should be a file or link, not a directory [10:37] I might have to use dpkg-backup [10:38] bekks: http://kopy.io/tVd1t apachectl configtest output [10:41] ShekharReddy: Either use port 3000 in line 1 or in line 2. [10:41] jamespage: ah, I didn't realise dpkg would touch it again after a rename, sorry. [10:41] jamespage: yeah, dpkg-backup should be fine [10:41] rbasak, yup - works ok [10:42] rbasak, just thinking I should actually remove it if there is noting in the directory post move [10:43] jamespage: agreed [10:44] rmdir appears to have an --ignore-fail-on-non-empty [10:44] any mysql guru's in here? [10:44] Probably not POSIX but shouldn't matter for Debian [10:44] Xin: Depends on your question. [10:45] I need to make my mysqld accessible ONLY to local network devices [10:45] Which is the default configuration. [10:45] oh yeah? [10:45] sweet [10:45] Yeah, look at the config. [10:45] surprisingly easy answer haha [10:45] double-check it with netstat [10:46] true, good tip! [10:47] hmm, nah doesn't seem to be [10:48] whats its port number by default? like 3309 or something? [10:48] Look at the config :) [10:50] I looked at everything in /etc/mysql and /etc/mysql/conf.d but there didn't seem to be anything except it pointing to a .sock file === admcleod_ is now known as admcleod [10:50] bekks: should the document root be set to myproject/public [10:50] ? [10:50] but even that file had like 20 lines [10:50] is it compulsory that a public folder be there ? [10:51] ShekharReddy: Depends on how you want to setup your webserver. [10:51] like... [10:51] http://prntscr.com/bfs9by [10:51] wat is this lol [10:51] i mean i don have the public directory in the repo [10:51] Xin: An empty config section. [10:52] bekks; the mysqld.cnf in the folder above that just includes this folder [10:52] oh hang on [10:52] yeah wtf lol [10:53] ok found it haha [10:53] that was utterly needlessly overengineered [10:54] What exactly are you talking about? [10:54] yeah default setting is localhost only as I suspected [10:54] You did not suspect it, I told it to you ;) [10:55] theres a full 3 folders of cnf files, some that just include other folders that only have empty config sections etc [10:55] for what is, in the end, like a 100 line config file [10:55] how did you tell me that? :| [10:55] Xin> I need to make my mysqld accessible ONLY to local network devices [10:55] Which is the default configuration. [10:55] local network as opposed to internet.. [10:56] not local host [10:56] xD [10:56] Configure your firewall then. [10:56] nah, found it, set the bind-address, just gotta restart hopefully [10:56] also configure your bind or listen addresses appropriately [10:56] belts and suspenders and so on [10:57] root@dev:/home/swim# service mysqld restart [10:57] Failed to restart mysqld.service: Unit mysqld.service not found. [10:57] wat is this [10:57] lool [10:57] ok its just mysql [10:57] CRISIS AVERTED [10:58] bekks: i want to configure this locally https://github.com/ShekharReddy4/atlas-mock-id which is a mock auth for an app i am gonna develop [10:58] hmm no port 3306 as per the config in my netstat [10:59] hm hm hm [10:59] the problem is when i login as said in the readme in the above link i always redirected to the same page again which is login page [10:59] Xin: Because it is using a socket. [10:59] it is capable of doing both [11:00] i mean i am kind of in a loop (login page --> i enter the credentials --> press the login and---> login page ) [11:01] guys this is irritating me from yonks [11:01] Xin: Yes. Either port or socket. [11:01] bekks: did u get my problem ? [11:01] No theres no reason it cant support ALL IPC methods simultaneously [11:01] like any db server [11:02] but, what setting would you propose I change except the bind-address [11:02] Xin: Well, if you know better, I am out. [11:02] there doesnt seem to be a lot here [11:03] I think mysql lets you add new lines as needed, right? [11:04] sure [11:04] sarnold: can u also please look into my issue ? [11:05] I just added an iptables rule to allow it [11:06] but that wouldnt have been the issue anyway [11:07] ShekharReddy: sorry, I've never heard of atlas before [11:07] hey atlas is my project, i feel it is something related to config [11:07] rbasak, I'll do a little conditional check first prior to removal [11:08] ok uploaded to ppa for xenial testing -thanks for the reviews... [11:08] oh dope you can use % for chunks of the ip addr [11:09] seems u din get my problem sarnold , i will explain it again, i setup the server locally for the above mentioned project [11:09] and when i ran it, is working fine , but i am unable to move out of the login page [11:09] stilllll no entry in netstat for mysql OR 3306 [11:09] hm hm hm [11:11] ohh the recommended way is apparently to ssh in [11:11] I wonder how that works with php etc.. [11:16] oh, if I try to connect to it [11:16] I get; ERROR 1130 (HY000): Host '10.0.0.3' is not allowed to connect to this MySQL server [11:16] hrm [11:17] not sure why it wouldnt show up in netstat though [11:18] 10.0.0.3 isnt localhost [11:18] thats the vm I just tested it from [11:18] :P [11:18] Can you pastebin your config then? [11:18] I can, but all I have done as of right now is comment out the bind-address line [11:18] which supposedly makes it bind to all interfaces [11:19] And which port? [11:19] 3306 [11:19] default [11:19] The default is a socket ;) [11:19] from that error though [11:19] it is contacting the server [11:19] its just not allowed [11:20] so I need to use some form of GRANT line [11:20] which I have nfi about [11:27] apparently I cant do CREATE USER 'username'@* [11:27] CREATE USER 'username'@'*' seemed to work [11:27] How can I disable mdadm? I get "W: mdadm: /etc/mdadm/mdadm.conf defines no array". mdadm is installed, since the ubuntu-server packages depends on it. 16.04 [11:28] I can only guess thats the correct syntax [11:29] nah that didnt work either [11:29] hmm [11:29] this is starting to annoy me [11:33] how can we change a file mode in a project [11:33] does that effect the installation of the project in anyway === sveinse_ is now known as sveinse [12:02] ShekharReddy: ? [12:03] bekks: i am unable to get out of that login page [12:03] ShekharReddy: So check your logs. [12:09] bekks: http://kopy.io/K106E [12:10] bekks, what is undefined index [12:10] ? [12:11] A not defined index. [12:12] can u provide a solution or some guidance on how to resolve it [12:12] bekks: ^^ [12:13] No. I have no clue about which project you are using there nor how you configured it upon installation. [12:16] coreycb, ddellav: I think the b1 announce is good to go... [12:17] bekks: atleast ideally what is the cause of this error ?? [12:17] ShekharReddy: I have no clue about which project you are using there nor how you configured it upon installation. [12:18] ok [12:41] http://kopy.io/1iVsR error ?? any hacks on how to resolve this [12:43] are there errors? I see three notices and one warning [13:08] sarnold: I think it's a hopeless case, they are assuming we would be doing their job of configuring Apache, MySQL and PHP [13:08] dasjoe: all things I've gone out of my way to avoid using.. :) [13:36] http://kopy.io/i6cDs facing this error, any help ? [13:38] ShekharReddy: What did you do before? [13:38] removed the mcrypt from php.ini [13:38] Why did you remove it? [13:39] i get a warning, googling made me remove it http://stackoverflow.com/questions/21923534/php-warning-module-mcrypt-already-loaded [13:40] bekks: ^ [13:40] mycrypt...ugh [13:40] roll on php 7.1 [13:40] So what exactly did you remove? [13:40] in php.ini i removed the extension [13:41] extension = mcrypt.so something like that [13:42] ShekharReddy: please explain why you think that is an error. [13:45] it says module mcrypt is already loaded so i thought that might be the issue of redundancy [13:45] ShekharReddy: Did you remove something else? [13:46] sarnold: ^ [13:46] bekks: no [13:47] i added a line in the code but that is related to unidentified index [14:03] hey peeps I asked in #ubuntu hours ago but we didnt really get anywhere [14:03] I have an ubuntu server box and a xubuntu box [14:03] the xubuntu box can ssh into the server box [14:03] BUT only using its IP [14:03] So fix your DNS. [14:04] BUT, the server box pings fine [14:04] nah dns works on every other box, and resolves fine for pings [14:04] Obviously it doesnt work that fine. [14:04] ohh unlesssss [14:04] iirc ping includes its own resolver code and doesn't use the nss resolver [14:04] I installed a dns server on my ubuntu boxxx [14:04] how do I uninstall that [14:05] How did you install it? [14:05] marked "DNS Server" in the net installer [14:06] Check if bind is installed, uninstall it. [14:06] no bind :( [14:07] So check the services running, and uninstall the DNS server software. [14:08] try dpkg -l bind9 [14:08] sarnold; nice work [14:09] rebooting and hopefully thats all good [14:09] cant believe I didnt remember that before [14:09] No reboot needed. [14:10] cant hurt :P [14:10] It kills kittens. [14:10] ohshiiii [14:10] I DIDNT SIGN UP FOR THIS [14:11] If I made an OS [14:11] id call it Putin [14:11] Computin the putin bootin [17:00] does anyone know how to do ip aliasing on the new ubuntu server 16.04? [17:00] as in more than one address per interface? [17:00] yes [17:00] I've tried the old method with the :0 or :1 [17:01] but it's not working [17:01] in /etc/network/interfaces [17:01] just add multiple address lines to the interface def [17:04] so it would look like: [17:04] iface eth0 inet static [17:04] address x.x.x.x [17:04] address x.x.x.x [17:04] ? [17:04] yeah [17:05] tried that it's not happy with me [17:05] huh [17:05] so I have: iface enp3s0 inet static [17:05] doesn't work for me either [17:05] I don't have all my good prompts with me right now to cut and paste from our production machines right now [17:05] I'm used to add additional IPs with a "post-up" line [17:07] and it doesn't show the additional address aliases when I type in ifconfig a- [17:07] https://wiki.debian.org/NetworkConfiguration#Multiple_IP_addresses_on_one_Interface [17:08] multiple iface defs [17:08] but not using the :'s right? [17:08] no, not the : [17:08] so I have tried : auto enp3s0 [17:09] iface enp3s0 inet static [17:09] address x.x.x.x [17:09] iface enp3s0 inet static [17:09] address x.x.x.x [17:09] and it says nope [17:09] try adding "up ip addr add x.x.x.x/mask" at the end [17:10] and btw iproute2 isn't "new" it's 10+ years old [17:10] yeah, but people are still using :1 and that's ifconfig syntax [17:10] well the last time I configured aliases, I did it with the :'s and it worked [17:10] and I was happy [17:11] dr4c4n: and now you're dreadfully unhappy because someone wrote a better interface? ;) [17:12] better is questionable [17:12] and I don't like change [17:12] just piss off, ok? we don't need complaints about what's better or worse [17:13] thanks for their help [17:13] doing it in an up line is a hack, I think setting an address and a netmask is supposed to work [17:13] jelly: yeah, it's supposed to work, but I don't have access to any of my xenial machines right now to test, so, well.. [17:13] jelly: it's not a hack, it's the new way [17:14] RoyK: it's a hack that existed before ifupdown 0.7 [17:14] jelly: sometimes things improve, like going from LILO to Grub to Grub2 etc [17:14] (0.7 is when ifupdown switched to iproute as backend) [17:14] * RoyK installs minix on jelly's box [17:14] calling an arbitrary command is not improvement [17:15] does anyone know where the documentation for iproute2 is? [17:15] * RoyK gives up [17:15] yeah, but configuring multiple addresses without fake-naming the interfaces :n is an improvement [17:15] dr4c4n: that debian wiki page is the best one I've found [17:16] dr4c4n: you don't want iproute docs, you want ifupdown docs, ie. interfaces(5), to see what syntax is actually supported before resorting to calling ip manually [17:17] thanks a lot jelly [17:18] RoyK: I've had a for loop with ip a ad... in an up line back in 2003 on... sarge, maybe? Yeah it's a hack, and not a new one [17:20] perhaps we should create a registry like windows with all the configuration stuffed into a database and make jelly and all happy [17:21] maswan: link? [17:22] dr4c4n: https://wiki.debian.org/NetworkConfiguration[6~ [17:22] dr4c4n: https://wiki.debian.org/NetworkConfiguration [17:22] RoyK: I'm not sure you're getting the point, {pre-,post-,}{up,down} are workarounds by definition, being able to run an arbitrary command is left as a fallback option when you can't configure something any other way [17:24] maswan and jelly: I found this as a really good reference, but my question is: how do I make the ip address add permanent other than modifying the /etc/network/interfaces file? [17:25] no other way [17:26] well, I'm lying a bit there. Technically, there's systemd-networkd, and there's NetworkManager too and its nmcli, but the support you're going to get in ubuntu 16.04 for those may vary. [17:27] picking any of those basically means abandoning /e/n/interfaces [17:28] is there a legacy mode for keeping the ifconfig method because the manual approach is retarded [17:28] well, you *could* write your own custom init script that does an "ip addr add" command once the interface is up, to make it permanent. but that seems an unnecessarily complicated way of diong an up-hook. [17:28] why change what was working??!?!?! [17:29] dr4c4n: the :n stuff didn't really work, it never worked for ipv6 addresses for instance [17:29] dr4c4n: but if multiple instances with the same name doesn't work, like it shoudl from the debian wiki instructions, I think that's a bug. [17:29] dr4c4n: the ifconfig method was never permanent, either [17:33] so i followed the debian wiki example exactly. [17:33] and I get cannot assign requested address [17:34] dr4c4n: show your work: pastebin the complete file, and pastebin the output of "ip a" after boot or after "ifdown someiface; ifup someiface" [17:34] jamespage, ddellav: for my ci shift today I pushed cinder updates for newton, uploaded python-oslo.policy 1.9.0 to yakkety (fixes keystone), and uploaded python-oslo.concurrency 3.7.1 to mitaka (sru fixes nova/keystone failures) [17:34] how to shutdown ubuntu server without having to login if I decided to do so at the login terminal [17:35] it's seems very stupid to login just shutdown the server [17:35] :/ [17:36] LJHSLDJHSDLJH: personally I don't like unauthed personnel to be able to shut down a server [17:37] such a person offered a login terminal can unplug the power the cables [17:37] LJHSLDJHSDLJH: if you can send an ACPI signal to the machine it should obey to it [17:37] that should be simple [17:37] just make a poweroff user that runs poweroff :) [17:38] or, why not just have it obay the power button? [17:38] one press should start the poweroff scripts [17:39] patdk-wk: yeah, the power button should send the proper ACPI signal [17:39] jelly: [17:39] http://pastebin.com/xdb6KmsA [17:39] ddellav, is bug 1587589 tested? [17:39] bug 1587589 in keystone (Ubuntu Wily) "[SRU] liberty point releases" [Undecided,Fix released] https://launchpad.net/bugs/1587589 [17:39] would be nice to release that [17:44] dr4c4n: you shouldn't have both auto and allow-hotplug for the same interface [17:45] dr4c4n: forgot the output of "ip a" after attempting to configure [17:46] dr4c4n: I'd remove or comment out the allow-hotplug line [17:46] jelly: ip a shows that I have .201 and .202 [17:46] I will remove that line [17:46] thanks :) [17:46] says 201/8 and 201/24 and a 202/8 [17:47] remove the /8 manually before trying again, or reboot [17:48] * jelly runs away [17:48] thanks === afkthairus is now known as athairus [18:23] hi, new install and user here. doing "service radvd start" appears to start it and then exits. radvd appears to work in non-daemon mode with /etc/radvd.conf how do I figure out what's causing this? [18:24] sudo systemctl list-units|grep radvd radvd.service loaded active exited LSB: Router Advertising Daemon [18:47] coreycb sorry, yes, it's tested and passed no problem. [18:47] ddellav, ok [18:49] ddellav, in that case can you request a promotion to liberty-updates from beisner or jamespage? then we'll be done on our end with that task. [18:51] coreycb beisner jamespage can one of you guys promote ceilometer and keystone point releases to liberty-updates. The updates have been tested successfully. [19:03] Strykar: does journaltl or systemctl info tell you more? [19:03] *journalctl [20:08] thanks sarnold for your help earlier [20:08] and bekks [20:08] it was the dns server [20:08] all working now, with nat and bridging [20:08] yay [20:51] hello, im trying to provision a VM using packer with a preseed file, and ive tried various ways, but i have no idea how to set a hostname properly [20:51] my host gets a weird ip-192-168-1-X as hostname === prince is now known as Guest95291 [21:25] d-i netcfg/get_hostname string $myhostname [21:25] in your preseed file [21:29] nacc: but that seems to be ignored if dhcp is used [21:30] Ergo^: your resulting VM, does it have a /etc/network/interfaces file? what is in it, if so (paste.ubuntu.com)? [21:33] nacc: https://friendpaste.com/45Az0a7zF1WYTCHqCwLE7u [21:36] nacc: is it possible to set this on boot at kernel level? [21:37] I tried setting hostname=foo-bar but it seems to be ignored [21:39] i have in my preseed [21:39] d-i netcfg/get_hostname string unassigned-hostname [21:39] d-i netcfg/get_domain string unassigned-domain [21:39] d-i netcfg/hostname string foo-bar [21:40] everything seems to be completly ignored [21:42] Ergo^: are there any files in /etc/network/interfaces.d/ ? [21:43] Ergo^: well, you can pass the above preseed value during the boot to the installer [21:47] nacc: those valies i pasted ARE from my preseed === Guest95291 is now known as prince [22:34] Ergo^: is your VM recognizing the other values in your preseed file? i don't really konw what packer is [22:34]