[02:24] <thumper> stgraber, hallyn: if I create a "template" container that doesn't autostart, is it possible to change a clone of that machine to autostart?
[02:25] <thumper> also, is the lxc.conf file from the original template copied anywhere where I can poke it?
[02:51] <hallyn> thumper: sure, just edit its config file.
[02:51] <hallyn> i.e. if you did sudo lxc-clone -s -o u1 -n u2, edit /var/lib/lxc/u2/config.
[03:03] <GeekDude> New install of server, with xfce as a desktop environment. I can't mount drives through it, but I hear I need something called a polkit. Can anyone point me in the right direction to get a polkit (such as lxpolkit) isntalled and set to start with xfce?
[03:06] <darthanubis> GeekDude, why did you not just install xubuntu?
[03:07] <GeekDude> darthanubis: Uuuuh
[03:07] <GeekDude> *best answer*
[03:09] <GeekDude> Because I'm inexperienced with this as a whole, and am lucky to even have gotten as far along as I have
[03:09] <darthanubis> ok
[03:09] <darthanubis> just was wondering, because it would have been easier for you
[03:09] <GeekDude> I don't necessarily want it to be easy. I'm building good experiences for future reference
[03:10] <GeekDude> That, and I'm out of CDs to burn
[03:10] <GeekDude> It can't boot from usb either :/
[03:11] <darthanubis> understood
[03:12] <GeekDude> Should I just grab https://www.archlinux.org/packages/?name=lxpolkit ? It downloads as an xz archive
[03:14] <GeekDude> then set it up to autostart as per https://wiki.archlinux.org/index.php/Autostarting
[03:17] <darthanubis> don't think that is the way you want to go
[03:17] <darthanubis> just grab it from the ubuntu repos?
[03:17] <GeekDude> lxpolkit isn't in the repos
[03:17] <GeekDude> from what I can tell
[03:18] <thumper> hallyn: edit before start and all is good?
[03:18] <thumper> if so, awesome
[03:19] <darthanubis> http://packages.ubuntu.com/raring/x11/lxpolkit
[03:19] <thumper> hallyn: is it ok to stop a container, edit the config to add a mount, and restart, and the mount will work?
[03:19] <thumper> if the mount is an external user directory
[03:27] <thumper> hallyn: when I cloned and started 50 lxc containers, I ended up hitting the "too many open files", is this likely to be for the root user on the host?
[03:46] <hallyn> thumper: for autostart?  you dont' have to editbefore you start;  just before you run lxc-autostart or shut down
[03:46] <hallyn> thumper: on restart and mount: yes
[03:46] <GeekDude> darthanubis: sudo apt-get install gksu-polkit would be what I'm looking for, if I want to install lxpolkit?
[03:47] <thumper> hallyn: ok, but editing for new mounts is fine if stopped, and when started all is good?
[03:47] <hallyn> thumper: as for 'too many open files' - hm.  i'm not sure.
[03:47] <thumper> yeah, load testing FTW
[03:47] <thumper> "juju add-unit -n 50 ubuntu" with the local provider :)
[03:47] <hallyn> these are root started containers?
[03:47] <thumper> I'm making juju btrfs aware for fast snapshotting
[03:47] <thumper> hallyn: yes
[03:47] <thumper> well, jujud started containers, and jujud is running as root
[03:47] <GeekDude> darthanubis: or at least close enough
[03:48] <hallyn> is this the /proc/sys/fs/file-nr limit, or ulimit, i wonder?
[03:48] <darthanubis> sounds right
[03:48] <darthanubis> which comes with the desktop install
[03:49] <thumper> hallyn: ulimit -n for root is still 1024
[03:49] <hallyn> you really shouldn't be restricted to 50 continers...
[03:49] <thumper> lxc said they were all started
[03:49] <thumper> however the cloud init logs were full of inotify problems
[03:49] <thumper> with too many open files
[03:49] <thumper> and upstart failed to recognize new jobs
[03:49] <GeekDude> darthanubis: the package lxpolkit doesn't show up until you get to quantal, and I'm on pangolin
[03:49] <thumper> probably due to inotify issue
[03:49] <hallyn> ok, so you can up the limit :)
[03:50] <GeekDude> darthanubis: I think that gksu-polkit would be the proper alternative, though
[03:50] <GeekDude> I'm unsure, but trying things is part of learning
[03:50] <thumper> hallyn: the jujud upstart jobs sets to 20000
[03:51] <thumper> since it is calling the lxc commands
[03:51] <thumper> is it likely that limit is inherited?
[03:51] <thumper> or something else?
[03:51] <thumper> this is well outside my area of expertise
[03:51] <hallyn> i suspect every caller of lxc-start needs to set it
[03:52] <thumper> in this case, there is exactly one caller of lxc-start
[03:52] <thumper> and that is the jujud process
[03:52] <thumper> hallyn: was very impressed though, that with 50 containers on btrfs loopback device mounted to /var/lib/lxc, it was only taking 2.5GiB
[03:53] <thumper> although really, just the ubuntu charm
[03:53] <thumper> which doesn't to a lot :)
[03:53] <thumper> s/to/do/
[03:53] <hallyn> thumper: yeah but a full ubuntu release in each :)
[03:53] <thumper> sure, cloud image, but yeah
[03:54] <thumper> pretty cool
[03:54] <hallyn> thumper: yeah, playing around with 'sleep 200 & cat /proc/$!/limits | grep file' suggests you should be fine with jujud
[03:54] <hallyn> cloud images is bigger than the base ubuntu image
[03:54] <thumper> about 900M for my template image
[03:55] <thumper> much better than 50G :-)
[03:55] <thumper> which I would have without the snapshots
[03:58] <GeekDude> darthanubis: That didn't seem to help. I'm trying apt-get install xfce4-mount-plugin now
[03:58] <GeekDude> Hmm. Didn't seem to help either
[03:59] <darthanubis> try apt-get install xubuntu-desktop
[04:00] <GeekDude> right now I'm trying xfce-goodies
[04:00] <darthanubis> has nothing to do with it
[04:00] <GeekDude> no?
[04:00] <darthanubis> are you researching what you are doing or just install stuff?
[04:01] <GeekDude> uuh
[04:01] <GeekDude> I'm researching a bit, then installing anything tangentially related until it works
[04:01] <GeekDude> I'm a terrible admin
[04:02] <GeekDude> darthanubis: If I install xubuntu-desktop, will it just completely redo my desktop?
[04:04] <GeekDude> Well, I've gotta go :/
[04:04] <GeekDude> See you all later
[04:31] <jak2013> how to  open a port? on ubuntu 12.04
[05:36] <usr13_> jak2013: It would have had to have neen previously closed in order to open it.  So, how did you close it, with netfilter?
[05:37] <usr13_> jak2013: ... or did you just kill the service?
[05:38] <usr13_> jak2013: What port are you wanting to use?  And for what servcie to you want to use it?
[05:39] <jak2013> mysql port: 3306
[05:39] <usr13_> jak2013: By default, they are all open.
[05:42] <usr13_> jak2013: you might want to /join #netfilter
[06:00] <jak2013> usr, uwf is the firewall right?
[06:02] <sarnold> jak2013: ufw is a frontend to the iptables firewall framework
[06:14] <jak2013> mmmmm cant connect 3306 arent opened :(
[06:14] <jak2013> thanks
[06:17] <sarnold> jak2013: is something listening on 3306? check netstat -nlp | grep 3306
[06:30] <jak2013> sarnold:
[06:30] <jak2013> tcp        0      0 127.0.0.1:3306          0.0.0.0:*               LISTEN      18205/mysqld
[06:31] <sheptard> so it's just listening on localhost
[06:31] <sheptard> which is why you can't connect to it remotely
[06:31] <sarnold> jak2013: good. and what happens when you run the mysql client command to connect to it?
[06:31] <jak2013> canc onnect
[06:31] <jak2013> but locally
[06:31] <jak2013> and remotelly? how to open the port?
[06:31] <sheptard> edit your my.cnf
[06:32] <jak2013> or for access not only locally
[06:32] <jak2013> sheptard and?
[06:32] <sarnold> tell mysql to listen on an external interface
[06:32] <sheptard> http://www.cyberciti.biz/tips/how-do-i-enable-remote-access-to-mysql-database-server.html
[06:32] <sarnold> I strongly recommend using a firewall of some sort to restrict who can connect to your mysql server
[06:34] <sarnold> you don't want it sitting on the public internet waiting for someone to either brute-force guess passwords or worse smack it with an exploit.. there've been more than one remote unauthenticated mysqld vulnerability in the psat
[06:34] <jak2013> then how to connecto to my db?
[06:36] <sarnold> jak2013: this is a good tool for testing: http://manpages.ubuntu.com/manpages/precise/en/man1/mysql.1.html
[06:42] <jak2013> ok, thanks
[07:53] <wkmanire> Hello, I'm having trouble installing 12 LTS on an older 32 bit machine (built in 2000). I want to use it as a DNS server.
[07:53] <wkmanire> GRUB loads, but then all I get is a flashing underscore.
[07:53] <wkmanire> Nothing changes even after 10 minutes of waiting
[07:54] <wkmanire> I'm installing from a live USB device.
[07:54] <wkmanire> Which I formatted from the 32 bit ISO using dd.
[07:55] <sheptard> I'd try burning a CD if you can
[07:55] <cfhowlett> wkmanire, did you md5sum the .iso?
[07:55] <cfhowlett> sheptard, usb may not be supported on a 2000 machine
[07:55] <wkmanire> cfhowlett: I didn't see an MD5 checksum on the download page so I did not
[07:55] <cfhowlett> !md5sum
[07:55] <wkmanire> No CD drive available, and the BIOS list USB-HDD as a boot option.
[07:55] <sarnold> if it made it to grub, I'd hope that would be enough..
[07:56] <cfhowlett> sarnold, 60 seconds to verify integrity isn't too much to ask
[07:56] <wkmanire> I was reading in the official ubuntu server book that sometimes you have to change some boot args in GRUB to get things to load
[07:56] <sarnold> cfhowlett: no, indeed, that's a good idea; I was sresponding to the "usb may not be supported"
[07:56] <wkmanire> I'll go verify the MD5 checksum. I hope that is what's wrong.
[07:56] <cfhowlett> sarnold, :)
[07:57] <sarnold> wkmanire: if you hit shift of whatever it is to force grub to prompt, can you edit the kernel command line?
[07:57] <wkmanire> sarnold: One moment
[07:57] <sarnold> wkmanire: if so my first attempt would be adding something like vga=ask to the command line. that helped me back in the day on odd equipment
[07:59] <wkmanire> I calculated the checksum but I can't find the checksum on the ubuntu website for the iso I downloaded.
[08:00] <wkmanire> 2a846f77b21bddd732d75f7f574c3045
[08:00] <wkmanire> sarnold: I'm going to try getting into the boot options now
[08:00] <wkmanire> I have 6 more identical machines to do after this.
[08:01] <cfhowlett> wkmanire, wait one
[08:01] <wkmanire> cfhowlett: Ok
[08:01] <sarnold> sigh all I can find are the md5sums for the 12.04.4 release http://cdimage.ubuntu.com/releases/12.04.4/release/MD5SUMS
[08:02] <wkmanire> sarnold:  I'm using the 32bit LTS "server" release.
[08:02] <cfhowlett> wkmanire, 12.04.4 version?
[08:03] <cfhowlett> wkmanire, 12.04.4 ????
[08:03] <wkmanire> cfhowlett: Just a sec please
[08:03] <sarnold> wkmanire: every few months we press a new iso with updated kernel and so forth, 12.04, 12.04.1, 12.04.2, 12.04.3, 12.04.4, soon will come 12.04.5, and last will be 12.04.6 if I'm not mistaken...
[08:03] <wkmanire> cfhowlett: http://www.ubuntu.com/download/server/thank-you?distro=server&bits=32&release=lts
[08:03] <wkmanire> this one
[08:03] <wkmanire> I just downloaded it 20 minutes ago
[08:03] <sarnold> oh good :)
[08:03] <wkmanire> sarnold: I'm in the GRUB editor.
[08:03] <wkmanire> What did you suggest adding?
[08:04] <sarnold> wkmanire: vga=ask was the first thing that came to mind
[08:04] <wkmanire> What does that do?
[08:04] <cfhowlett> wkmanire, 12.04.4 386 server iso 14ad92270218a8925d802b3d3b6e140f
[08:04] <wkmanire> cfhowlett: Thank you, will confirm
[08:04] <sarnold> if I'm not mistaken it'll ask the kernel to promt you with a list of vga video modes, hopefully that'll show how much the kernel is alive or dead..
[08:04] <cfhowlett> wkmanire, http://releases.ubuntu.com/12.04.4/MD5SUMS
[08:05] <wkmanire> That's what i needed
[08:05] <cfhowlett> wkmanire, DOH!  bad info!  here's 32 bit server 0081e57fb8c7e4094fb9767384f087c6
[08:06] <wkmanire> I used md5sum to produce the value, but it looks like my checksum is different than your substantially
[08:06] <wkmanire> I'll redownload the ISO
[08:06] <wkmanire> yours*
[08:06] <cfhowlett> wkmanire, don't.  use torrents instead.  integrity checking is better
[08:07] <sarnold> I wish the silly download page made it easy to grab the SHA256SUMS and SHA256SUMS.gpg files from the same directory
[08:08] <wkmanire> Ok, I'm downloading the torrent
[08:08] <wkmanire> Holy crap, that downloaded in 10 seconds
[08:08] <sarnold> just the .torrent file? or the whole thing?
[08:09] <wkmanire> The whole thing!
[08:09] <wkmanire> At least one of the peers seems to be connected to a fire hose
[08:09] <sarnold> wow. what a connection. :)
[08:09] <cfhowlett> wkmanire, run the md5sum again
[08:11] <wkmanire> cfhowlett: https://www.refheap.com/52889
[08:11] <wkmanire> Why is my number different again?
[08:12] <wkmanire> sarnold: That option you suggested did not help.
[08:12] <cfhowlett> wkmanire, try direct download
[08:13] <sarnold> wkmanire: did you just get the blinking cursor again?
[08:13] <wkmanire> sarnold: Yes.
[08:13] <wkmanire> cfhowlett: It's going to be 5 to 10 minutes for the direct download
[08:14] <cfhowlett> wkmanire, do it.  check it.  I'm curious to know results so I'll be around
[08:14] <wkmanire> Ok
[08:15] <sarnold> wkmanire: okay, a few more ideas. (I'm running low.) you could also try removing the "splash" argument from the kernel command line; the 'splash' asks plymouth to start, it's the graphical thing that 'owns' the console. I dunno if its even on the server disks, but it might be worth a shot: https://wiki.ubuntu.com/Plymouth
[08:15] <sarnold> wkmanire: another thing I remember from machines of that era is turning off acpi or apic; I had one machine that I had to boot with apic=off
[08:15] <wkmanire> I'll do both and see what it does
[08:16] <wkmanire> This is a shining example of why you don't let developers do sys admin work :P
[08:16] <cfhowlett> LOL.
[08:17] <sarnold> lol
[08:18] <wkmanire> There is no splash argument to be found
[08:19] <sarnold> well, okay, that's at least one fewer thing to try :) hehe
[08:19] <wkmanire> sarnold:  Should I be adding these options anywhere in the boot script?
[08:19] <wkmanire> Or is there a particular line they should be on?
[08:20] <sarnold> wkmanire: hrm, i'd put them at the end of the kernel command line..
[08:20] <wkmanire> What does the kernel command line look like?
[08:20] <wkmanire> linux /boot/........
[08:20] <wkmanire> that line?
[08:21] <wkmanire> Or is it after the call to initrd?
[08:21] <wkmanire> cfhowlett: My checksum is correct now.
[08:21] <wkmanire> sarnold: I'm going to reimage this USB device and try again
[08:21] <sarnold> wkmanire: on my machine, near 'ro' and 'quiet' ...
[08:22] <jpds> wkmanire: The linux one.
[08:22] <sarnold> wkmanire: cool
[08:22] <cfhowlett> wkmanire, try installation
[08:22] <wkmanire> It's actually an SD card in a USB SD card reader
[08:22] <wkmanire> It's all I had on hand.
[08:23] <sarnold> oof :)
[08:24] <wkmanire> What an adventure. And I thought setting up netboot would be more work.
[08:24] <sarnold> wkmanire: oh. uh.
[08:24] <sarnold> wkmanire: does this CPU do PAE?
[08:24] <wkmanire> PAE?
[08:24] <sarnold> wkmanire: somewhere along the way we stopped supporting kernels without PAE, and I do'nt know when that was.
[08:24] <sarnold> wkmanire: extended page tables
[08:24] <wkmanire> I don't know
[08:25] <wkmanire> The image completed transfer so I'm going to try to boot again
[08:25]  * sarnold crosses his fingers
[08:26] <wkmanire> sarnold:  Legacy `ask' parameter no longer supported
[08:26] <sarnold> wkmanire: oh, pae was introduced in 95. wow: http://en.wikipedia.org/wiki/Physical_Address_Extension
[08:27] <sarnold> wkmanire: oh so now I'm legacy! lol
[08:28] <wkmanire> Whelp, I'm back in the same boat again.
[08:28] <wkmanire> I'm going to have to go dig up a CDROM drive
[08:28] <wkmanire> And then crack all of these boxes open :(
[08:29] <sarnold> wkmanire: try adding 'nomodeset'?
[08:31] <wkmanire> sarnold: Should I put that together with apic=off?
[08:31] <sarnold> wkmanire: you could, but it'd be nice to know if one specific thing is sufficient
[08:32] <wkmanire> Absolutely
[08:32] <sarnold> wkmanire: even though that's the most annoying and slowest way to get there :) heh
[08:32] <wkmanire> I wish I could get into a terminal so I could see the log output
[08:32] <sarnold> yes
[08:32] <wkmanire> trouble-guessing sucks
[08:33] <sarnold> i like that, i'll have to remember it :) thanks
[08:33] <sarnold> wkmanire: bed time for me, good luck, and if you sort it out, i'd like to find out what it was..
[08:34] <wkmanire> sarnold: I just tried the last possible combination of arguments and it also failed.
[08:34] <wkmanire> I *MIGHT* have a CDROM in a dead PC in the garage. So I'm going to go on a scavenger hunt
[08:34] <wkmanire> sarnold: I'll come back to the channel after I get this working
[08:35] <sarnold> wkmanire: good luck! thanks :)
[08:35] <wkmanire> sarnold: Thank you for the help. Good night.
[08:42] <wkmanire> In the grub terminal, lscpi brngs up one device as "Unkown Bridge"
[08:42] <wkmanire> Unknown*
[08:43] <wkmanire> Could that cause the installer to fail to load?
[08:49] <wkmanire> I think I found what the hold up is.
[08:49] <wkmanire> ACPI Warning: Detected an unsupported executable opcode at module-level: [0x007B]
[08:50] <wkmanire> There's two messages like that in GRUB recovery mode.
[08:53] <wkmanire> Yep.
[08:53] <wkmanire> Got it.
[09:41] <wkmanire> Did you know that a CD won't load if the CD tray is not 100% closed?
[09:41] <wkmanire> Who would have guess that it doesn't operate while open?
[09:43] <cfhowlett> wkmanire, yeah, they're funny that way ...
[09:48] <alimj> Hi everyone. I am looking for advises on recommended average e-mail box size.  I have to setup 100+ user e-mails for a client.
[09:49] <AtuM> alimj, I usually give users 1GB/mailbox
[09:49] <alimj> AtuM: Thanks
[09:50] <AtuM> alimj, that's to give admins time to react when some users go over limit - if users use imap
[09:51] <ikonia> alimj: you can't take a "recommendation" it depends on the spec/needs of the mail users and host much storage/backup you have available to you
[09:51] <alimj> I am going to forbid IMAP. Only POP3
[09:51] <AtuM> alimj, if users use pop3 and download mail to their desktops, then you can give much less space per mailbox
[09:52] <AtuM> alimj, I'd recommend setting 100-200MB/mailbox so that mailboxes don't get full when ppl go to vacation.. are you about to use quotas or just limit the space for all mailboxes?
[09:53] <alimj> AtuM: A combination of both. For different users, VIPs, etc
[09:53] <alimj> I am also looking into fetchmail and getmail daemons. For a local storage solution
[09:54] <alimj> ikonia: Thanks for suggestion. I am looking for a starting point
[09:55] <AtuM> alimj, what are users using as a client software?
[09:55] <alimj> AtuM: Will be outlook, thunderbird or mobile devices.
[09:55] <AtuM> alimj, how come you're not setting up a service like zimbra/zarafa?
[09:56] <alimj> AtuM: I am not familiar with those. I will look to them.
[09:57] <alimj> For VIPs, I was planning to setup POP on clients to leave mails on server for one week. So they can use IMAP on their mobiles to check their mails while traveling.
[09:57] <AtuM> alimj, is there not enough resources on server side? I use zimbra since all mailboxes get stored on a single server and it's then easy to backup, no worries about desktop mailbox when computers crash, and you get webclient acces from anywhere ;)
[09:58] <ikonia> alimj: you can't have a "starting point" with no requirements
[09:58] <spidernik84> hi! how am I supposed to update an interface ip address after changing it in the interfaces file? Ifup says the interface is already configured. If I append --force it will add a secondary address! And service networking restart is deprecated :(
[09:58] <alimj> AtuM, ikonia: Thanks for your time and answers. I have a visitor. I will be back. Thanks again.
[09:58] <spidernik84> ifdown first and ifup?
[09:58] <AtuM> but zimbra needs quite a lot of resources for 100+ users.. at least 16GB RAM + SSD disks on a single server (capacity requirements depend on user quotas)
[09:59] <AtuM> spidernik84, use ifconfig to reconfigure manually..
[10:01] <AtuM> spidernik84, the interfaces file will be used at the next boot..  you could also reboot the machine, but that's a bit of a waste of time imho
[10:06] <spidernik84> AtuM, thanks. We use a puppet module to modify the networking file but we designed it to trigger an /etc/init.d/networking restart after the change. I realized it's dangerous
[10:06] <spidernik84> it will take down the network for a fraction of time
[10:07] <AtuM> alimj, the approximation is difficult without any real requirements, as ikonia pointed out. it's best if you set up LVM2 and then leave some space in reserve.. mailboxes can get huge very fast - it all depends on what users are doing with it
[10:07] <spidernik84> so I was looking for a way to "refresh" the interfaces file (i need to add the dns records to it as suggested starting from ubuntu 12.04)
[10:08] <AtuM> spidernik84, aren't you changing ip ?
[10:09] <spidernik84> that's also a scenario, but less used
[10:09] <spidernik84> I'm mixing up things, apologies.
[10:10] <AtuM> spidernik84, oh.. dns records..  perhaps resolvconf might help
[10:11] <spidernik84> the ifup script is supposed to parse the dns-nameservers line in the interfaces file and tell resolvconf to use them
[10:11] <spidernik84> I'll do some tests. Thanks
[10:13] <AtuM> spidernik84, that's true. even if you restart network for that fraction of time it should not have any side-effects.. tcp/ip can handle it if the interruption is not too long
[10:13] <spidernik84> I agree. It's just that they say it's deprecated :(
[10:13] <AtuM> i've run ifdown && ifup within ssh session and I didn't get disconnected :)
[10:13] <spidernik84> lol nice :D
[10:14] <spidernik84> I'll try a bit
[10:27] <stetho_> Obscure question. My company purchased an HP StoreEasy 1630. It's a nice bit of kit with a built in 3Par array. However, it came with Windows Storage Server 2012. The overhead of setting up an Active Directory server with UNIX support, migrating existing users and their UIDs/GIDs and maintaining everything in two places has become an absolute pain. The HP website lists "Linux" as a supported operating system. Does anyone know if I'm going to hit
[10:27] <stetho_>  any problems (in terms of drivers etc) if I flatten the box and install 12.04 on it?
[10:28] <mardraum> I'd say it depends on storage server to work
[10:29] <mardraum> but hopefully HP do actually have a solution. You are going to need to raise a ticket with them and try escalate until you get someone who knows. They do have some talented storage peeps.
[10:30] <rbasak> hallyn: FYI, bug 1287140 - when I fix it, we'll need to modify your auto-package-testing MP a little bit, to add a --ssh-private-key-file option to when you call "uvt-kvm wait"
[10:40] <AtuM> stetho_, why don't you ask HP? it's a new machine and they should have a great customer service support..  will it be used as a CIFS, NFS or ISCSI ?
[10:41] <stetho_> mardraum: Thanks - I just had a lightbulb moment and checked the HP site for Linux drivers and couldn't find any. :-(
[10:42] <AtuM> stetho_, with windows storage server it's all "point-and-click". when u use linux as a storage os, you really have to dive into each technology.. there's no "point-and-click" configuration.
[10:42] <stetho_> AtuM: That was my next port of call. Google didn't show anything so I thought I'd ask in here. Usually full of knowledgeable people. The answer is "All 3" - I just want to remove the AD part from my mainly Ubuntu and a bit of OS X network.
[10:43] <stetho_> The only option the device gives in its current setup to get round AD is to create NFS shares and mount them as root.
[10:44] <AtuM> stetho_, community can help, but you should not use any step-by-step advice - you should dive into each of those services and understand the inner workings.. linux does not complain when you mess up your config - it will try to make it work as you set it up.. even if all data goes away.. you should know what you're doing.
[10:47] <AtuM> stetho_, our company also makes iscsi/infiniband storage devices.. but we've never made "step-by-step" documentation because of those facts. if you're willing to take the time and learn, then I say go for it. it's all pretty simple and straightforward, but it takes some skill to manage
[10:48] <stetho_> Atum: I've got no problem with doing it unless I fall at the first hurdle - not being able to recognise the RAID controller.
[10:49] <AtuM> stetho_, if you can see the disks as JBOD devices, you can use software raid.. if anything, you can boot some "server/enterprise" distro to see if linux sees the raid devices
[10:50] <AtuM> stetho_, I can't really tell you more, since I have no idea where the HP stores the OS on this model.. it could be a CF card, and internal HDD... who knows
[10:51] <AtuM> stetho_, usualy the disks that are hot-pluggable do not store the OS (firmware).. so there must be some other media to install to
[10:53] <stetho_> AtuM: If I have to put a disk in to install the OS that's fine - I can live with that. I've just found something on the HP site that says the P822 (which is what's in this machine) is supported by RHEL and SLES so it's not a total non-starter.
[10:55] <jamespage> Daviey, did you have a chance to review the pxc packages in NEW yet?
[10:55] <AtuM> stetho_, the latest devices are more quickly supported by linux then by windows.. i see that a lot these days
[10:57] <AtuM> stetho_, HP usually uses somewhat trusted and well supported components.
[10:59] <AtuM> stetho_, you know that installing linux on this machine will brick its warranty
[11:00] <AtuM> stetho_, the part where the documentation says "compatible with windows, linux.." is for the client part, not for the machine itself
[11:03] <stetho_> Yes - I realised that (the client part) which is why I went of looking for drivers.
[11:05] <AtuM> stetho_, iscsi, nfs, cifs don't need drivers
[11:06] <stetho_> I know - this is just one of those "it should work" things. Like you say, if I do it I brick the warranty and if it fails it's an expensive failure.
[11:07] <stetho_> Maybe I need to investigate a better way of integrating AD in to my network. If AD had been here first it would have been much easier. Adding Linux to AD is a lot easier than adding AD to Linux.
[11:07] <AtuM> stetho_, linux includes drivers for the most hw.. they come as modules.. it's rare that you need special drivers from vendors.. it can happen, but it's rare
[11:09] <AtuM> stetho_, integrating AD into linux environment is pain.. how come the company bought this box?
[11:10] <wmp> hello, i have problem with MdRaid
[11:10] <wmp> on normal ubuntu boot, i have error with degradated raid
[11:10] <wmp> on resovery all works good
[11:11] <AtuM> stetho_, as you pointed out.. setting up AD, unix support, user mapping.. i am playing with this too these days :)
[11:11] <AtuM> wmp, does mdraid get degraded after each reboot?
[11:12] <wmp> AtuM: yes
[11:12] <AtuM> wmp, it's a bug.. what ubuntu version do you use?
[11:12] <wmp> always on normal boot i have degrated raid, always on recovery raid os good
[11:12] <wmp> 12.04
[11:12] <wmp> updated
[11:13] <AtuM> can you check this: http://www.fractalizer.ru/frpost_350/warning-mismatch_cnt-is-not-0-on-devmd0/
[11:13] <AtuM> if mismatch_cnt is not 0, then you might have a disk failure
[11:15] <stetho_> In short, we were a small company, we were going to move to AD, then everything went mad because someone had a good idea, we quadrupled in size in 6 months and then got acquired. Now things have settled down again I have the opportunity to do things "properly" (Puppet, monitoring etc) and I wanted to remove (or improve) this administrative overhead/pain it the butt.
[11:15] <wmp> AtuM: i have 0
[11:16] <wmp> AtuM: this is my grub.cfg config: http://wklej.org/id/1291241/
[11:21] <wmp> AtuM: and this is screenshots from KVM: http://wstaw.org/m/2014/03/06/makeconfig6.png http://wstaw.org/m/2014/03/06/makeconfig7.png
[11:24] <wmp> AtuM: ok, my server has efi...
[11:35] <wmp> AtuM: maybe bug is in kernel, becouse on old kernel all works good
[11:35] <wmp> 3.11.0-17 dont works
[11:36] <wmp> 3.11.0-15 works
[12:11] <AtuM> wmp, I'd say it's a bug.. I've hit this one on fedora a year ago.. when they've merged dmraid and mdraid with dm-mapper.. the problem apparently lies with the upstart/systemd shutdown/reboot procedure, where the dm-mapped doens't flag the mdraid as "clean".. therefore it goes into resync each time the system boots up
[12:11] <wmp> AtuM: no, this is bug on kernel, on older version all works good
[12:12] <AtuM> wmp, then it might have something to do with its module for dm-mapper.. it's part of the kernel
[12:12] <wmp> AtuM: i dont know, now i reporting this bug on launchpad
[12:14] <AtuM> anyway I haven't found the solution for it and got tired of waiting for a fix.. this bug ate my nerves across two fedora releases.. when i got tired of it i switched distros ;)
[12:14] <wmp> AtuM: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1288703
[12:15] <AtuM> I must say that I haven't had these issues on ubuntu yet..  if old kernel works for you, I'd suggest u use the old version until the fix comes along.. 14.04 will come soon enough ;)
[12:15] <wmp> yes, of course ;)
[12:16] <AtuM> wmp, great.. however I do believe that 12.04 might become obsolete before they fix it.. so my hopes are on 14.04
[13:26] <Daviey> jak2013: pxc = percona?
[13:26] <Daviey> err, jamespage
[13:27] <Daviey> jamespage: I assume it does... What is the status of it getting in Debian?
[13:33] <jamespage> Daviey, its in the NEW queue in Debian as well
[13:33] <jamespage> the packages in the Ubuntu NEW queue are the Debian packages with a 0ubuntu1 version number
[13:50] <zul> Daviey:  pingaroo
[13:52] <zul> Daviey:  so ceilometer added a new dependency yesterday (right before a release yay) called python-croniter, i packaged it last night, can you review it for me?
[14:04] <zul> Daviey:  https://bugs.launchpad.net/ubuntu/+bug/1288761
[14:31] <hallyn> rbasak - i'm giving up on the idea of ditching vmbuilder before trusty
[14:32] <hallyn> taking too long, not gonna happen.  resigned to spending some time maintaining it.
[14:33] <rbasak> hallyn: ok. Though I hope it's not because of the uvtool bug. I can rush a fix if that will help you.
[14:33] <hallyn> so the adt conversino should still happen, as it takes a lot of fragile code out, but i don't want to rush it
[14:34] <hallyn> (and if someone else does, they can take my MP and run with it)
[14:35] <hallyn> rbasak: no, it's not just that bug.  if you l ook at the pace of the open bug for that MP, you'll see it lives on a slower timeline :)
[14:36] <rbasak> OK. If anything changes and you need a fix quickly, let me know. Your specific use case is a small part of the general problem but easy to fix permanently, but I would prefer to sort out the other aspects of it while I'm there otherwise.
[14:37] <Daviey> zul: hey
[14:37] <zul> Daviey:  heylo
[14:39] <hallyn> rbasak: ok.  would prefer a proper fix, which is exactly why i don't want to push the mp :)
[14:39] <hallyn> Daviey: zul: hey!  (can anyone join this party?)
[14:40] <zul> hallyn:  no you arent invited :)

[14:42] <hallyn> oh wait, if forgot i'm a loner :)
[14:42] <Daviey> It's the no hallyn's party... Which on this occasion means we are allowed one.
[14:42] <hallyn> that's confusing
[14:43] <Daviey> (simpsons joke)
[14:43] <hallyn> oh.  one day i'll sit down and start watching that at season 1
[14:44] <Daviey> I wouldn't bother. :)
[14:48] <smoser> hallyn, hey.
[14:49] <smoser> um. is there a way to lx-attach without the thing running ?
[14:49] <smoser> i just want to do this:
[14:49] <smoser> $ sudo lxc-attach -n test1 /bin/bash  < myscript
[14:49] <zul> brb need to reboot
[14:50] <smoser> and have it basically chroot me into that thing so that myscript can run and do something to that root
[14:51] <hallyn> smoser: no
[14:52] <hallyn> attach means attach to a running container
[14:52] <hallyn> now, you could perhaps do 'lxc-execute -n container -- sleep 1200'
[14:52] <hallyn> but generally lxc-start -n container -d goes fast enough
[14:52] <smoser> hm..
[14:53] <smoser> ah.
[14:53] <smoser> but i guess i could just lxc-start
[14:53] <smoser> with a command of /bin/bash ?
[14:56] <smoser> hallyn, the goal would be to do something like this:
[14:56] <smoser>  lxc-create -n original-trusty
[14:56] <rbasak> hallyn: the fix for you *would be* a proper, permanent fix; it's just that it's a subset of a bigger issue. I can do the subset first if need be, that's all.
[14:56] <smoser>  lxc-clone -o original-trusty -n trusty-test
[14:57] <smoser>  lxc-<something> -n trusty-test < my-update-script
[14:57] <hallyn> smoser: yeah that should be fine
[14:57] <hallyn> smoser: you just need the namespaces to exist
[14:58] <smoser> i'm confused.
[14:58] <smoser> on a clone of overlayfs (or anything) i want it to do all the mounts and such.
[14:58] <smb> jamespage, ping, nearly forgot you wanted to finish up iscsitarget.
[14:59] <jamespage> smb, yeah - so I think it generally all looks OK for  straight backport
[14:59] <jamespage> smb, I think we need todo it for all releases prior to trusty right?
[14:59] <hallyn> rbasak: oh, in that case, i guess please make your best judgement based on how long you think it'll take.
[14:59] <smb> jamespage, Not, sure. isn't the userspace side part only required for precise
[15:02] <smb> I mean the main reason for the backport is to allow HWE kernels which are newer to work with the dkms coming from the user-space package. All the releases in between on their own do not get newer kernels
[15:05] <rbasak> kirkland: see ubuntu-devel for someone in need of some entropy :)
[15:05] <rbasak> (I hear you have a supply?)
[15:09] <jamespage> smb, agreed - but I'm not sure exactly how the versioning would work
[15:09] <jamespage> if a 12.04 user with the newer iscsitarget upgraded to quantal say
[15:09] <jamespage> they would stick with the package from precise
[15:09] <jamespage> as its a newer version than the one in quantal
[15:16] <smb> jamespage, Hm yeah that is true. But on the other hand the dkms build has multiple patch stages by kernel version. So the new version should work with anything in between. But there is the risk of it being confusing
[15:16] <jamespage> smb, that was why I was proposing backporting everywhere if that makes sense
[15:17] <kirkland> rbanffy: :-)
[15:18] <kirkland> rbasak: :-)
[15:18] <smb> jamespage, Yeah. Though if we drag our feet a little bit more, there isn't much left in between :)
[15:18] <kirkland> rbasak: looks like a fair request, to me
[15:19] <smb> jamespage, Right now it would only be Quantal if I am not wrong. ITs backported from Saucy and Raring is already out of support
[15:19] <rbasak> kirkland: sure. I know little about ubiquity. For d-i, I'd suggest preseeding (and enhancing the cryptsetup udeb) for some of his needs, and a pollinate udeb
[15:20] <xnox> you want to polinate ubiquity?
[15:20] <kirkland> rbasak: the pollinate prng seeding is currently ubuntu cloud (server) images only
[15:20] <rbasak> Though, other, hardware sources may be better, for someone who really cares about this.
[15:20] <kirkland> rbasak: yeah, I don't see any point, really, in bringing that up on his thread
[15:21] <kirkland> rbasak: it's pretty clear to me that he wants more than just an improved prng seed
[15:21] <jamespage> smb, ah - yes - I forgot
[15:21] <jamespage> I was thing we where backporting from trusty
[15:24] <smb> jamespage, I was doing that so long ago (at least feels that way) that now we could go there. Though I am hoping that the user-space changes between S and T are not having effects on the dkms part. So hopefully to allow the T kernel in P we only have to pick some patches there.
[15:30] <gQuigs> could I get my blueprint accepted by one of the server track leads: https://blueprints.launchpad.net/ubuntu/+spec/servercloud-1403-networking ?
[15:37] <rbasak> gQuigs: could you start a mailing list thread on that, maybe? I have no objection to having a vUDS session, but I have immediate "why not just ..." responses to much of what you've said. Perhaps a ML thread could cover the basics, so as not to waste a session doing that?
[15:37] <rbasak> (BTW, I can't approve blueprints myself)
[15:38] <gQuigs> rbasak: there have been several internal Canonical discussions..
[15:39] <gQuigs> rbasak: it was last discussed at a team sprint (that I wasn't a part of) and the next step was to do a public discussion
[15:39] <rbasak> gQuigs: sure, OK. I'm proposing that a public discussion start on a mailing list in advance of a UDS session, because that makes a session more useful.
[15:42] <gQuigs> rbasak: recommend ubuntu-devel or ubuntu-server?
[15:42] <rbasak> gQuigs: up to you - I'm not sure of the context yet! If it's a server blueprint, then ubuntu-server makes sense. If it's a wider thing, then ubuntu-devel, but then maybe it should be a foundations blueprint?
[15:43] <jamespage> rbasak, gQuigs: if its networking post ubuntu-devel
[15:43] <gQuigs> jamespage: will do
[15:58] <zul> Daviey:  thanks
[16:33] <zul> jamespage:  https://code.launchpad.net/~zulcss/neutron/new-plugins/+merge/209722
[16:33] <jamespage> zul, missing new files
[16:34] <zul> jamespage:  there now
[16:45] <jamespage> zul, neutron update needs some work still
[16:46] <zul> jamespage:  ok ill have a look at it after lunch
[16:51] <jamespage> zul, doing horizon now
[16:51] <jamespage> coreycb, use this channel for co-ordination pls
[16:52] <coreycb> jamespage, ok
[16:55] <jamespage> zul, coreycb: horizon done
[16:58] <jamespage> zul, coreycb: doing keystone
[17:02] <zul> jamespage:  ack
[17:03] <jamespage> zul, as most of these are fairly trivial, going straight for push and upload rather than review first
[17:03] <zul> jamespage:  ditto
[17:03] <zul> nova is done
[17:03] <jamespage> anything non-trivial I'll MP
[17:04] <jamespage> coreycb, obviously you'll still need to raise a MP
[17:04] <jamespage> zul, I pointed coreycb at cinder to start with
[17:04] <zul> jamespage:  ack...can you both update the list as you as well?
[17:05] <jamespage> zul, will do
[17:05]  * jamespage twiddles his fingers while keystone tests itself
[17:27] <zul> jamespage:  building swift
[17:32] <Repox> Hello. I have an executable which relies on me typing in my choices in steps. Is it possible to pipe these choices? Like 'echo "1" | echo "myemail@example.com" | etc | ./create_user.sh' ? (Already tried that, didn't work).
[17:38] <smoser> rbasak, https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1235231 is going to be fixed!
[17:41] <tych0> hi hallyn, i'm getting: http://paste.ubuntu.com/7045342/
[17:41] <tych0> any ideas?
[17:41] <tych0> just from doing an lxc-start
[17:43] <hallyn> tych0: ?  no...  so you start the contaienr, and it immediately shuts down?
[17:44] <tych0> yeah
[17:44] <hallyn> look through syslog.  oom maybe?
[17:44] <hallyn> things being indiscriminately TERMed says oom to me
[17:45]  * tych0 looks
[17:51] <tych0> hallyn: http://paste.ubuntu.com/7045385/ so it looks like it gets an IP
[17:51] <tych0> then the bridge device it is on does something weird
[17:56] <rbasak> smoser: \o/
[17:57] <coreycb> jamespage, zul: cinder started
[17:58] <hallyn> tych0: but anything in syslog or kern.log?
[17:58] <zul> coreycb:  awesome
[17:59] <tych0> hallyn: that was syslog, trying to recreate and will check kern.log
[18:10] <jamespage> zul, coreycb: keystone uploading
[18:10] <jamespage> zul, found one bug in horizon
[18:10] <zul> jamespage:  heh
[18:10] <jamespage> well two actually
[18:10] <zul> jamespage:  ok..working on ironic
[18:10] <jamespage> https://bugs.launchpad.net/ubuntu/+source/horizon/+bug/1288904
[18:10] <jamespage> zul, and the ubuntu theme is really stuffed due to new theme from upstream
[18:11] <zul> ack
[18:11] <jamespage> zul, I'll see if I can wrangle some design team for that
[18:12] <zul> jamespage:  that would be great...since the quality of my css skills is umm...crappy :)
[18:12] <jamespage> ditto
[18:12] <jamespage> gaughen, ^^ fyi
[18:19] <rostam> hi I am using ubuntu 12.04 update 4. The output of ifconfig displays an interface which I am not sure what it is:   ax0  192.168.4.1   what is ax0 interface? thx
[18:20] <coreycb> jamespage, zul: cinder started for real this time
[18:21] <coreycb> :)
[18:21] <zul> sweet
[18:37] <coreycb> zul, jamespage: cinder pushed and merge requested
[18:38] <jamespage> coreycb, just leave the reviewer blank in future - it will default to the team
[18:38] <jamespage> zul, you ok to land that?
[18:39] <zul> jamespage:  yeah i can
[18:42] <gQuigs> rbasak: jamespage, can either of you let my message in to ubuntu-devel :)
[18:48] <coreycb> jamespage, ok
[18:48] <coreycb> zul, jamespage: starting python-swiftclient
[19:21] <coreycb>  zul, jamespage: python-swiftclient pushed and merge proposed
[19:21] <zul> coreycb: cool ill have a look
[19:21] <coreycb> zul,  thanks!
[19:22] <zul> cinder uploaded
[19:26] <MavKen> can someone recommend a tutorial for getting mail server and a web interface up and running on a vps that has several virtual hosts?
[19:29] <jamespage> zul, https://launchpadlibrarian.net/168544611/buildlog_ubuntu-trusty-i386.keystone_1%3A2014.1~b3-0ubuntu1_FAILEDTOBUILD.txt.gz
[19:29] <jamespage> gah
[19:30] <zul> jamespage:  gah...
[19:30] <zul> swine testr
[19:30] <jamespage> zul, its the inactivity caused kill thing
[19:31] <jamespage> zul, maybe run_test is noisier
[19:31]  * jamespage looks
[19:31] <zul> jamespage:  yeah we need to make it more verbose
[19:32] <zul> jamespage:  coreycb found a bug in the python-heatclient package
[19:32] <zul> jamespage:  we'll keep him ;)
[19:33] <coreycb> zul, lol
[19:35] <DeltaHeavy> Hey, so I'm trying to configure netfilter with iptables and I'm wondering how I can insert a rule, between two othe rules. As far as I can tell this is impossible?
[19:37] <sarnold> DeltaHeavy: try the iptables -I with the optional [rulenum] parameter
[19:42] <zul> jamespage:  ceilometer needs python-pysnmp4 now :(
[19:43] <coreycb> zul, fix pushed
[19:43] <zul> cool
[19:48] <zul> coreycb: heatclient needs fixing
[19:49] <gQuigs> rbasak: posted https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2014-March/014877.html
[19:50] <DeltaHeavy> sarnold: Awesome, thanks!
[19:50] <sarnold> DeltaHeavy: you're welcome :) have fun!
[19:51] <coreycb> zul, ok I can do that.  I was going to do 2 different commits.
[19:51] <zul> coreycb: nah just quash them
[19:51] <coreycb> zul, gotcha, will do
[19:54] <wmp> zul: you are from Poland?
[19:55] <zul> wmp: no im in canada
[19:55] <wmp> zul in poland mean a habitual drunkard, an alcoholic
[19:56] <wmp> zul: http://tiny.pl/qlfr4 :D
[19:59] <zul> oh
[20:31] <shredding> Hello all.
[20:31] <shredding> I'm writing my first shell script.
[20:32] <shredding> I'm trying to hook into the cd-function in order to activate a ruby gemset on specific directory or activate venv for python.
[20:32] <shredding> this is what i got: http://pastebin.com/dJqXaqVk
[20:33] <shredding> It works, in terms of it tells me that it activates the gemset and that it activates venv on the correct directories ...
[20:33] <shredding> ... but they are not really applied.
[20:33] <shredding> I have to add gemsets / venv again.
[20:33] <shredding> What may be the reason?
[20:33] <shredding> Do i have to delay or so?
[20:42] <rostam> hi I am using ubuntu 12.04 update 4. The output of ifconfig displays an interface which I am not sure what it is:   ax0  192.168.4.1   what is ax0 interface? thx
[20:43] <patdk-wk> ax.25
[20:51] <rostam> patdk-wk,  how that could be plumbed on my system?
[20:52] <patdk-wk> how should I know? I didn't configure your system
[21:03] <rostam> The output of the "axi-cache depends libssl1.0.0" has some lines starting with "Breaks:"  what that means? thx
[21:16] <patdk-wk> rostam, ask any webpage about how debian packaging works
[21:16] <patdk-wk> what it isn't compatable with
[21:30] <rostam> patdk-wk, thx
[21:37] <sarnold> rostam: https://www.debian.org/doc/debian-policy/ch-relationships.html
[21:53] <justizin> is it just me, or is it awkward that the docker package has been renamed from lxc-docker (upstream) to docker.io (ubuntu maintained) in trusty?
[22:06] <rostam> sarnold, thx
[22:51] <med_> smoser, is it safe to say that cirros-0.3.1 isn't built with the virtio driver (and thus can't respond to pings if libvirt/kvm are using virtio)
[22:51]  * med_ is pretty sure that's the case
[22:51] <med_> or rather that the cirros KERNEL isn't built with it
[22:53] <med_> hmmm, looks like it is being built and tested with virtio
[22:53] <med_> (yet it doesn't have the side effect of actually working.)
[23:36] <apb1963> I want to take system (ubuntu 12.04) generated mail and be able to read it with thunderbird.  So I'm trying to use my (residential roadrunner) ISP as a relay to send it to my normal user account.  Since I'm on a dynamic IP, I fall under RR spamhaus policy and it (refused to talk to me: 554 5.7.1 - ERROR: Mail refused).  RR used to have a writeup on how to get around this, but they moved the page and of course didn't leave a pointer.
[23:38] <apb1963> I seem to recall it may have had something to do with SMTP auth, but I don't remember for sure.  I'm using postfix & dovecot
[23:39] <apb1963> to be honest, I didn't want this to be a major fuss... I just wanted to be able to read mail in the manner described.
[23:44] <sarnold> apb1963: they may also support pop-before-smtp or imap-before-smtp, maybe that will help you find their documentation?
[23:54] <apb1963> sarnold: Thank you.  The only thing I can find is something about entering the username properly.  Perhaps I'm asking the wrong question.  The mail is destined for a user (me) at that domain.  The fact that it's my account I don't believe to be relevant.  I just want to send mail to one of their users.
[23:55] <sarnold> apb1963: yeah, sometimes the ability to send mail from a specific account is dependant upon loging in as that account from the same IP address a few seconds earlier
[23:55] <sarnold> (logging in over pop3 or imap4)
[23:57] <apb1963> sarnold: I can pickup and send mail through thunderbird just fine... it's only when I try to send system mail by way of postfix that it fails.  I suspect it may have to do with the fact that I'm sending mail as another domain name.  i.e. My dynamic host is registered with freedns to be another domain... at the RR supplied ip address.
[23:57] <apb1963> pop3
[23:57] <apb1963> but that's just a wild guess.
[23:59] <apb1963> sarnold: ok lets look at this from a different angle... how does a normal everyday system send mail to a RR user?
[23:59] <sarnold> apb1963: ah, I hadn't realized that the From: or From_ address didn't match the account in question..
[23:59] <sarnold> apb1963: by having a mail server not on a dial-up line or modern equivalent..