=== markthomas is now known as markthomas|away === Lcawte|Away is now known as Lcawte [07:21] Good morning. [08:14] teward: I would "Won't Fix" bug 1194074, but up to you. [08:14] bug 1194074 in nginx (Ubuntu) "Default index.html blindly overwritten" [Medium,Triaged] https://launchpad.net/bugs/1194074 [08:52] hey all [08:54] rbasak: a package overwriting user files doesn't seem like a sane Won't Fix to me :P [08:55] So I'm trying to allow a user to access CIFS mounts. This is my FSTAB http://paste.ubuntu.com/11922616/ these are the mounts http://paste.ubuntu.com/11922620/ [08:56] the user zachary who is part of the "mediashare" group CAN read and acess the mounts, good [08:56] however, the service/application subsonic, using "subsonic" user, ALSO part of the "mediashare" group, CANNOT access the mounts for some reason with permission denied. [08:56] RoyK: packages *own* files in /usr/share. They are supposed to overwrite them on update. Users are not supposed to change package-shipped system files in /usr and expect them to not be overwritten on update. [08:57] RoyK: nginx shouldn't default to using /usr/share/nginx/www/index.html IMHO, but that's a Won't Fix in Debian. So either we should decide to diverge in Ubuntu, or decide not to, but either way make a decision. === G_ is now known as G [09:12] rbasak: IC === masACC is now known as maswan [09:22] hello? [10:42] hi anyone around here that could explain metro-ethernet and MPLS to me? [10:48] sarnold, yo [10:51] metro-ethernet is just that, they just handoff ethernet uplink to you [10:51] mpls is a vpn type service, normally used for faster switching to get from one location to another [10:52] could you explain a bit more patdk? [10:52] but you don't want to pay for a private dark fiber [10:52] you want me to explain what ethernet is? 802.3? [10:53] i do knowwhat etherent is, i just wondered if you could eloborate more on metro-ethernet and MPLS [10:53] well, if you know what ethernet is [10:53] you know what metro-ethernet is [10:53] it just means they give you an ethernet cable to plug into [10:53] not dsl, cable, t1, ... [10:54] well if whoulmed know that i wouldn't have ask, i presu [10:55] I guess more accurately, mpls is more like a vlan on the telephone network [10:55] The wikipedia page says metro-ethernet is ethernet for a MAN. [10:56] Whereas ethernet is used for a WAN or LAN. [10:58] yes? [10:58] I don't see why people need to bring up wan/lan/man/... into it [10:58] does it matter if your network is in your house/lan, or outside it/wan [10:58] it's still ethernet [10:59] metro-ethernet is just that, ethernet is provided as your upstream [10:59] vs dsl, that is not ethernet [10:59] and you need a device to convert and use it [10:59] "Metro-Ethernet" is more a branding label; the CPE delivery is Ethernet but the underlying network technology can be almost anything, including Ethernet over MPLS over Ethernet [10:59] and since it's not directly ethernet, you have no possibility to directly connect multible locations as a single l2 zone [10:59] like pure ethernet would be able to do === JanC_ is now known as JanC [11:33] kickinz1|afk: any news yet on whether the docker 1.7 backport to trusty will need a golang toolchain backport? === kickinz1|afk is now known as kickinz1 [11:35] rbasak, on the Dockerfile used to build docker it is till using go-1.4.2 [11:36] kickinz1: so does that mean that it looks like we'll be OK, or that we have a problem? [11:37] rbasak, but there is an upstream bug that can be problematic for now, so I post-poned it for now. This bug fixed I would say we should be OK. [11:37] kickinz1: OK. What's the bug reference please? [11:41] rbasak, https://github.com/docker/docker/issues/14160, I encountered it on snappy, and apparently other people on trusty, I didn't had it on vivid. But seems not an easy one. [11:41] Thanks [12:40] can Ubuntu Server for ARM be installed on Banana Pi? [12:42] Fyr, why not [12:43] ok, where do I find the review? [12:43] review ? [12:43] google doesn't show anything about it. [12:43] yes, I want to read the manual and see screenshots. [12:44] the Internet is full of Fedora ARM and Bananian, but there is no entry for Ubuntu Server for ARM on Banana Pi. [12:44] well, you install it like any arm board ... set up bootloader and kernel yourself, use debootstrap to bootstrap a rootfs and put it in place on SD/USB/whatever [12:45] ogra_, where do I read the manual? [12:45] Is there a good way to see which runlevel starts apache? [12:45] i think there are pre-made ubuntu snappy images for the bananapi as well [12:45] for Fedora ARM I use just fedora-arm-installer and an SD card. [12:45] jrtappers, debian based systems do nt use runlevels [12:45] *not [12:46] (well, they do, but all of them are identical) [12:46] ogra_, Is there a best way to guarantee running a command before a service starts each boot? [12:47] jrtappers, depends on your version ... with upstart based releases you can just create an upstart job with somethin like: "start on starting apache" ... that will exec whatever you put in it before apache is started [12:48] for systemd you need to likely do it differently [12:48] DISTRIB_DESCRIPTION="Ubuntu 14.04.1 LTS" [12:48] thats upstart then ... take a look at the upstart cookbook [12:48] Ah, [12:48] Thanks [12:57] looking for some suggested approaches to this: hips://bugs.launchpad.net/ubuntu/+source/nginx/+bug/1194074 Debian believes that this is "Won't Fix", I'm not sure how to approach it, this has always been a problem, where users just expect things to 'work' and don't take the time to protect their data [12:57] s/users/endusers and novice sysadmins/ [12:58] I think we need to either diverge in Ubuntu or decide not to diverge and mark it Won't Fix in Ubuntu. I can't think of any other solution that'll be effective. [12:58] rbasak: nor can I. I hate to say "Hey, Sysadmins, learn proper administration of your servers, for once" but I can't see a method to approach/fix that [12:59] What Apache does seems to work much better IMHO [12:59] rbasak: i was about to say, what does Apache do lol [12:59] Are you asking? [12:59] mhm [12:59] It creates /var/www/html/index.html (formerly /var/www) and sets the default path to that [13:00] Users change stuff in /var/www/html. Package upgrades leave it alone. [13:00] This mirrors what for example MySQL does with /var/lib/mysql [13:00] perhaps we should take a page from Apache and diverge, then [13:00] The only difference being that the sysadmin arranges to change /var/www directly, rather than going via the app like changing /var/lib/mysql [13:01] I think diverging would be reasonable here, but maintaining the delta could be a pain, as it involves upgrade path considerations for the future [13:01] mmm [13:01] (as well as a conffile change, so maybe even upgrade path from the past) [13:01] So it's quite a bit to take on in a delta, so I think it's also reasonable to say that we can't take it on. [13:02] mmm [13:02] Separately, you could push in Debian, but that'd mean being involved in a long debate. [13:04] rbasak: i think we're at the point where we may need to take it up to Debian, but we're going to get pushback again. I can try and go over the last maintainer's head to comment on this, take it to nginx upstream and have them chastise Debian [13:05] 'cause I know they won't like it if nginx upstream ends up saying "Listen, you're doing it wrong, fix it" [13:15] teward: if it were me then I'd try and openly re-open the debate in Debian, but with more clarity and evidence. [13:16] teward: but it's not me, so up to you :) [13:16] teward: I still think just leaving it "Won't Fix", even if for the moment, is fine. [13:16] teward: leaving the bug open does set an expectation of "patches accepted, may be fixed soon", which is false here I think. [13:26] rbasak: i just tried this - i'll PM you the response I got [13:33] rbasak: i'm making a decision on this - we're going to diverge this for Wily, or at least X [13:36] rbasak: whos an Apache maintainer I can reach out to to ask them why they did it that way? To get a Debian answer on this === trippeh_ is now known as trippeh [13:48] or not, we'll discuss this further === lukasa is now known as lukasa_away === lukasa_away is now known as lukasa === abhishek is now known as Guest80797 [15:48] hello I have a multipath device , I am unable to issue pvcreate command to mpath5 device [15:50] what is mpath5p1 in my /dev/mapper ? [15:51] mpath5 is multipath device I know === Vodka_Gobalsky is now known as SL89 === markthomas|away is now known as markthomas [16:13] security.ubuntu.com at 2001:67c:1562::16 is unresponsive for me. Anyone else? [16:19] jrwren: Get:1 http://security.ubuntu.com trusty-security Release.gpg [933 B] [16:19] oh wait, lemme check if its going over ipv6 [16:20] Sling: http://security.ubuntu.com started working here. I think it had to timeout and try again on a different address. 2001:67c:1562::16 is still iin the DNS and unresponsive, but I'm not blocked. [16:20] seems to work on port 80 at least for me [16:21] dunno if this is anycast dns, probably not [16:36] Hello All! [16:36] I was setting up vlan virtual NICs on an ESXi virtual host running ubuntu 14.04. when I try to get it an IP through DHCP, it never gets past DCHPDISCOVER. Note that the interface I’m cloning is physically connected directly to a modem, not a router. If I do the same procedure with an interface connected to a router, it works fine. Anyone know what the problem is, or how to fix it? [17:12] brotoes, use a modem that works :) [17:13] patdk-wk, is it a modem problem, then? [17:13] modems lock to the first mac address they see [17:13] and will NOT talk to any other mac address [17:13] the first mac address it will see, is esxi [17:13] then your vm [17:14] Ah yes, that. I’ve gone through all these shenanigans some time ago. I’ve gotten a good number of virtual hosts working on the modem. this is one of five virtual hosts on the same modem, all of which work [17:14] configure esxi/switch/... to not send lldp, stp, monintor packets, status reports, .... [17:14] and it will work === lukasa is now known as lukasa_away [17:15] and how many mac addresses are you allowed? [17:15] if you asked for 5, and esxi takes one [17:15] that leaves 4 working vm's [17:16] sorry, one of four on the modem, the virtual NIC being the fifth mac address. I’ve configured the mac addresses on the modem directly. [17:16] I have never seen a modem that lets you assign mac addresses on it [17:16] It’s a business modem [17:17] so it's not a modem, but a router? === lukasa_away is now known as lukasa [17:18] it’s very much a modem. 3v.A2010tel Telus DSL Modem [17:19] Sorry, I don’t set the mac addresses directly on the modem. the mac addresses are set through a settings portal Telus gives you. My guess is that the modem interfaces with a Telus backend to get the MACs is supposed to talk to. [17:22] no idea, they are doing something strange [17:23] If I just added a new sudo user on my server is a new set of keys required for SSH login? [17:24] JaguarDown: no, but depending on the situation they may be prudent. [17:24] Well when I try to do "ssh 192.168.0.110 -l " it says permission denied public key [17:25] the username is allowed in sshd_config [17:26] I am sure there is a simple solution but I am just a newbie. [17:26] ah you finally fixed the grub time out -1 30 is a bit much tho, 5 or 10 would have been fine [17:26] whenever something like that happens, I find it’s always permissions [17:26] JaguarDown: check to make sure the new user owns .ssh and its contents. [17:27] thanks [17:27] well...root owns it and the user is has sudo privileges [17:28] 30 secs is a lot of time when booting 50 server individual, so I still need to edit grub [17:28] just like my main user. [17:28] chmod 750 .ssh and chmod 600 .ssh/* [17:28] the world should not be able to write to your keys or read the private ones [17:29] why still the grub time out when set to less than zero tho? [17:29] warning [17:30] because it aint, its now 30 and not -1 [17:30] it's set that way already [17:31] is it because the new user doesn't have an .ssh directory? [17:31] change the owner to the new user. if that’s a problem, see what /var/log/auth.log says [17:31] if there’s no .ssh directory, where’s the authorized_keys file going? [17:31] yep, you need one ;) [17:32] any chance of less reboots Ubuntu? [17:33] good point [17:33] ;) [17:33] so I assume I will just copy the .ssh directory from the other sudoer to this new one [17:35] if you only want to be able to log IN from the same users as you can to the other one, just copy authorized_keys [17:36] if you want the new user to be able to log in to all of the same stuff as your old sudoer can, then copy id_rsa. however, keep in mind that everyone who can use the new user will have complete access to everything the old user does because they’ve got your old private key. [17:41] Well this more of a for fun experiment for expanding my knowledge and I have no concern for access because I will be using both accounts. [17:42] so to reiterate, the new sudo user is a user on the server and I want to login to the server as this new sudo user [17:42] so all it needs is the private key? [17:42] er [17:42] the public key I mean [17:43] yes. the private key you log in with needs its corresponding public key to be in the authorized_keys file in the users home directory you’re logging in to [17:44] and for future reference, I find 90% of ssh problems are due to bad permissions/file ownerships somewhere [17:44] when diagnosing this yourself, /var/log/auth.log is your friend. [17:46] ah [17:46] thanks. [17:56] guys lower the grub timepout to 5 or 10 [17:56] timeout [17:56] 30 is over the op [17:57] top [18:02] my /var/log/auth.log just says failed public key and gives the fingerprint [18:02] along with username, LAN ip, ports, etc [18:17] brotoes: Thanks for the help sir I finally got it working and you were right it was file permission problems. I just had to use the -a option to preserve read/write/user permissions then I just chown to the new sudoer and it works perfectly. [18:18] While copying the .ssh directory, that is. [18:18] wonderful! glad I could help [18:37] I have a server that refuses to boot. After the boot agent, nothing appears, it just halts. No grub menu, nothing. I have tried reinstalling grub from a live CD (by chrooting in, then doing grub-install /dev/sda) but it doesn't change anything. what can I try next? [19:05] no answers here as usual [19:50] hi there, is anyone here who can help me with building and configuring apache2 and php5? [19:50] why do you say 'building' [19:51] what do you need to 'build' those for [19:51] i have to make a server with several php versions [19:51] its for testing purposes [19:51] they'll all conflict with each other [19:51] so you can't sanely have 'multiple versions' available, AFAIK [19:51] not without multiple testing platforms [19:52] mhm, how does phpbrew handle that? [19:52] I mean, it seems to do the same.. [19:52] it uses userspace [19:52] so it runs as normal user, right? [19:53] that's also third party software. [19:53] yeah I know, just read about that.. [19:53] doing the same with a pure Ubuntu server builidng everything from source... [19:53] that's a lot trickier [19:53] okay, let me go a bit deeper [19:54] building apache2 and php5 (one version) is just for me, learning a bit about compiling from sources and setting the right parameters [19:54] the server with the multiple versions will be a jenkins server [19:54] jenkins needs the php version for tests [19:54] Pupp3tm4st3r: well, IDK how jenkins does that. [19:55] but building form source is just "read the documentation as it explains it" [19:55] for the most part [19:55] * teward disappears to a meeting [19:55] (MAYBE someone knows better than I do on this...) [19:55] thanks teward [19:56] whats the real difference between fcgid and php-fpm? [19:56] so many questions :( all I read was not that much helpfull [19:58] This is probably a good use case for LXC and/or Docker, I think. [19:58] One or the other of those would let you have separate environments with different PHP versions. [19:59] tonyyarusso: that doesnt fix Jenkins though? [19:59] Pupp3tm4st3r: uncanny timing! I just found out about this http://3v4l.org/ [19:59] teward: No idea how that works. [20:00] Our developers already use a Jenkins machine with 3 php versions in /opt/php/... [20:00] these were self compiled, but the one who built this machine has gone now [20:00] and I want to understand more... [20:01] so it generally works [20:01] think that jenkins only uses the path to binarys for testing purposes [20:01] *binaries [22:08] can someone help me set up a mail server that sends emails from users on a local network [22:10] mailserver: https://help.ubuntu.com/14.04/serverguide/email-services.html [22:10] sarnold: Utopic Is Dead, now i can free up a few hundred gigs of disk space xD [22:12] a few hundred gigs?? ouch :) [22:13] yeah i have a few VMs sitting around [22:13] not my fault [22:14] AND I can close 3 Ubuntu Bug Tasks against NGINX now that the announce went out and it's been marked Obsolete! [22:14] can't do that for utopic-backports but bleh [22:14] heh [22:20] Isn't Utopic EOL now anyways? [22:21] genii: just died today, officially as of about what 20 minutes ago? [22:21] https://lists.ubuntu.com/archives/ubuntu-announce/2015-July/000198.html [22:21] that went out... [22:21] wow about 30 minutes ago [22:21] 20-30 minutes ago [22:21] Ah, I need to check my email more often [22:22] yup [22:28] :-) === YamakasY is now known as YamakasY_ === YamakasY_ is now known as Yamakasi === Yamakasi is now known as YamakasY === Lcawte is now known as Lcawte|Away