| keithzg | Hmmph. I feel like this entire scheme was based on using a hosts file, and everything we've done technologically since it was no longer viable to keep a copy on every machine is just a rickety hack keeping a now-untenable scheme alive :P | 00:16 |
|---|---|---|
| bekks | What exactly is this "scheme" your a re talking about? | 00:17 |
| bekks | Ensuring name resolution with a host file in case of loosing the DNS servers is quite viable, though. | 00:17 |
| keithzg | bekks: I mean literally everything since we moved past that first scheme of using a hosts file, so mostly I guess I mean DNS. | 00:20 |
| bekks | What exactly is this "scheme" your a re talking about? | 00:20 |
| keithzg | Scheme: a plan, program, or policy officially adopted and followed | 00:21 |
| bekks | The only "schemes" I know of are inside a database which is capable of schemes and roles. | 00:22 |
| bekks | So actually you are talking about "we are using host files"? | 00:22 |
| keithzg | Nope | 00:22 |
| keithzg | I'm talking about "gee, it's way easier to remember names instead of numbers" and then the scheme that was invented to implement that. | 00:23 |
| bekks | So you are talking about host files. | 00:23 |
| keithzg | Nope. | 00:23 |
| keithzg | Kindof. | 00:23 |
| bekks | Then you are talking about WHAT exactly? :) | 00:23 |
| keithzg | That was the initial iteration of the scheme. | 00:23 |
| bekks | Of which scheme? | 00:23 |
| keithzg | But that became untenable, so now we try to automate it with DNS. | 00:24 |
| sarnold | bekks: I suspcet it's some thing they've written themselves .. | 00:24 |
| keithzg | The scheme to map names and numbers together for ease of human understanding. | 00:24 |
| bekks | I am giving up - you are puzzled inside theoretical terms and you cant even explain what you are doing, in particulat. | 00:24 |
| bekks | *particular. | 00:24 |
| bekks | So I am resting your case, sorry. | 00:24 |
| keithzg | bekks: that's because I *am* talking in generalities. | 00:24 |
| keithzg | I'm saying "this entire set of systems and technologies we've developed to ease human usage of networking addresses" | 00:25 |
| bekks | Can you define your actual Ubuntu support question? | 00:26 |
| bekks | Despite generalities? | 00:27 |
| keithzg | Yeah, I did earlier but everyone ignored me for hours so I just started lamenting the generalities of it, heh | 00:27 |
| bekks | Which led to nothing, as you can see. | 00:27 |
| keithzg | Hosts on our network are increasingly resolving as "hostname.local" | 00:27 |
| sarnold | is it this? < keithzg> Arghhh, I just simply cannot figure out why an increasing number of hosts on my network are resolving as "hostname.local" instead of "hostname.our.fqdn" . . . | 00:27 |
| keithzg | sarnold: yup | 00:27 |
| sarnold | apt-get purge avahi on all your systems | 00:27 |
| keithzg | Already did. | 00:27 |
| sarnold | apt-get destroy all your OS X and Windows systems? :) | 00:28 |
| keithzg | Heh | 00:28 |
| keithzg | Oh how I wish! | 00:28 |
| sarnold | I think apple brought us this .local insanity | 00:28 |
| maxb | You may wish to uninstall libnss-mdns (or unconfigure it in /etc/nsswitch.conf) if you don't want it to be used | 00:29 |
| keithzg | Yeah, it generally comes from bonjour, eh? | 00:29 |
| bekks | It does, yes. | 00:31 |
| keithzg | I'm just slightly confused as to why it's overriding all the instances of"dns-search our.fqdn" defined in all the /etc/network/interfaces files on our local servers, and other such configs. | 00:31 |
| sarnold | "search" just tells the resolver library what domain to append to all queries | 00:32 |
| keithzg | maxb: Yeah, I *though* I did that everywhere too, but it's probably worth double-checking, thanks. | 00:32 |
| bekks | keithzg: the order in nsswitch.conf rules. | 00:32 |
| sarnold | so it'll ask for e.g. www.google.com.our.fqdn. before www.google.com. | 00:32 |
| sarnold | it doesn't say anything about how reverse lookups are done, e.g. to turn 192.168.1.1 to router.our.fqdn. | 00:33 |
| keithzg | sarnold: ah, fair enough. | 00:34 |
| keithzg | bekks: so if it was trying to use mdns that'd be listed there in the hosts line, right? ex. "hosts: files mdns4_minimal dns" or such, right? (I do vaguely remember tangling with this before, now) | 00:38 |
| keithzg | Well, I've gone around and made double-sure, and there was indeed still a service or two still running (the server setup here is a jungle I've only slowly tamed, alas), Seems to be free of .local hostnames now, thanks sarnold, bekks and maxb :) | 01:00 |
| sarnold | keithzg: good luck keeping it that way :) | 01:00 |
| keithzg | sarnold: heh yeah | 01:01 |
| keithzg | Now if only I could figure out why reverse lookup is being so awfully slow, but since DNS is still handled here by the isc dns service on an ancient Trustix server that nobody at work wants me to remove yet I'm not sure there's much the ubuntu-server channel can help me on for that :P | 01:01 |
| sarnold | trustix, wow :) | 01:02 |
| keithzg | Yeah, if I want outside support for that I'd have to hop in my time machine ;) | 01:02 |
| Logos01 | Heheh... | 01:33 |
| Logos01 | I love fucking with apt-get sometimes. | 01:33 |
| Logos01 | "183 upgraded, 0 newly installed, 0 to remove and 4 not upgraded." "Need to get 65.0 MB/416 MB of archives." "After this operation, 8312 kB of additional disk space will be used." | 01:33 |
| Logos01 | (This is for a brand new, never-before-initiated, apt-get upgrade.) | 01:33 |
| bekks | Whats wrong at that point? | 01:35 |
| Logos01 | Nothing. | 01:35 |
| Logos01 | I know why it's doing it. | 01:35 |
| Logos01 | It's just amusing because superficially it makes the process seem schizophrenic. | 01:35 |
| Logos01 | "I am psychic enough to only need 65 out of 416MB to update your software for you!" "Of course, once I grab all 416 MB, you'll have 8MB extra diskspace used." | 01:36 |
| bekks | Do you have a specific support question? | 01:38 |
| Logos01 | Sorry, no -- I'm just amused by it. | 01:40 |
| Logos01 | NFS repositories are a good thing. | 01:40 |
| === markthomas is now known as markthomas|away | ||
| === zz_DenBeiren is now known as DenBeiren | ||
| lvmer | um kind of confusing question to explain: 1) Want user greg & user john to be able to edit all samba files. But all other users as members of group: publicshare to only be able to read/execute. Which I think I have setup nearly correctly. | 02:31 |
| lvmer | Should I make a new group called: smbeditors (or similar) and put both greg and john in that group and then chown smbeditors:pubshare -R /sharefolder ? Right now all folders are: john:pubshare & chmod = 0755 for most folders and 0750 or 0700 for others. | 02:31 |
| lvmer | This will let both john & greg change file names? Or is there a better way. Right now, john & sudo edit, but greg does not. But I cannot make greg a member of group john because greg should not have access to other folders that are john:john. | 02:31 |
| lvmer | kinda confused because file permissions are owner:group:other... and Idk how this will work because I think it makes both greg and john now in the group category. | 02:33 |
| lvmer | ah I'll ask in samba | 02:35 |
| linuxmint | So I installed NFS server on Ubuntu server. I have 2 computer clients to transfer files to the NFS server. I add 4 HDDs to the NFS server (1 running Ubuntu server OS). I'm not clear how to transfer files to the correct HDDs and format the HDDs? | 03:01 |
| linuxmint | I think I need to install NFS client on the computer client too, which I'm researching now. | 03:02 |
| bonhoeffer | thinking of installing a process monitor that kills and reboots bad processes. anyone install monit? is it a good way to go? | 03:30 |
| keithzg | bonhoeffer: Personally I've always felt that if you need to monitor for, and then kill, bad processes, then you have some underlying problems that need fixing! | 03:39 |
| Turner | This is a support channel for the ubuntu server right | 03:52 |
| lvmer | Turner: yes | 03:52 |
| Turner | Ok sorry just wanted to make sure. So I have a question, is there a way to like say reset everything to the way it was when i first logged into the server? The reason is was I was installing multicraft and the installation messed up so I had to reinstall it, then it kept giving me this one thing and so then i tried some other stuff and long story short it doesnt really work now. (anything that I did during the install) Im runni | 03:54 |
| sarnold | Turner: you're cut off at "Im runni" | 03:55 |
| Turner | What do you mean? At the end where it says Im running ubuntu 14.04 | 03:55 |
| sarnold | irc has line length limits, and you hit it :) hehe | 03:56 |
| sarnold | it's so often nearly nothing.. | 03:56 |
| sarnold | Turner: you can use dpkg --purge to remove individual packages and their configuration files | 03:56 |
| sarnold | Turner: so if multicraft was installed via dpkg or apt-get you can probably remove all its configuration easily | 03:56 |
| Turner | it was installed with apt-get | 03:57 |
| Turner | or that was in the front of the command | 03:57 |
| sarnold | nice. then you can take those package names and run apt-get purge <packagename> and it'll delete that package | 03:57 |
| sarnold | so you can then re-run apt-get install <packagename> and it'll re-install giving you a blank slate to work with | 03:58 |
| Turner | and i take it the package names would be what ever came after apt-get install | 03:59 |
| sarnold | yes | 03:59 |
| Turner | Ok let me try this | 04:00 |
| Turner | Ok so I think I removed all the packages i installed is there a way to list the ones that are on it atm? | 04:05 |
| Turner | Ok so I removed all the stuff in added. But theres one thing that I cant remove with the apt-get purge its muticraft and I think why is because i used wget http://www.multicraft.org/download?arch=linux64 -O multicraft.tar.gz to get it | 04:08 |
| Turner | know how to delete that file? if I go to cd multicraft its there | 04:08 |
| Turner | Im sorry im a noob and I want to be a programmer -.- good luck to me | 04:08 |
| Turner | oh wait maybe i got it. let me try and install everything now | 04:12 |
| Turner | Oh wait this was the error I was getting " * Starting web server apache2 AH00558: apache2: Could not reliably determine the server's fully qualified domain name, using (IP NAME WAS HERE). Set the 'ServerName' directive globally to suppress this message" any ideas? | 04:14 |
| linuxmint | Hello, my ubuntu server names the HDDs incorrectly. E.g, SATA 1,2,3,4 should be named /dev/sda,/dev/sdb,/dev/sdc,/dev/sdd, but /dev/ is all mixed up. Can I rename to correct this? | 04:44 |
| andol | linuxmint: Mixed up, as in with completely different devices names, or just in the "wrong" order? | 05:21 |
| andol | linuxmint: Dependings on the issue being caused, it might be worth using UUIDs instead. | 05:23 |
| linuxmint | andol: wrong order. I've moved on as apparently the /dev naming can't be managed. | 05:24 |
| linuxmint | Prob I have now is, my Ubuntu server install was running. I just pulled the physical HDD out of the computer bay and the monitor went snowy then black. Now the computer power LED doesn't turn on and monitor is black. CPU fan runs thought. No access to BIOS. Did I wreck the HDD or MOBO? | 05:24 |
| linuxmint | ok, HDD works in another computer. | 06:12 |
| linuxmint | Other computer won't download gparted, stopping me reformatting the HDD. | 06:13 |
| linuxmint | I suspect the MOBO is blown, but no clear way to tell. CPU fan does run. | 06:15 |
| zatricky | Hey, guys - possibly not an Ubuntu-specific query (though this is for an Ubuntu server). I'm attempting to track down outbound connections to the local Apache server. Unfortunately netstat shows dashes in the pid column - super useless. tcpdump shows the traffic is going over the loopback interface using the public IP addresses. | 06:42 |
| zatricky | To answer my own query - lsof showed the culprit - I had to put it into a loop and quit as soon as it found something. I suspect the connections were so short-lived that typically the pid never existed by the time it was requested. But lsof appears to be more robust or "quicker" to get that data out. :) | 07:13 |
| zatricky | while true ; do lsof -Pni | grep "$local_ip:80 " && break ; done | 07:15 |
| linuxmint | So, HDD not found on server. Put HDD into LinuxMint and finds HDD. Reformatted HDD to FAT32, Mint finds HDD. Put HDD back into server and doesn't find HDD? | 07:20 |
| === Lcawte|Away is now known as Lcawte | ||
| linuxmint | Ok, walkthrough for NFS server setup says # exportfs -ra. I receive error: exportfs: /etc/exports:1: unknown keyword "re" | 08:18 |
| soren | linuxmint: Clearly the first line of /etc/exports is busted. | 08:20 |
| soren | linuxmint: Dude, seriously, read the error message. | 08:20 |
| linuxmint | soren: I fixed it, coding typo. | 08:21 |
| linuxmint | soren: I read the error message, but didn't understand what is was referring to, but now I do. | 08:21 |
| soren | It's pretty clear. Line 1 of /etc/exports says "re" somewhere where it shouldn't. | 08:22 |
| linuxmint | Can I ask though, the NFS walkthrough says to edit /etc/default/portmap to get rid of "-l" or "-i 127.0.0.1". When I open the file there's no data in it? | 08:23 |
| lordievader | Good morning. | 08:23 |
| linuxmint | Ok, NFS server restarted, but portmap is a worry. Hopefully remote clients will be able to access it. | 08:24 |
| linuxmint | So, final problem I'm stuck on is an NFS client won't recognise an HDD. The HDD and SATA data cable are tested ok. I swapped the SATA port connection and now both HDDs aren't recognised? I think I'll have to check the BIOS. | 08:25 |
| linuxmint | ok, checked BIOS which also doesn't find the HDD. | 08:38 |
| lordievader | linuxmint: If you stick the hdd in another pc, does that one detect it? | 08:48 |
| linuxmint | lordievader: yes, HDD works find in LinuxMint. | 08:48 |
| linuxmint | *fine | 08:48 |
| linuxmint | This Ubuntu server has a new MOBO so I assume BIOS is up to date...although maybe too modern BIOS for the HDD? | 08:48 |
| linuxmint | lordievader: I tried different SATA ports, which all work, but don't recognise the HDD. | 08:49 |
| linuxmint | lordievader: it's a Seagate Barracuda | 08:50 |
| lordievader | linuxmint: Does dmesg say anything about it? | 08:50 |
| soren | It's obviously not "LinuxMint" that finds it. | 08:51 |
| soren | You could install Mint on the other server all day long, but if the BIOS doesn't see it, you won't see it once it's booted. | 08:51 |
| linuxmint | lordievader: dmesg is too quick to see any message. I'll have to load the computer and try a command like $ dmesg or dmesg | tail -30. I still get stuck with viewing all the output of dmesg. | 08:52 |
| linuxmint | soren: ye.s | 08:52 |
| linuxmint | *yes. | 08:52 |
| lordievader | linuxmint: Yes, search through it... | 08:53 |
| linuxmint | Anyone recommend an alternative to pastebin? I'm not confident Terminal lets me see all of $ dmesg. I scroll to the top of Terminal and can't see my original command $ dmesg. | 09:01 |
| linuxmint | lordievader: perhaps line 346? http://pastebin.com/wzP0GxkM | 09:03 |
| lordievader | linuxmint: Do you have four ssd's in your system? | 09:16 |
| lordievader | I see a bunch of Intel SSD's and one Hitachi HDD. | 09:17 |
| zetheroo1 | "implementation of systemd in upcoming Ubuntu releases might mean we should dump Ubuntu Server as systemd is bloated and not really what system admins need/want" - So I have heard ... I am not sold ... hence I am consulting the Oracle - you all! :D | 09:27 |
| linuxmint | lordievader: yes, I'm thinking or learning how to make RAID. I thought I'd try RAID 5 but not sure if it will work or be worth it? | 09:38 |
| linuxmint | lordievader: they're small, so I could afford them. | 09:39 |
| lordievader | linuxmint: I have no experience with raids. | 09:41 |
| linuxmint | lordievader: glad I'm not the only one. Thought I'd try, but don't see the worth in it really, apart from bragging rights. | 09:42 |
| lordievader | Oh, I can see the advantages. It is simply that I never wanted to spend money on it ;) | 09:47 |
| === matsubara-afk is now known as matsubara | ||
| gnuoy | jamespage, If I'm creating a package merge request bug with an associated bzr branch I assume I don't need to upload deb diffs to the bug? | 09:58 |
| jamespage | gnuoy, bzr or debdiff | 09:59 |
| jamespage | not and :-) | 09:59 |
| gnuoy | tip top | 09:59 |
| gnuoy | thanks | 09:59 |
| zetheroo1 | so apparently systemd is only good for desktops and not server ... | 10:02 |
| zetheroo1 | anyone!? | 10:03 |
| ogra_ | zetheroo1, what alternative server distro would you use then | 10:03 |
| zetheroo1 | well I am being told to switch to BSD | 10:03 |
| ogra_ | i dont think any but slackware and gentoo will come without it | 10:03 |
| zetheroo1 | but is it really true that systemd is only useful for Desktops? | 10:05 |
| ogra_ | nonsense | 10:05 |
| ogra_ | thats like asking if upstart is only good for desktops ... | 10:07 |
| zetheroo1 | ok | 10:09 |
| zetheroo1 | another thing I am being told is "if you need to fix stuff, you need to recompile C++ programs" ... | 10:09 |
| zetheroo1 | true? | 10:09 |
| ogra_ | its is just an init system ... its is different to others and yes it swallowed (and replaced) a bunch of core functionality with new stuff ... and yes, it offers some extra features for better desktop integration ... but it will be the default in most linuxes in the future which means documentation can be unified etc etc | 10:10 |
| ogra_ | why dont you ask these questions in a systemd channel ? | 10:11 |
| zetheroo1 | ok ... sorry, didn't know there was one ... | 10:11 |
| zetheroo1 | is it on freenode? | 10:11 |
| ogra_ | no idea | 10:12 |
| ogra_ | it just doesnt seem like the right place to ask questions an upstream developer could answer you better | 10:12 |
| zetheroo1 | ok :) | 10:14 |
| gnuoy | jamespage, Are you happy to sponser Bug #1401461 ? and if so should I assign it to you? | 10:22 |
| uvirtbot | Launchpad bug 1401461 in python-logutils "Please merge python-logutils 0.3.3-2 (main) from Debian unstable (main)" [Undecided,Confirmed] https://launchpad.net/bugs/1401461 | 10:22 |
| johncarper_ | Hello, this question might be abit offtopic here, but I recently purchased a 5.25 inch SATA Hard Drive mobile rack that has a molex connector for the power: http://i61.tinypic.com/2556k4n.jpg will I be able to power it trough my PSU with this cable: http://i59.tinypic.com/25i0p5v.jpg ? | 10:39 |
| maxb | 5.25 inch SATA kit, really? | 10:43 |
| maxb | The connector on the left does look like a molex power one | 10:44 |
| linuxmint | So, can a faulty HDD stop a working OS HDD booting? Seems to make computer load into GRUB rescue? | 12:02 |
| YamakasY | can't we run a centos mirror on Ubuntu ? we can run a Ubuntu mirror on centos | 12:19 |
| === Lcawte is now known as Lcawte|Away | ||
| makara | hi. I'm trying to relink a deleted file inode using debugfs, but it doesn't recognize the filesystem /dev/simfs. I have to use debugfs. What can I do? | 12:30 |
| makara | i see this guy has the same problem: http://stackoverflow.com/questions/26326123/dev-simfs-no-such-file-or-directory-while-opening-filesystem | 12:32 |
| linuxmint | I'm stuck in GRUB rescue, can't get out? | 12:35 |
| linuxmint | I try boot, but error: Unknown command 'boot' | 12:37 |
| === Lcawte|Away is now known as Lcawte | ||
| dvargek | hi guys, anyonen tried to install kernel 3.16 kernel on ubuntu 14.04 ? | 13:44 |
| dvargek | i had a problem last week, with a missing raidcontroller module | 13:45 |
| dvargek | it seems, the module 'megaraid_sas' is not included in the package | 13:47 |
| smb | hallyn, stgraber, I subscribed you both to bug 1401148 so we can figure out things. Last comment is a question which you probably can answer more quickly | 13:55 |
| uvirtbot | Launchpad bug 1401148 in linux "Re/starting an lxc container corrupts all network namespaces on the same physical host" [Undecided,Confirmed] https://launchpad.net/bugs/1401148 | 13:55 |
| stgraber | smb: gave you a vague reply on there, hallyn would have more details | 14:01 |
| smb | stgraber, ok, we may want the more info then. Like I noted a bit above the bad things don't seem to happen when one lets aa only complain about stuff | 14:04 |
| caribou | rbasak: do you think that it would make sense to add a uvtool-simplestreams command to sync local images ? | 14:18 |
| rbasak | caribou: what do you mean exactly? | 14:21 |
| caribou | rbasak: to be able to load a local cloud image (i.e. an image file downloaded to local disk) as an image available for uvtool to use | 14:22 |
| caribou | rbasak: for instance, I build my own debian jessie image that I want to use with uvtool | 14:22 |
| rbasak | caribou: ah, I see. Yes - I'd like to have something like that. Better, integrate everything so you sync to only one place and everything can use it (including libvirt, tgz users and qcow users) | 14:22 |
| rbasak | caribou: for your own image, you can use --backing-image-file | 14:23 |
| caribou | rbasak: ah, let me look at that | 14:23 |
| alex88 | hi guys, https://gist.github.com/alex88/26b064ec9ce50bc6d961 I'm trying ot downgrade a package, it says it will be installed but it doesn't do anything | 14:26 |
| alex88 | any idea? | 14:27 |
| === wedgwood1 is now known as wedgwood | ||
| caribou | rbasak: ah, it's the only option missing in man uvt-kvm! | 14:32 |
| caribou | rbasak: want a bug for that ? | 14:32 |
| rbasak | caribou: ah. Yes please - I must have missed it and I think the manpage should be comprehensive. Thanks! | 14:33 |
| caribou | rbasak: ok, I'll create the bug & fix it | 14:33 |
| jrwren | anyone ever have a bash variable behave readonly but its not in the readonly list? http://paste.ubuntu.com/9476396/ | 14:34 |
| jrwren | not being able to set SSH_AUTH_SOCK is pretty terrible. | 14:34 |
| caribou | rbasak: do you want the bug on upstream uvtool or the ubuntu/uvtool ? | 14:40 |
| caribou | rbasak: looks like hallyn has already done it : bug #1317266 | 14:41 |
| uvirtbot | Launchpad bug 1317266 in uvtool "add an option to specify a backing store disk file" [Undecided,New] https://launchpad.net/bugs/1317266 | 14:41 |
| jrwren | nevermind. I'm a fool over ssh_auth_sock | 14:44 |
| === ikonia is now known as Christmas_Jesus | ||
| === Christmas_Jesus is now known as ikonia | ||
| === ikonia is now known as Christian_ikonia | ||
| === Christian_ikonia is now known as fictional_wizard | ||
| === fictional_wizard is now known as ikonia | ||
| rbasak | caribou: just a warning about uvtool. There's a spike for snappy that hasn't landed yet. | 14:51 |
| caribou | rbasak: ok | 14:52 |
| rbasak | caribou: https://github.com/smoser/uvtool is the spike - I need to review, modify and merge as required, and that's the top priority for now. Though I imagine a man page fix should merge in fine. | 14:52 |
| caribou | rbasak: I'll ping you before doing anything | 14:54 |
| caribou | rbasak: this --backing-image-file is just want I wanted. Was a good idea to ask first :-) | 14:57 |
| rbasak | caribou: :) | 14:59 |
| rbasak | caribou: I added it when I needed it once :) | 14:59 |
| caribou | rbasak: I use uvtool to test makedumpfile & kdump-tools kernel dumps & I needed that for debian | 15:00 |
| caribou | had to enable kdump-tools to work with systemd | 15:02 |
| === Lcawte is now known as Lcawte|Away | ||
| === MeltedDed is now known as MeltedLux | ||
| apw | hallyn, hey, this netns corruption thing. the behaviour changes if apparmour is put in moan mode, what it prevents is your attempt to remount /run/netns -slave; if you are able to do that things work, so i think this is something you need permission from aa to do | 16:13 |
| hallyn | jjohansen: ^ what would the rule be to allow the ms-slave mounts? (we turn all pre-existing mounts into slave mounts in our own ns to avoid umounting things on the host when starting a container) | 16:16 |
| apw | hallyn, i should say "appear to work to me" but i didn't do extensive tests there | 16:18 |
| Azaril | hey | 16:25 |
| Azaril | im getting | 16:25 |
| Azaril | http://pastie.org/private/newqkpx9vyhhbqfia40jtq | 16:25 |
| Azaril | i dont understant where this is coming from... | 16:25 |
| hallyn | apw: it makes sense | 16:27 |
| hallyn | apw: what is the bug# again? | 16:42 |
| hallyn | oh got it | 16:42 |
| apw | https://launchpad.net/bugs/1401148 | 16:43 |
| uvirtbot | Launchpad bug 1401148 in linux "Re/starting an lxc container corrupts all network namespaces on the same physical host" [Undecided,Confirmed] | 16:43 |
| apw | i am not sure if this is a new behaviour mounts wise from "ip" | 16:43 |
| avalon | i was installing server 14.04.1 onto a dell poweredge r300 earlier today and chose the manual partitioning option, and every step of that process took a long time for something that seems instant on other distros - is that normal, or is there a reason it might be laggy? | 16:44 |
| teward | Azaril: looks like maybe there's an issue with the mirrors somewhere... if i remember right i get that sometimes in the US mirrors, and it might resolve itself | 16:44 |
| hallyn | apw: start-container profile already has: | 16:45 |
| hallyn | mount options=(rw, slave) -> /, | 16:45 |
| hallyn | oh | 16:45 |
| apw | mount(NULL, "/run/netns", NULL, MS_SLAVE, NULL) = -1 EACCES (Permission denied) | 16:46 |
| apw | but regardless, it fails according to strace | 16:46 |
| Azaril | teward: its being doing it for a couple of weeks... | 16:48 |
| Azaril | i get a valid response from the ip which is weird | 16:48 |
| teward | Azaril: i poked the mirrors team to see if there's any known issues about it | 16:49 |
| teward | Azaril: in the mean time all I can say is maybe try using a different archive mirror? | 16:49 |
| Azaril | hmmm | 16:49 |
| hallyn | apw: adding "remount options=slave," doesn' thelp. so this seems like either an apparmor bug, or a misunderatnding of how to specify the policy | 16:49 |
| hallyn | sarnold: ^ do you know offhand? | 16:50 |
| === markthomas|away is now known as markthomas | ||
| apw | hallyn, (ro, slave) perhaps ? | 16:52 |
| apw | mount options=(slave) -> /run/netns, | 16:52 |
| apw | or | 16:52 |
| apw | mount options=(ro, slave) -> /run/netns, | 16:52 |
| apw | would be my guesses | 16:52 |
| Azaril | different mirror seems to have worked, cheers | 16:53 |
| tyhicks | apw, hallyn: this sounds like bug #1350947 | 16:53 |
| uvirtbot | Launchpad bug 1350947 in linux "apparmor: no working rule to allow making a mount private" [Medium,Confirmed] https://launchpad.net/bugs/1350947 | 16:53 |
| tyhicks | apw, hallyn: I'm only vaguely familiar with that bug but I think it may be what you're bumping into | 16:53 |
| apw | tyhicks, i suspect that that is indeed, in part what we are bumping into, in that i think we need that rule, and could not specify it even if we want to | 16:56 |
| hallyn | tyhicks: agreed, | 16:58 |
| hallyn | just added a comment to 1401148 , only 'mount,' works for me | 16:59 |
| hallyn | tyhicks: haha, and i noted htat in a comment in that bug | 17:01 |
| hallyn | (so this was known in august) | 17:01 |
| tyhicks | We've been prioritizing other bugs/features higher than that one | 17:05 |
| tyhicks | hallyn: is it now a blocker for something you're doing? | 17:05 |
| hallyn | tyhicks: not me. jamespage ^ | 17:11 |
| jamespage | hallyn, ? | 17:12 |
| hallyn | jamespage: you filed bug 1401148 | 17:13 |
| uvirtbot | Launchpad bug 1401148 in linux "Re/starting an lxc container corrupts all network namespaces on the same physical host (dup-of: 1350947)" [Undecided,Confirmed] https://launchpad.net/bugs/1401148 | 17:13 |
| uvirtbot | Launchpad bug 1350947 in linux "apparmor: no working rule to allow making a mount private" [Medium,Confirmed] https://launchpad.net/bugs/1350947 | 17:13 |
| hallyn | you didn't assign it a priority though | 17:13 |
| hallyn | so the q is is this blocking something for you | 17:14 |
| hallyn | tyhicks: i guess i'tll block use of containers in neutron? | 17:15 |
| hallyn | for zul's nc-lxd, the answer may be that since the containers run unprivileged, we do in fact just allow "mount," | 17:15 |
| hallyn | (as workaround) | 17:15 |
| tyhicks | ok | 17:16 |
| tyhicks | we'll (the sec team) will discuss if we can give it more attention in the short term | 17:16 |
| hallyn | in fact based on the Description I guess it actually prevents some basic setups (without containers as guests, just containerizing some services on the host) are bein prevented | 17:20 |
| hallyn | really i'd argue this comes down to poor design in 'ip netns', which is very limited, and should not have been used by neutron | 17:24 |
| hallyn | (wonder if making /run/netns itself unbindable would be useful) | 17:30 |
| === martinst is now known as martins-afk | ||
| sbtechcom | anyone here deal with MaaS? | 17:38 |
| jamespage | hallyn, not immediately | 17:38 |
| jamespage | hallyn, but its needs resolving for the target architecture we have for deploying openstack this cycle | 17:38 |
| jamespage | right now putting anything under lxc on a neutron gateway node is not a great story | 17:39 |
| === martins-afk is now known as martinst | ||
| === Lcawte|Away is now known as Lcawte | ||
| hallyn | jamespage: tyhicks found the problem | 17:41 |
| jamespage | hallyn, ok - so at least I'm not going mad | 17:44 |
| jamespage | hallyn, medium is fine for now | 17:44 |
| hallyn | cool, thx | 17:46 |
| jamespage | hallyn, remind me again how I fix the mtu of the veth interfaces for an lxc container to 9000? | 17:54 |
| === bilde2910|away is now known as bilde2910 | ||
| hallyn | jamespage: lxc.network.mtu = 9000 | 18:02 |
| jamespage | hallyn, hmm yes | 18:25 |
| jamespage | hallyn, OK - so I have another interesting bug then | 18:26 |
| jamespage | hallyn, if I stop/start using the lxc-* commands - all's good. | 18:27 |
| jamespage | hallyn, if I reboot within the container, the veth on the host gets reset to 1500 | 18:27 |
| jamespage | but inside it still thinks its 9000 | 18:27 |
| jamespage | that has some odd effects | 18:28 |
| jamespage | I'll raise a bug shortly | 18:28 |
| hallyn | jamespage: hm, drat, not sure where that'll be happening. got a afeeling that may become a hairy interaction of upstart jobs | 18:29 |
| === markthomas is now known as markthomas|away | ||
| johncarper | Hello, I'm trying to allow apache on iptables as i'm using a cups print server, I've tried adding serveral rules to my iptables but none of them worked. I'm trying to access it from my pc on tesame network, anyone know what might be wrong? Everything works fine with iptables disabled | 19:21 |
| johncarper | here are my iptables rules: http://pastebin.com/0yT5u1eT | 19:21 |
| Madkiss | hi folks | 19:24 |
| Madkiss | If I device to install systemd on ubuntu 14.04, does that systemd come with a compatibility layer for upstart? | 19:25 |
| lordievader | Madkiss: In 15.04 I could enable systemd without any modification. That said, systemd is to my knowledge not supported on 14.04 or 14.10. | 19:45 |
| Madkiss | okay, thans | 19:47 |
| dmsimard | Is there a specific channel for Ubuntu core ? #ubuntu-core is empty :) | 19:59 |
| soren | #snappy | 20:02 |
| soren | dmsimard: ^ | 20:02 |
| sacarde | hi | 20:02 |
| sacarde | is possible to re-exec network configuration by consolle ? | 20:03 |
| soren | Sure. restart networking | 20:05 |
| sacarde | no no | 20:06 |
| sacarde | I have to change config | 20:07 |
| sacarde | by consolle, by script automated | 20:07 |
| soren | Well, then say re-config, not re-exec. | 20:07 |
| sacarde | yes | 20:07 |
| sacarde | re-config | 20:07 |
| sacarde | like during installation | 20:07 |
| dmsimard | soren: ty | 20:11 |
| === markthomas|away is now known as markthomas | ||
| jamespage | hallyn, https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1401658 | 20:38 |
| uvirtbot | Launchpad bug 1401658 in lxc "Host veth mtu not preserved during container reboot" [Undecided,New] | 20:38 |
| hallyn | jamespage: ok, thx | 20:53 |
| lordievader | sacarde: You mean edit /etc/network/interfaces? | 20:54 |
| hallyn | jamespage: though i'm not observering the same (on utopic) | 20:56 |
| jamespage | hallyn, thats on 14.04 with stock 3.13 | 20:56 |
| * hallyn tries on a vm | 20:56 | |
| jamespage | hallyn, prob worth pointing out | 21:02 |
| jamespage | eth0 (9000) <-> juju-br0 (9000) <-> veth (9000) | 21:02 |
| hallyn | jamespage: and it's the host side veth that gets its mtu reset? | 21:07 |
| jamespage | hallyn, yes | 21:07 |
| hallyn | jamespage: oh, what guest release? | 21:07 |
| jamespage | hallyn, all 14.04 | 21:08 |
| hallyn | k | 21:08 |
| hallyn | jamespage: trusty VM with trusty containre still don't reproduce, using lxcbr0. | 21:19 |
| jamespage | hallyn, I'll try make a reproducer up | 21:19 |
| jamespage | but not tonight | 21:19 |
| hallyn | jamespage: do you have other containers running with mtu 1500 by chance? | 21:20 |
| hallyn | or, VMs attached to the juju bridge with 1500 mtu? | 21:20 |
| jamespage | hallyn, all containers are set to mtu 9000 | 21:20 |
| jamespage | well at least the ones attached to the juju-br0 are | 21:20 |
| hallyn | interesting. yeah reproducer will be appreciated - thanks, good night | 21:21 |
| blacknred0 | I think I know the answer to this..... but is there a way to have crontab sending an email without having to install a mail server? | 22:15 |
| keithzg | blacknred0: yup, in fact I swear by default it does; certainly you can use utilities like mailx | 22:16 |
| blacknred0 | thanks keithzg! - I tried without it and no mail was sent out. I will give mailx a try. | 22:18 |
| keithzg | blacknred0: no problem; worth mentioning, I don't think the traditional bsd-mailx can send without an MTA, you'll probably need to specifically use heirloom | 22:24 |
| medecau | Evening all, I am trying to disable Logstash as a service. I whish to control it through supervisord. Not sure how to go about it. #ubuntu said #ubuntu-server is the place to ask. | 22:27 |
| rberg_ | I am trying to track down why the password on my iDRAC keeps getting reset during a debootstrap based install. looking at the time stamps it appears to be happening during package installation, has anybody here experienced this before? | 22:39 |
| rberg_ | BTW I agree that doesnt make much sense :( | 22:42 |
| rberg_ | NM found it, it was a custom package | 22:47 |
| linuxmint | Could I get some help with formatting disks? I'm not sure whether to use ext1,2,3,FAT32. I only need the disk for storing backups, so maybe ext3. The command I have is # mksf.ext3 /dev/sde. Could someone confirm as it's a bit scary. | 22:51 |
| linuxmint | Actually, how to check if the new disk is formatted or what's on it? I tried # cd /dev/sde, but error: Not a directory. | 22:55 |
| keithzg | linuxmint: you can use fdisk or parted to look at how drives are partitioned or formatted from the terminal. If you have a GUI and you're new at this, you might want to use GParted instead. | 23:05 |
| keithzg | linuxmint: fdisk is the classic way of listing partitions, but parted is the more modern one; "sudo parted --list" will show you how things are currently partitioned, and you can use to to format the drives/partitions themselves too. | 23:13 |
| keithzg | (main reason to use parted instead of fdisk is that fdisk doesn't support GPT, which is the partition table scheme that is often used these days) | 23:14 |
| linuxmint | keithzg: thank you. Yes, I'm familiar with # fdisk -l and GParted, however I'm SSHing into the computer, so I don't think the GParted GUI will work. The new HDD doesn't show any partitions...just 500 GB, so I thought I'd see if it's formatted or has any files, before trying to format to perhaps ext3? | 23:17 |
| keithzg | linuxmint: Yeah, then gparted or fdisk are your friend (although to see if there are any files, you'll then have to mount any partitions you see). But you *can* get GUIs working, if you ssh with the -X flag! (Although that's slightly tricky since you'll probably need root to use GParted). | 23:20 |
| keithzg | err, sorry, by gparted first there I just meant parted. | 23:21 |
| keithzg | GParted is the frontend for parted. | 23:24 |
| keithzg | Or more accurately, I think they both use libparted. | 23:25 |
| === Lcawte is now known as Lcawte|Away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
