[00:31] <andygraybeal> when i login with ldap auth onto ubuntu.. my home directory is getting made with the wrong permissions :(  i don't know how to even begin to tell you what i need to for you to help trouble shoot :)  just commiserate with me for now!
[00:37] <andygraybeal> gah
[00:38] <sarnold> andygraybeal: which permissions does it get? which permissions would you like it to get instead?
[00:38] <andygraybeal> i found the problem
[00:38] <sarnold> andygraybeal: is there anything educational in /var/log/ that might indicate a config to change?
[00:38] <sarnold> oh yay :)
[00:38] <andygraybeal> two of the users in my ldap directory have the same UID number.... because did an import wrong
[00:39] <andygraybeal> err.. i said the wrong thing.. i meant 'owner' and not permissions
[00:39] <andygraybeal> thank you for responding
[00:40] <sarnold> ugh, I've even seen that one before, mismatched local vs ldap-provided uids... ls -l always looked fine, but ls -ln showed te problem clearly....
[00:40] <andygraybeal> ls -ln ..hmmmmm
[00:41] <andygraybeal> aah you rock, thank you for that
[00:41] <andygraybeal> n means numbers?
[00:41] <sarnold> yeah
[00:41] <andygraybeal> nice
[00:41] <sarnold> similar to netstat or some other commands that could look up numbers to names or not...
[00:41] <andygraybeal> okay, i'm not usd to the convention
[00:41] <andygraybeal> i'm still new-ish
[00:42] <andygraybeal> even tohugh i've used this for more than 10 yrs
[00:42] <sarnold> well, like many things on unix, you learn one and find it's different elsewhere :)
[00:42] <sarnold> ah, then you're already familiar with that, hehe :)
[00:43] <andygraybeal> well really i'm 34.. i startd in highschool-ish with bsd of some sort
[00:43] <andygraybeal> anyway.. it's all good
[00:43] <andygraybeal> i'm still totally new :)
[00:44] <andygraybeal> this is my first time trying to configure ldap auth client logons :)
[00:44] <andygraybeal> kerberos is still another bag of worms to me.
[00:44] <andygraybeal> even though i know i need to understand it.
[00:51] <bmorriso> I've been told that growing the root volume on an EC2 instance takes hours, if impossible...is this true? Any assistance/guidance/etc?
[00:53] <andygraybeal> i have an ec2 instance too :)
[00:54] <andygraybeal> what storage container type do you have?
[00:54] <andygraybeal> i have the simple basic one
[00:54] <andygraybeal> i forget what it's called
[00:59] <andygraybeal> amazon naming scheme is a little rediculous
[01:02] <Patrickdk> bmorriso, where is this root volume located?
[01:03]  * Patrickdk always found changing out a ebs only took minutes
[01:03] <andygraybeal> yea, i think i have ebs
[01:03] <Patrickdk> ebs is the default in ubuntu image
[01:04] <Patrickdk> ebs is not a simple/basic one though :)
[01:04] <Patrickdk> but it is, the safer one
[01:05] <bmorriso> Patrickdk, it is ebs /dev/sda1
[01:05] <bmorriso> it's only 8G and needs to be grown
[01:07] <cheese1756> Does anyone have a good script to rsync an entire Ubuntu server in full?
[01:07] <cheese1756> For super-easy referencing or restore
[01:11] <andygraybeal> okay, i think the owner is set correctly on my home folder, now i'm getting an auth error saying 'nologin' where do i change this to say /bin/bash (or something... i'm assuming)
[01:15] <Patrickdk> andygraybeal, chsh?
[01:16] <andygraybeal> Patrickdk, hmm.. okay
[01:16] <andygraybeal> i'm trying to login auth from a ldap server.
[01:17] <andygraybeal> so i'm a little confused.
[01:17] <Patrickdk> chsh?
[01:17] <Patrickdk> if it doesn't work, your ldap config is highly unusual
[01:17] <andygraybeal> no, i understand.
[01:18] <andygraybeal> :)
[01:18] <andygraybeal> i'm sure it is unusual
[01:19] <andygraybeal> it says the 'user' doesn't exist in passwd
[01:19] <andygraybeal> it's an ldap user... in my ldap config i have an entry: 'loginshell' it's value is /bin/bash
[01:21] <andygraybeal> i've never ever been able to login with ldap client auth in my life.... if i get this it will be an event.
[01:21] <Patrickdk> login with ldap client?
[01:21] <Patrickdk> you just need to configure your pam correctly
[01:22] <andygraybeal> ah pam
[01:22] <andygraybeal> pam i sso evil
[01:22] <andygraybeal> thank you for the hint though
[01:22] <sarnold> definitely underdocumented..
[01:37] <andygraybeal> i apologize, i'm reading this: https://help.ubuntu.com/community/LDAPClientAuthentication
[01:37] <andygraybeal> it doesn't say anything about a shell
[01:38] <Patrickdk> andygraybeal, that works, but it gives you no mangement, except doing it directly via ldap
[01:38] <Patrickdk> atleast I think that is how it is
[01:38] <andygraybeal> i have an entry in ldap, 'loginshell' = '/bin/bash'  .. i hoped that'd cover it :(
[01:38] <andygraybeal> but ihave no idea, i'm grasping :)
[01:39] <andygraybeal> i thin everythign is or should be managed from my ldap :)
[01:39] <Patrickdk> no, that isn't the issue
[01:39] <Patrickdk> the issue is, do you want it to be read only
[01:39] <Patrickdk> or do you want normal tools to update it, like passwd, chsh, adduser, ...
[01:40] <Patrickdk> if not, you will have to use or create your own user admin system
[01:40] <andygraybeal> well, i am using zentyal as my backend.
[01:40] <andygraybeal> it's got user mgmt built in.  if your not familiar with it.. it's a webified server admin thing.
[01:40] <Patrickdk> personally I perfer to build them myself
[01:41] <andygraybeal> i know, i know, everyone does.   i'm not exactly smart.
[01:41] <andygraybeal> i'm not exactly sure how this is going to roll... i was hoping to give it a test drive
[01:42] <andygraybeal> see how i feel about it
[01:43] <andygraybeal> any thoughts you have would be helpful, if you have any more.
[01:43] <andygraybeal> omg.. it works
[01:44] <andygraybeal> i don't know why it didn't work for a second there.. but it is working now.. holy crap!!!!!
[01:44] <andygraybeal> this is the first time i ever got this to work!
[01:44] <andygraybeal> how exciting
[01:44] <andygraybeal> i don't understand any of it :)
[02:31] <andygraybeal> anyway, i'm celebrating :)
[03:18] <InteliWasp> i have a server loaded with 13.04 beta 2, just installed vsftpd and enabled local user (after restarting service)but i am not able to log in
[03:58] <phelps> Logging into my server and it has started being very slow to give me a prompt
[03:59] <phelps> the last line from the ssh login is
[03:59] <phelps> debug1: Entering interactive session.
[03:59] <phelps> so it isn't actually connecting thats slow
[04:00] <phelps> happens when I su to another user as well
[04:01] <phelps> about 10 seconds
[04:07] <phelps> it happens after the motd
[04:33] <EvilEyeCorp> How would anyone recommend learning more about ubuntu server, are there any lessons or tutorials online I should visit?
[04:34] <ScottK> EvilEyeCorp: The server guide (mentioned in the channel topic) is a good place to start.
[04:34] <EvilEyeCorp> lol thanks
[04:35] <ScottK> You'd be amazed how often people just assume some random page on the internet is better than the actual documentation.
[04:35] <ScottK> Don't worry about it.
[04:37] <EvilEyeCorp> Yeah I just dove into it about a week ago and I am just getting the basics down
[05:35] <hammommah> greetings I am having some troule connecting via ssh from outside my local network. I have port 22 forwarded. I am currently running 4 csgo servers and a minecraft server. I can ping my server from outside. I can ssh into server from within my local network no worries. Any suggestions on where to look. When I try to connect from outside I get "No route to host"
[06:25] <koolhead17> hi all
[07:15] <Olivier1977> My name is Olivier and since a couple of day my Ubuntu Server 12.10 is having a problems
[07:16] <Olivier1977> I have a headless server which I could reach with putty
[07:16] <Olivier1977> I had it configured as a LAMP server
[07:16] <Olivier1977> with Webmin configured on
[07:17] <Olivier1977> When I connect a screen I see the server booting (RAM tests, disk mounting, ... all going fine)
[07:18] <Olivier1977> but after that nothing
[07:18] <Olivier1977> no connection with any tool anymore ftp,putty,webmin not reachable
[07:19] <Olivier1977> Can anybody help me on how to debug this issue
[07:19] <Olivier1977> This is the best I can tell about my problem, if more info is needed feel free to ask
[07:21] <Olivier1977> my machine is a SIEMENS primergy econel 100
[08:50] <jamespage> smoser, roaksoax: https://wiki.ubuntu.com/RaringRingtail/TechnicalOverview#Ubuntu_Server
[08:50] <jamespage> can you guys take a run at MAAS and simple streams please
[12:12] <AtuM>  How can i find out what's causing my system to "block" tty1 on ubuntu server 12.10- there's no login screen after bootup. i can type but i can login only by accessing some other tty
[12:13] <AtuM> the last output is "Skipping profile /etc/apparmor.d/disable: usr.sbin.rsyslogd "
[12:22] <AtuM> I'll try to disable update-motd on bootup...
[12:28] <AtuM> it helps not.. gets stuck right after apparmor or during
[12:29] <ogra_> check your upstart logs
[12:30] <ogra_> (and check if the tty1 job is even there etc)
[12:42] <zul> jamespage:  ping https://code.launchpad.net/~zulcss/nova/nova-final/+merge/160626
[12:49] <jamespage> zul " * debian/control: Add iptables dependency."
[12:49] <jamespage> I thought the missing dep was ebtables?
[12:49] <jamespage> or was that a typo?
[12:49] <zul> its for nova-api
[12:50] <zul> jamespage:  its in the api-metadata.filters
[12:51] <jon__> hi, having trouble getting apache running on ubuntu server
[12:51] <jon__> do i need to change any permissions for the default setup to work
[12:52] <jon__> using var/www
[12:55] <jpds> jon__: No, it should just work.
[12:55] <jon__> hmm
[12:55] <jpds> jon__: Have you looked at /var/log/apache2/error.log ?
[12:55] <jon__> ok
[12:55] <jon__> no
[12:55] <jon__> i will
[12:56] <jon__> basically i have cloud server with rackspace
[12:56] <jon__> i installed apache
[12:56] <jon__> pointed my nameservers at rackspace dns
[12:57] <jon__> but when i go to required domain in broswer get an error saying firefox cant find server
[12:58] <jon__> ive used dig / whois etc and seems ok there
[12:58] <jon__> so thinking its my apache setup
[12:59] <pmatulis> jon__: does it work by ip address?
[12:59] <jamespage> zul, +1
[12:59] <zul> jamespage:  thanks
[12:59] <zul> just giong to do a test build before uploading
[13:02] <jon__> pmatulis: yes it does
[13:02] <jon__> it seems
[13:04] <Diegonat> hi guys... What's the best free program to design networks ??
[13:04] <Nafallo> Diegonat: whiteboard and marker
[13:05] <[tla]> hi. i have a 12.10 server install and have accidentally overwritten the passwd and shadow files. passwd- is there but obviously i cannot move it back and pwconv because i cant sudo anymore. concerned that as the root volume is on device mapper i may not be able to recover by rebooting into recovery mode kernel. any advice / guidence appreciated before i reboot. :s
[13:05] <Diegonat> Nafallo, ahahahah
[13:05] <Diegonat> come on
[13:05] <Diegonat> you know what I mean =)
[13:05] <jon__> thank you pmatulis
[13:05] <Nafallo> no seriously... that's my advice when designing networks :-)
[13:06] <jon__> i just need to figure out why domain name doesnt resolve
[13:06] <Diegonat> Nafallo, tell me the second best then =D
[13:07] <Diegonat> because i cannot take the whiteboard with me
[13:07] <Diegonat> =)
[13:07] <Nafallo> Diegonat: spreadsheet, pencil and rubber :-)
[13:07] <Diegonat> COME ON!!
[13:07] <Diegonat> =D
[13:08] <Nafallo> Diegonat: dia? gns3?
[13:08] <Nafallo> depending on what you're trying to achieve.
[13:12] <pmatulis> jon__: it's a DNS problem.  either your configuration (nothing to do with apache btw) or the DNS servers themselves do not have the answer to your query.  ① try to query some other name (ex: host google.com) and ② did you tell the DNS servers about your domain name?
[13:12] <pmatulis> ...either your configuration is wrong (nothing...
[13:29] <patdk-wk> roy_k?
[13:29] <patdk-wk> RoyK?
[13:29] <thelamest> sopp
[13:44] <stgraber> zul: hey there. I was wondering if you made any progress on that libvirt-lxcapi driver? (just wondering as it's now on the LXC 1.0 roadmap)
[13:44] <zul> adam_g/jamespage: I have more cloud updates for you when you get a chance http://people.canonical.com/~chucks/ca
[13:45] <zul> stgraber:  no ive been busy with other things
[13:45] <RoyK> patdk-wk: ?
[13:45] <patdk-wk> pm
[14:06] <LargePrime> I got a question about raid configuration
[14:12] <gema> LargePrime: you can share it and see if someone have an answer
[14:12] <gema> has*
[14:12] <LargePrime> ya.  just getting my thoughts together...
[14:13] <LargePrime> I am ordering a dedicated server with 2 SSDs
[14:13] <LargePrime> I would like to configur them with part of the space as raid 1 and the rest raid 0
[14:13] <LargePrime> and have no idea what i am doing
[14:14] <LargePrime> so 1) can this me done?  2) guide?  link? what to google?
[14:34] <hachre> yes you can do it
[14:34] <hachre> butwhy
[14:35] <hachre> just make two partitions on each drive, same layout, and then make a mdraid from one pair of partiton with raid1 and the other with raid0
[14:35] <hachre> mdadm
[14:35] <hachre> is the tool
[14:36] <hachre> haha
[14:36] <hachre> oops
[14:36] <hachre> mt
[14:39] <hachre> LargePrime: .
[14:41] <LargePrime> ok
[14:41] <LargePrime> again I have no idea
[14:41] <LargePrime> so if that is what i want to do.... is there a guide?
[14:55] <Daviey> zul: to check, python-melange is also deprecated ?
[14:55] <zul> Daviey: it is
[15:15] <plars> matsubara: ping
[15:15] <plars> matsubara: are you testing maas on raring isos?
[15:16] <matsubara> plars, yes
[15:16] <matsubara> hehe I think you are the third person that asked me that today
[15:16] <plars> matsubara: you're very popular today I guess :)
[15:17] <plars> matsubara: ok, some people were wondering about results for it
[15:17] <yolanda> jamespage, looking at that bug you commented: https://bugs.launchpad.net/keystone/+bug/1073569
[15:17] <matsubara> plars, our CI tests does most of the testing described in the wiki page, but I'll try it manually installing the ISO from scratch just to be sure (and cover some other test cases that are not covered by CI)
[15:18] <yolanda> there is a CVE reference for CVE-2012-4573
[15:20] <plars> matsubara: but does your ci test install from the latest raring iso?
[15:20] <plars> matsubara: with iso testing, we want to obviously check what's really on the iso, as opposed to what *should* be on it :)
[15:21] <jamespage> yolanda, I see - well the CVE references on that bug look wonky
[15:22] <matsubara> plars, it uses a testbed for each release. the test bed comes from: CLOUDIMGURL=http://cloud-images.ubuntu.com/$RELEASE/current/
[15:22] <matsubara> which I suspect is not what you want tested
[15:22] <yolanda> ok, so i remove it?
[15:23] <jamespage> yolanda, yeah
[15:23] <jamespage> yolanda, I think something odd happened in LP - https://bugs.launchpad.net/bugs/cve/2012-4573
[15:23] <jamespage> there is no way its that many bugs :-)
[15:23] <jamespage> yolanda, nova looks OK btw
[15:24] <yolanda> nice!
[15:26] <yolanda> i did a minor fix also in horizon now, the same problem with email, i have something misconfigured
[15:27] <caribou> jamespage: I see your name in rsyslog's changelog; are you still looking afther this pkg ?
[15:27] <jamespage> caribouwell I touched it once
[15:27] <jamespage> caribou, ^^
[15:27] <caribou> jamespage: too bad for you ;-)
[15:28] <jamespage> whats up?
[15:28] <caribou> jamespage: I got something fixed upstream that I'd like to have into our pkg
[15:28] <jamespage> caribou, critical bug?
[15:28] <caribou> jamespage: we provide 5.8.11 which is no longer supported
[15:28] <caribou> jamespage:
[15:29] <caribou> jamespage: $PreserveFQDN functionality is broken
[15:30] <jamespage> yolanda, one comment on keystone
[15:38] <yolanda> jamespage, i'll recheck it
[15:39] <yolanda> for some reason quilt showed me that the patch wasn't applying, but maybe there was some problem with the merge
[15:39] <caribou> jamespage: remember bug #1066845
[15:40] <jamespage> caribou, I do yes
[15:40] <caribou> jamespage: it's now in Raring, should I do the SRU process even if it's in cloud-archive ?
[15:41] <jamespage> caribou, well quantal folsom packages are currently verified for most recent SRU and should be accepted +7 days
[15:41] <jamespage> so we could stack it after that
[15:41] <jamespage> if its important enought to SRU
[15:41] <caribou> jamespage: what about precise ,
[15:41] <caribou> ?
[15:42] <jamespage> caribou, folsom packages go through quantal SRU process before going to precise
[15:42] <jamespage> cloud archive that is
[15:42] <caribou> jamespage: ok, I see
[15:42] <caribou> jamespage: so what should I do ? format the bug for SRU ?
[15:43] <jamespage> caribou, we kept getting bumped by security fixes - which is why the last lot have been pending for 2 months
[15:43] <caribou> jamespage: yeah, I noticed that
[15:43] <jamespage> caribou, I would really debate whether its SRU worthy TBH
[15:44] <jamespage> caribou, can you sit on it for a bit - there is another point release for folsom to get out after the current one
[15:44] <jamespage> I'll make sure it get swept up with that
[15:44] <caribou> jamespage: sure
[15:46] <hallyn> zul: stgraber: drat, my dnsmasq.d 'fix' for libvirt in raring and precise are wrong.
[15:46] <zul> *sigh*
[15:46] <zul> whats wrong/
[15:48] <caribou> jamespage: ok, will do. thanks
[15:48] <hallyn> /etc/dnsmasq.d-available was being created in debian/tmp, never got cpied into the package
[15:48] <hallyn> got my first 0-day raring sru
[15:49] <hallyn> (the fix is trivial, at least)
[15:54] <yolanda> jamespage, i refreshed the patch, actually only a bit of all was needed, the other bits were already upstream
[15:59] <yolanda> jamespage, i see your approval, so i do the merge? not sure about permissions
[16:00] <jamespage> yolanda, you don't need todo anything for approved things - I'll take care of the upload and pushes as sponsor
[16:01] <yolanda> so i link my branches to the tracking bug?
[16:10] <halvors> When will 13.04 be released?
[16:10] <halvors> Tomorrow 00.00?
[16:10] <Pici> halvors: tomorrow, when its done.
[16:11] <genii-around> Generally sometime after dinner in the UK
[16:27] <parallel21> I've just built a VM using KVM but am unable to ssh into the machine
[16:47] <Daviey> roaksoax: can you look at bug 1172336 urgently please
[16:47] <Daviey> roaksoax: I'd like to understand where the page went.
[16:48] <markthomas> parallel21: I assume you've checked the obvious (that sshd is running in the guest)
[16:48] <roaksoax> Daviey: i think upostream never removed this message
[16:48] <roaksoax> Daviey: since we decided to drop the avahi boot image
[16:48] <parallel21> markthomas: I have no way of doing so
[16:48] <parallel21> to my knowledge
[16:48] <roaksoax> didn't we?
[16:49] <markthomas> parallel21: I believe there's a command from the host to access the VM directly.  One sec.
[16:49] <Daviey> roaksoax: Hmm, possibly
[16:50] <markthomas> parallel21: Do you have virsh?
[16:50] <parallel21> markthomas: yes
[16:50] <markthomas> I believe you can use virsh to access the console.
[16:50] <roaksoax> Daviey: yeah, we did drop the image to do this... (Remember it was the image to be placed on USB stick and make it discover a MAAS server and enlist the machjine))
[16:50] <parallel21> markthomas: Is this it? virsh --connect qemu+ssh://user@host/system
[16:51] <markthomas> parallel21: Let me check my notes
[16:51] <markthomas> parallel21: try virsh console
[16:51] <markthomas> back in 2 min.
[16:53] <Daviey> roaksoax: well, I don't think it is a release blocker.  Can you reintroduce a wiki page explaining that the functionality is deprecated, and raise an upstream task?
[16:54] <roaksoax> Daviey: Already in the process :)
[16:55] <Daviey> perfect!
[16:55] <parallel21> markthomas: "Escape character is ^]" "error: internal error cannot find character device (null)"
[16:56] <markthomas> parallel21: Hmm...
[16:56] <parallel21> one sec
[17:02] <parallel21> I found a tip online saying I need to add a serial console declaration in the xml file
[17:03] <parallel21> I've edited the vm using "virsh edit vm3"
[17:03] <parallel21> But when I boot the vm it rewrites the xml file
[17:04] <markthomas> parallel21: the changes made with virsh edit don't survive a reboot?
[17:07] <parallel21> trying virsh reboot now
[17:07] <parallel21> Yeah, my edits never take
[17:07] <markthomas> Hmm...
[17:08] <parallel21> The vm is running: That's when I do "virsh edit vm3" Make my changes and save
[17:08] <parallel21> Could it be the xml is not being placed in the right place
[17:10] <markthomas> I wouldn't think so.  Where do you keep the XML?
[17:10] <RoyK> anyone here that knows where mdadm is initialized, as in finding raids and assembling them? nested raids are broken and seem to be broken since at least precise. see bug 1171945
[17:10] <parallel21> It's just the a tmp file. "/tmp/virshtY9V2t.xml"
[17:13] <xnox> RoyK: as far as I know we only support 0, 1, 5, 6, 10 raid levels, the rest is questionable in its usefulness.
[17:13] <xnox> RoyK: my recommendation to use LVM instead of "0" level nested or not, and that works quite well.
[17:14] <markthomas> parallel21: try virsh dumpxml, then redirect to a file.  Edit the file, and run virsh define against the file.  Not sure if that will work, but I don't konw why your changes are not persisting.
[17:14] <markthomas> parallel21: Actually, if you have the XML in tmp, skip the first part.  Do the virsh define against your edited file.
[17:16] <markthomas> parallel21: meanwhile, let me ask you this: do you have access to a system with xorg and virt-manager installed?
[17:18] <parallel21> markthomas: I do not. I was hoping to be able to do it all through the command-line. I found my error, looks like it was getting compiled into the xml file. But still hasn't solved the "Escape character is ^]" error
[17:18] <markthomas> parallel21: Yeah, we need to try to connect to the console.  Without virt-manager, we'll have to try to do this the hard way
[17:20] <parallel21> I have this error when using the virsh console command: "qemuDomainOpenConsole:10906 : internal error cannot find character device (null)" from /var/log/libvirt/libvirtd.log
[17:24] <parallel21> Success! kinda
[17:25] <parallel21> markthomas: I did the dumpxml to a different file. Edited that file and redefine the vm using that xml
[17:25] <markthomas> parallel21: can you post some of your /var/log/libvirt excerpts to pastepin.ubuntu.com?
[17:27] <parallel21> http://pastebin.ubuntu.com/5598873/
[17:28] <parallel21> not it hangs at virsh console
[17:31] <markthomas> parallel21: looking into this...
[17:32] <sarnold> markthomas: wow, parallel21's logs sure look ugly. I don't know where you are in debugging it, but perhaps some debsums disk-checking would be useful...
[17:33] <markthomas> sarnold: thanks.  Haven't gotten too far with it, yet.
[17:33] <sarnold> markthomas: thanks :)
[17:33] <parallel21> Running virsh ttyconsole vm3 returns /dev/pts/4
[17:37] <Praxi> Working on a shell script, how do I echo a variable with other text? Does this look correct? echo "The backups were succesfully copied from $backup_files to $destination
[17:37] <Pici> Praxi: Yes.  You may also want to check out #bash, and the links in the topic of #bash
[17:38] <Praxi> its own channel, well ok then!  ty Pici !
[17:39] <markthomas> parallel21: let's back up a step and try sarnold's suggestion:  Run debsums against your qemu-kvm package, and other kvm-related packages you have installed.
[17:41] <markthomas> parallel21: and libvirt-bin, libvirt0, etc
[17:42] <parallel21> Is there a way to list all kvm related packages?
[17:42] <parallel21> debsums for libvirt-bin, libvirt0, kvm all pass with OK
[17:43] <markthomas> parallel21: if there is one, I don't know it.  Just apt-cache depends or whatever it is.
[17:43] <markthomas> Okay.  One sec.
[17:44] <plars> roaksoax, matsubara-afk: fighting with trying to get maas to do something for me here, in an admitedly less than ideal setup.  I've got it at least seeming to respond to the tftp request when I boot a machine, but it seems to timeout. Any logs I can look at? nothing seems to show up under /var/log/maas/* that I can tell
[17:47] <parallel21> markthomas: I just ran debsums on everything and everything checks out
[17:47] <markthomas> parallel21: Okay.  Well, one sec.  I may have an idea, assuming the VM works properly.
[17:48] <markthomas> sarnold: do you know if KVM starts a VNC server process for a guest automatically?
[17:49] <sarnold> markthomas: I think it only creates the vnc if you pass -vnc on the command line
[17:49] <sarnold> markthomas: at least, the manpage says something about "normally, SDL is used .."
[17:49] <sarnold> .. and my running VMs all have -vnc on the command line
[17:49] <markthomas> sarnold: Merci.
[17:50] <parallel21> I could rebuild the instance passing the command
[17:50] <markthomas> parallel21: Can you restart your VM with -vnc so we can try to connect with a VNC viewer on another box?
[17:50] <markthomas> good.
[17:50] <parallel21> Could I just restart? Or does it require a rebuild?
[17:51] <roaksoax> plars: pserv.log maas.log
[17:51] <roaksoax> plars: did you run maas-import-pxe-files?
[17:52] <sarnold> parallel21,markthomas: you'll need a argument to -vnc as well, I think it wants the IP and X display number to use.. e.g. -vnc 127.0.0.1:4
[17:52] <plars> roaksoax: I did
[17:53] <plars> roaksoax: both of those are empty
[17:53] <markthomas> sarnold, parallel21: I believe if you include something like the following in the devices section in your XML it will achieve the same thing:
[17:53] <markthomas>  <graphics type='vnc' port='-1' autoport='yes'/>
[17:55] <sarnold> markthomas: woot :)
[17:55] <parallel21> This is the xml file running
[17:55] <parallel21> http://pastebin.ubuntu.com/5598942/
[17:55] <markthomas> sarnold: I had a guest that was built with virt-manager lying around.  Was just trying to figure out how it does what it does.
[17:56] <markthomas> parallel21: That will work.  If you can grab a box with any GUI and VNC viewer installed, you can tunnel a VNC session over SSH and you won't have to change the listen address.
[17:56] <markthomas> I think.
[17:56] <roaksoax> plars: ok i';m gonna restart the tests since i've been getting pulled to bugs
[17:56] <roaksoax> plars: in what test are you?
[17:58] <parallel21> I'll need to add a firewall rule for my machine to the server. What port does vnc run by default?
[17:58] <plars> roaksoax: well, I was mostly just going through the install one, but wanted to see if I could take it a step further and actually get it talking
[17:58] <parallel21> 5900 it looks like
[17:59] <roaksoax> clear
[17:59] <roaksoax> ls
[17:59] <roaksoax> err
[17:59] <roaksoax> smoser: do we still a maas devenv?
[17:59] <roaksoax> we do right?
[17:59] <roaksoax> plars: yeah just send me the logs
[17:59] <markthomas> parallel21: You shouldn't need a firewall rule if you tunnel over SSH.  The answer to your question is 5900-590x depending on the display number.  Should be :0 in this case, so 5900.
[18:00] <plars> roaksoax: the logs are empty
[18:00] <roaksoax> plars: weird... are you sure they pxe boot from maas
[18:00] <roaksoax> plars: can you send me a screenshot?
[18:01] <RoyK> xnox: erm - it works with lucid
[18:02] <parallel21> Does the vnc client allow ssh tunneling? or do I setup it up like so " ssh -L 5900:localhost:5900 -N -f -l user hostname"
[18:03] <xnox> RoyK: sure. not everything that works or worked is inherently supported. Can you please tell me why you use "0" level instead of lvm?
[18:03] <xnox> RoyK: with or without RAID levels for redundancy.
[18:04] <markthomas> parallel21: I believe so
[18:04] <RoyK> xnox: well, it's rather useful, raid-0 on top of raid-5 will redistribute the data if the raid is extended. afaik lvm will not
[18:05] <RoyK> xnox: also, it doesn't make sense of not supporting something that works, and worked in previous versions of ubuntu
[18:07] <mgz> can I see the reason for rejection of juju-core in raring anywhere?
[18:07] <smoser> roaksoax, we dont relaly have a maas devenv. virtual-maas really took over that need.
[18:07] <mgz> xnox: ^maybe you know? last time I just got an email, and I wasn't the owner of the packaging branch for this one
[18:07] <roaksoax> smoser: yeah the only problem is that virtual maas depends on juju :)
[18:07] <Daviey> mgz: Yes
[18:08] <Daviey> mgz: #ubuntu-release and the FFE tracking bug
[18:08] <smoser> why do you think virtual maas depends on juju?
[18:08] <smoser> ah. for setup. i see.
[18:09] <xnox> mgz: many reasons. But none of them are fatal, in a sense they should be fixed and reuploaded and accepted as an sru to introduce that new package.
[18:09] <smoser> yeah, we wanted to separate that out.
[18:09] <xnox> mgz: there is no ultimate requirement to have it in the release pocket.
[18:09] <mgz> ah, the bug gets updated? didn't last time...
[18:09] <Daviey> mgz: There was bit of a bun fight both on the bug and in #ubuntu-release
[18:10] <xnox> mgz: it's irregular. But SRU/release are happy to introduce the new source package juju-core as an SRU, such that it's top-notch the first time people can install it from the supported archive.
[18:10] <xnox> (teams that is)
[18:11] <mgz> oh, and isn't the bug fun...
[18:12] <xnox> mgz: i'm just relaying the discussions i have overheard in #ubuntu-release and further debates around the table here.
[18:12] <mgz> I
[18:12] <mgz> I'll read the log
[18:12] <roaksoax> smoser: yeah, well In my case I can setup my own env pretty quickly, but test cases such as plars might be more complicated
[18:12] <xnox> mgz: no problem.
[18:13] <parallel21> markthomas: I'm on a mac. I have vnc-viewer. I've tunnel the port over using ssh. I get a warning that this connection will be unencrypted without a vnc license. I hit continue and it connects briefly and then closes the connection
[18:13] <markthomas> parallel21: try Chicken of the VNC.  I find it less problematic than others like RealVNC Viewer.
[18:13] <markthomas> parallel21: free to download, and does work with KVM
[18:14] <parallel21> markthomas: will do
[18:15] <parallel21> So this must be the problem
[18:15] <parallel21> http://i.imgur.com/s2ulMiD.png
[18:16] <markthomas> parallel21: your VM is frelled.
[18:16] <parallel21> So rebuild
[18:17] <markthomas> parallel21: looks that way.  Did it have anything important on it that wasn't backed up?
[18:17] <parallel21> markthomas: Nay… this was just a test setup. Is there a place I can find templates to work from? I'm looking to setup a vm running just nginx and php on a single core
[18:20] <parallel21> markthomas: this is my build command http://pastebin.ubuntu.com/5599005/
[18:21] <markthomas> parallel21: I haven't used vmbuilder.  I'll have to look into it.
[18:21] <parallel21> What do you use?
[18:21] <smoser> roaksoax, plars http://bazaar.launchpad.net/~smoser/maas/maas-pkg-test/files
[18:21] <smoser> i'm following that right now for testing raring
[18:22] <markthomas> parallel21: I've been using virt-manager to build from ISO for the guests I have.
[18:22] <adam_g> zul, those CA packages LGTM
[18:22] <markthomas> parallel21: but it looks like it's time for me to learn vmbuilder
[18:22] <zul> adam_g:  cool thanks
[18:24] <parallel21> markthomas: I'm gonna rebuild once more and see if I get a working formula down
[18:32] <zul> Daviey: ping can we can cinder and nova through the queue please? they contain an upstart typo and fixes for the CA
[18:32] <bkerensa> pmatulis: hi
[18:33] <plars> smoser: ah, this looks useful
[18:33] <plars> smoser: do these still work on raring?
[18:33] <smoser> i got most of the way through before a system hung (it was an instance)
[18:35] <Daviey> zul: looking
[18:36] <Daviey> zul: i need a bug for iptables dependency
[18:39] <zul> Daviey:  damn it
[18:39] <zul> do you want me to make one now/
[18:40] <Daviey> zul: yes please.  It's far too late to not have an upload without a decent bug
[18:44] <zul> Daviey:  coming right up
[18:47] <zul> Daviey:  so i made a boo boo...adam_g already +1 while it was still stuck in the queue so the staging ppa has the changelog without the bug number can we remove the old one from the ppa and i can reupload it?
[18:49] <Daviey> zul: it's not possible to upload the identical version.  You will need to make it ~cloud1
[18:50] <zul> Daviey:  ok will do
[18:50] <Daviey> (which you can just do)
[18:52] <Daviey> zul: can you add a bit more meat to the iptables bug please
[19:02] <matsubara> plars, do you have the dhcp server running?
[19:03] <plars> matsubara: I do
[19:05] <matsubara> plars, and your node get an ip address but then doesn't get the image to pxe boot the node?
[19:08] <plars> matsubara: right, it gets an ip in the range I specified, tftp starts, then after a few seconds I get "TFTP open timeout"
[19:09] <matsubara> plars, did you run maas-import-pxe-files?
[19:09] <plars> yes
[19:09] <matsubara> did it finish successfully? what do you get in pserv.log?
[19:09] <plars> matsubara: nothing in pserv.log - it's empty
[19:09] <matsubara> when the node does the tftp request, what shows up in the pserv.log?
[19:09] <matsubara> oh
[19:10] <matsubara> I think I've seen that before. can't remember what the issue was. /me thinks
[19:11] <matsubara> plars, what do you have in your /etc/maas/dhcpd.conf?
[19:11] <plars> matsubara: In the interest of full disclosure, I *do* have another dhcp server on the network (my router) which I haven't touched. But the ip I get every time is the first one from the range that I specified in maas, and it didn't even start to dhcp/tftp from netboot until I installed maas-dhcp and set that up
[19:12] <matsubara> plars, I've seen the TFTP open timeout issue when the next-server is set to 127.0.0.1 in the dhcpd.conf
[19:12] <matsubara> did you run sudo dpkg-reconfigure maas and changed the URL for the pxe server?
[19:12] <plars> matsubara: http://paste.ubuntu.com/5599118/
[19:13] <plars> hmm... no I didn't
[19:13] <plars> I never saw anything about running dpkg-reconfigure, maybe I overlooked it
[19:13] <matsubara> plars, the package might pick the wrong interface during install
[19:14] <plars> matsubara: it doesn't ask me for anything
[19:15] <matsubara> plars, sorry, dpkg-reconfigure maas-region-controller
[19:20] <plars> no effect after setting it there
[19:20] <plars> I don't seem to have a next-server in that dhcpd.conf
[19:21] <matsubara> plars, I'm setting up the VM so I can try to reproduce. just a minute
[19:24] <adam_g> zul, whats the point of symlinking from /etc/bash_completion.d/ to /usr/share/doc/python-novaclient? why not just install the completion file into /etc/bash_completion.d/?
[19:24] <zul> adam_g:  thats the way people do it apparently
[19:27] <matsubara> plars, so you have two dhcp servers in the network segment? http://www.bootix.com/support/problems_solutions/pxe_e32_tftp_open_timeout.html says having the dhcp server running on different machines than the one the tftp is running from might cause the timeout problem
[19:27] <matsubara> could you try adding next-server 192.168.1.1 (assuming this is your pxe server address), restart maas-dhcp and then reboot the node?
[19:27] <plars> matsubara: the dhcp server I'm getting the ip from is *definitely* maas though
[19:27] <matsubara> plars, yeah, that's odd indeed
[19:28] <matsubara> according to the page above tftp client should assume the tftp server is the same it got its ip from
[19:29] <matsubara> plars, maybe a firewall rule is blocking the tftp requests?
[19:30] <Daviey> Oh dear.  You are seeing a timeout?
[19:31] <plars> matsubara: not a firewall between these systems, all on the same subnet
[19:31] <Daviey> plars: What hardware is the client?
[19:31] <plars> Daviey: don't get too paranoid just yet, this is my first time really trying to set up maas
[19:31] <plars> Daviey: a spare laptop I have around here for testing
[19:32] <Daviey> plars: So, it could be buggy pxe implementation.  We had this with an enterprise grade server, https://bugs.launchpad.net/maas/+bug/1155556
[19:34] <plars> interesting, not sure if I have another machine here I can use at the moment to try
[19:55] <parallel21> markthomas: I have since restart the process and I dunno what I've done differently. But everything seems to work. I think I didn't powerdown a machine one time and instead used the destroy to power down it. Anyway… thanks for all the help!
[19:56] <markthomas> parallel21: Glad you got it figured out.
[20:18] <LargePrime> ima ask a noob question.  how do i undo "iptables -A INPUT -m iprange --src-range ip.add.ress.0-ip.add.ress.255 -j DROP"  a link or search critera would be cool
[20:19] <LargePrime> and thanks you
[20:51] <maxb> LargePrime: 'iptables --help', look specifically at the --list, --line-numbers and --delete options
[20:54] <LargePrime> thanks maxb
[21:28] <med_> smoser, hallyn, stgraber, zul, Daviey et al: Why does the KVM FAQ say to not run NTP on kvm guests? It appears to just be out of date. Is there any tech reason not to these days (seems to be tech reasons to actually RUN ntpd on guests).   Ref: https://help.ubuntu.com/community/KVM/FAQ
[21:28] <med_> smoser, hallyn, stgraber, zul, Daviey et al: Why does the KVM FAQ say to not run NTP on kvm guests? It appears to just be out of date. Is there any tech reason not to these days (seems to be tech reasons to actually RUN ntpd on guests).   Ref: https://help.ubuntu.com/community/KVM/FAQ
[21:29] <med_> (looks like an artifact from 2008 on the FAQ that may no longer be valid in current kvm.)
[22:01] <hallyn> med_: yeah i've talked to him about that yesterday
[22:02] <hallyn> as you guessed, it's probably a very old suggestion, and it's also the community docs, not the server guide.  i had no idea that was being suggested.
[22:03] <JuanBre> hi, I want to know if ubuntu-server kernel comes with napi support enabled....
[22:25] <med_> hallyn, many thanks.
[22:26] <sarnold> hallyn: does that mean we -should- run ntpd on both guests and hosts?
[22:27] <sarnold> (I always thought the guests would wind up being very wrong about clock drift and drive the clocks further away from reality..)
[22:30] <med_> sarnold, the research indicates it's probably a good idea to do both. And if  you can point the guests at the host as master, that's ideal.
[22:30] <hallyn> sarnold: yeah I thin you want to run it on both
[22:31] <med_> (in  a cloud, you may have no clue about the host of course)
[22:31] <hallyn> sarnold: note that ntp only adjusts how jiffies are accounted in the kernel - it doesn't change the hwclock, so guests wont' affect host and vice versa
[22:35] <sarnold> hallyn: this is where I get mighty fuzzy on details... I thought that the cpu-based clock source that the kernel uses to keep track of time would propogate from host to guests when the host's time is updated.. (cpu counter? hrtimer? man it's been ages..)
[22:39] <hallyn> sarnold: my impression was it doesn't change the clock frequency, it's just that when clock interrupts arrive, it changes how often it bumps the jiffies
[22:39] <sarnold> hallyn: well, cool, that's one old folklorism I can remove from my brain :) thanks
[22:40] <hallyn> sarnold: i'm not 100% convinced.  and of course if yo'ure right then ntp in guests means they're fighting each other
[22:42] <hallyn> but no i'm pretty sure it's right
[22:43] <sarnold> pfew :)
[23:07] <LargePrime> so i al looking at installing/replacing the kernel on my dedicated server. 12.10  any recommendations?  desktop and server kernels are merged now right?
[23:07] <LargePrime> http://pastie.org/7711965 is the list i have i gues
[23:28] <RoyK> LargePrime: for a server, use 12.04
[23:28] <RoyK> the LTS releases are the best for servers
[23:29] <RoyK> and since 8.04 is out of support in a week, and 10.04 is oldish, 12.04 will probably be the best unless you have special needs
[23:30] <LargePrime> well i think it too late. and yes, generaly i fell like i am special needs, but that is not ubuntu related
[23:30] <LargePrime> well not specifically
[23:31] <RoyK> I only use LTS on servers
[23:31] <maxb> I would say that it depends on the server. If it's the kind of server where you value up-to-date-ness of software and don't mind upgrading every 6 months, then track the latest release. Otherwise, stick with LTSes
[23:31] <RoyK> the non-LTSes tend to be far less tested