[00:21] <diago> mdadm is screwed from an install...
[01:14] <t3chkommie> hey guys, i need some help with the system email my server trys to email out.
[01:15] <t3chkommie> when i log in the server tells me i have new emails, but i have no diea how to see these, or forward them to my email account on my imap server
[02:07] <shanezilla> anyone use SSH server
[02:25] <shanezilla> CLEAR -all
[02:26] <shanezilla> soory all new to irssi
[02:26] <shanezilla> sorry that is : )
[03:19] <lukehasnoname`> I wonder if Microsoft has an internal IRC server
[03:36] <shanezilla> windows dose not supply an installed IRC client
[03:37] <shanezilla> or an IRC server
[03:38] <myk_> No, it doesn't.
[03:41] <lukehasnoname`> they just use messenger and outlook, I suppose
[03:44] <myk_> What do i need to do to have an unattended setup/installation? Is there a special instal media, or a boot argument?
[04:20] <ccheney> my wife is now predicting she will be at the hospital within 24hr, heh
[04:28] <unewbie1> ccheney: why is that?
[04:30] <lukehasnoname`> I'm assuming baby, given the chuckle at the end
[04:34] <ccheney> unewbie1, baby was due on last friday but the doctor wasn't sure when it would actually come
[04:34] <shanezilla> disconnect
[04:34] <ccheney> the doctor thought it wouldn't come until induction which was scheduled for june 17
[04:53] <MTecknology> ccheney: gettin' close then?
[04:53] <MTecknology> ccheney: I wish you the best of luck on a happy healthy baby
[04:55] <ccheney> MTecknology, yea any day now :)
[04:56] <MTecknology> ccheney: congrats - you an ubuntu member? - if so - make sure to put that on the planet
[04:56] <ccheney> MTecknology, yea ubuntu core dev
[04:56] <MTecknology> ccheney: heh... i shoulda known that actually :P
[04:57] <ccheney> heh :)
[04:57] <MTecknology> not like i haven't read email from you
[04:58] <MTecknology> you should grab the hostmask though - then I can know for sure jsut from a whois :P
[04:58] <ccheney> MTecknology, yea i have one for my old nick, need to get an updated one
[04:59] <unewbie1> ccheney: is it a boy or girl?
[05:00] <ccheney> boy
[05:01] <unewbie1> how do you know if he is not born yet? :D
[05:01] <ccheney> ultrasound :)
[05:01] <unewbie1> yeah
[05:02] <unewbie1> i didn't know until my son was born
[05:02] <unewbie1> i don't trust usg
[05:04] <unewbie1> i try kvm on ubuntu 10.04 but it's keep complaining for error pci_add_option_rom: failed to find romfile "pxe-rtl8139.bin"
[05:04] <unewbie1> Could not initialize SDL - exiting
[05:04] <unewbie1> it's my first virtualization :D
[05:04] <unewbie1> and not work
[05:05] <MTecknology> unewbie1: odd..
[05:05] <MTecknology> unewbie1: do virsh list --all
[05:06] <MTecknology> unewbie1: do you see anything listed and if so, pastebin what you see
[05:07] <unewbie1> no
[05:07] <unewbie1> i try to create my first guest os
[05:07] <MTecknology> pastebin the command you use and output
[05:09] <unewbie1> http://pastebin.com/fQuh2C1K
[05:10] <MTecknology> !kvm
[05:10] <MTecknology> unewbie1: check out this page - https://help.ubuntu.com/community/KVM/CreateGuests
[05:10] <unewbie1> ok
[05:10] <MTecknology> unewbie1: personally - the kvm command is kind of a pain and i only use it if i'm not making an ubuntu guest
[05:14] <unewbie1> actually i dont understand virtualisation. i installed 10.04 yesterday and put a 400GB LVM partition. should i do this? :)
[05:15] <unewbie1> this partition created for the guests os
[05:15] <MTecknology> depends what you want to achieve :P
[05:15] <MTecknology> don't make a partition for just one guest
[05:16] <unewbie1> i made that 1 partition for 3 os
[05:16] <unewbie1> is that ok?
[05:16] <MTecknology> ya
[05:16] <MTecknology> I have a partition that's nearly 1TB on LVM for all guests at /virt
[05:18] <unewbie1> the partition was data and i create virt folder inside
[05:18] <unewbie1> is that ok?
[05:19] <MTecknology> depends if that's what you want
[05:19] <MTecknology> if that's what you want - then it's ok
[05:19] <unewbie1> i don't know what i want :D
[05:19] <unewbie1> i'm new
[05:20] <MTecknology> you want to play with virtualization - you know that much
[05:20] <unewbie1> i only want ubuntu servers in my machine
[05:20] <MTecknology> doing that mkes life much easier
[05:21] <unewbie1> i have few servers with no virtualisation so i wanna try
[05:21] <MTecknology> it's fun once you get the hang of it
[05:21] <MTecknology> libvirt is an interesting beast - your entire config is xml too
[05:22] <unewbie1> can i have web interface also? :D
[05:23] <MTecknology> You can.. but imo - you're much better off not doing that
[05:23] <unewbie1> why? isn't that easier?
[05:24] <MTecknology> My personal opinion is that you get a lot further learning to do things on cli - you're much more efficient most of the time if you can do something on cli as opposed to gui
[05:24] <MTecknology> in the case of libvirt - if you want gui I'd say get virt-manager
[05:26] <unewbie1> can i create my guest with libvirt too?
[05:26] <MTecknology> ya
[05:27] <MTecknology> just need to setup your storage pools correctly
[05:28] <unewbie1> how to setup the size? here https://help.ubuntu.com/10.04/serverguide/C/libvirt.html#libvirt-virt-install, i see: -s 4
[05:28] <unewbie1> what's that mean?
[05:28] <unewbie1> 4 GB?
[05:29] <MTecknology> yup
[05:30] <unewbie1> i made the file with dd command earlier for 100 GB. should i put -s 100?
[05:30] <MTecknology> ya
[05:30] <MTecknology> wait
[05:31] <MTecknology> you want to use an existing system?
[05:31] <MTecknology> -s will be part of making a new disk
[05:31] <unewbie1> i want to use the existing
[05:32] <MTecknology> I haven't don that ever - it's easily possible but I probably can't help you short of google
[05:33] <unewbie1> then create new disk is fine
[05:35] <unewbie1> Starting install...
[05:35] <unewbie1> Creating storage file bar 100% |[05:35] <unewbie1> Creating domain...                                                 0 B 00:00
[05:35] <unewbie1> Domain installation still in progress. You can reconnect to
[05:35] <unewbie1> the console to complete the installation process.
[05:35] <MTecknology> !pastebin
[05:35] <unewbie1> what is that mean?
[05:36] <unewbie1> sorry :D
[05:36] <MTecknology> ya know what i'd do.....
[05:36] <unewbie1> http://paste.ubuntu.com/445899/
[05:37] <MTecknology> make a vm that has a single 4GB disk - then swap that disk with the one you made - then you have everything else setup
[05:37] <unewbie1> where is the guest?
[05:37] <unewbie1> dont understand
[05:37] <MTecknology> it's in files
[05:38] <MTecknology> virsh list --all
[05:38] <MTecknology> do you see anything there yet?
[05:38] <unewbie1> Id Name                 State
[05:38] <unewbie1> ----------------------------------
[05:38] <unewbie1>   1 barracuda            running
[05:38] <MTecknology> your vm is running
[05:39] <unewbie1> and what to do then?
[05:39] <MTecknology> ssh into it
[05:39] <unewbie1> ssh? i don't even installed the os yet?
[05:39] <unewbie1> what is the ip then?
[05:40] <MTecknology> from your system you can do virt-viewer -c qemu+ssh://192.168.1.4/system barracuda     where 192.168.1.4 is the host ip - not the vm
[05:41] <unewbie1> how do you know my ip? :)
[05:42] <MTecknology> call it a good guess
[05:42] <unewbie1> i'll try
[05:43] <unewbie1> virt-viewer is not installed yet
[05:44] <unewbie1> 50MB to download
[05:45] <unewbie1> http://paste.ubuntu.com/445903/
[05:46] <unewbie1> error
[05:47] <MTecknology> you caught the part where you run this on your system, right?
[05:48] <unewbie1> dont understand
[05:49] <MTecknology> your system = what you're typing on now
[05:49] <unewbie1> yes
[05:50] <unewbie1> the host
[05:50] <MTecknology> no
[05:50] <unewbie1> o
[05:50] <MTecknology> use another system with X
[05:51] <unewbie1> another ubuntu?
[05:51] <MTecknology> yes
[05:51] <unewbie1> i dont have any ubuntu desktop
[05:53] <MTecknology> did you use virt-install to make the vm?
[05:53] <unewbie1> yes
[05:54] <MTecknology> you're going to need to recreate the vm and build it with --add-pkg openssh-server
[05:55] <unewbie1> http://paste.ubuntu.com/445915/ that's how i create the vm
[05:57] <MTecknology> ya.. you need a desktop system in order to complete that install
[05:57] <unewbie1> i dont have desktop
[05:57] <unewbie1> any alternative?
[05:58] <unewbie1> how to recreate to vm?
[05:58] <MTecknology> virsh destroy barracuda && virsh undefine barracuda
[05:59] <MTecknology> lemme give you a magical command
[05:59] <unewbie1> Domain barracuda destroyed
[05:59] <unewbie1> Domain barracuda has been undefined
[05:59] <MTecknology> !enter
[06:00] <MTecknology> How much ram do you wantto give it?
[06:00] <unewbie1> 1 GB
[06:01] <MTecknology> how much swap?
[06:01] <unewbie1> dont understand
[06:01] <MTecknology> !swap
[06:01] <MTecknology> oh, and how many CPU's?
[06:02] <unewbie1> 2 GB swap and 1 cpu
[06:02] <unewbie1> the server only have 1 processsor
[06:02] <unewbie1> quad core
[06:03] <MTecknology> that means you can give it up to 4 procs
[06:03] <unewbie1> ok, lets give it 4
[06:03] <MTecknology> 2GB swap is insane.. but ok
[06:03] <unewbie1> insane?
[06:03] <unewbie1> not enough?
[06:04] <MTecknology> very high
[06:04] <unewbie1> make it just enough
[06:04] <unewbie1> how much it should be?
[06:04] <MTecknology> I use 256MB on all of mine
[06:04] <blue-frog> on prod servers, the higher the better
[06:04] <unewbie1> it's for production
[06:05] <blue-frog> for desktop for people who are just playing witht their computers 256 is enough
[06:05] <MTecknology> blue-frog: not if it's a vm - then you just start thrashing the host
[06:05] <blue-frog> ah vm
[06:05] <blue-frog> hum
[06:05] <blue-frog> dunno about vm. sorry fr my talkings
[06:05] <unewbie1> ok let it 256
[06:06] <MTecknology> unewbie1: what's your full name and what do you want for a user name?
[06:06] <MTecknology> didi for a username?
[06:08] <MTecknology> Looks like you're from Indonesia - are there any package mirrors near there?
[06:12] <MTecknology> unewbie1: I'm trying to rig up a command for you that you can reuse over and over and over
[06:12] <MTecknology> you still exist?
[06:17] <unewbie1> sorry
[06:17] <unewbie1> a phone call
[06:17] <unewbie1> that's ok
[06:17] <unewbie1> let's make didi
[06:18] <MTecknology> and full name?
[06:18] <unewbie1> supriyadi
[06:18] <MTecknology> that's not a full name but ok
[06:18] <unewbie1> that's a fullname here :)
[06:18] <MTecknology> you don't have last names?
[06:18] <unewbie1> nope
[06:19] <unewbie1> that's the culture here
[06:19] <MTecknology> odd
[06:19] <MTecknology> there... you ready for your magic command
[06:19] <MTecknology> http://pastebin.ubuntu.com/445921/
[06:19] <MTecknology> enjoy
[06:20] <unewbie1> the longest command i've ever seen
[06:20] <MTecknology> lol... that's the trimmed version
[06:20] <MTecknology> the one I use is MUCH longer - and if I'm adding static IP's... - but I know EXACTLY what I'm getting
[06:21] <unewbie1> is this vmbuilder should be installed first?
[06:22] <MTecknology> ya.. if you don't install the app you're trying to use you can't use it
[06:22] <unewbie1> ok
[06:23] <unewbie1> E: Couldn't find package vmbuilder
[06:24] <MTecknology> !kvm
[06:24] <MTecknology> sudo aptitude install kvm libvirt-bin ubuntu-vm-builder bridge-utils
[06:25] <unewbie1> waiting for installation
[06:26] <unewbie1> how do you know everything? :D
[06:29] <MTecknology> i don't know that much
[06:29] <unewbie1> i have to configure email?
[06:29] <MTecknology> hm?
[06:30] <unewbie1> i got that screen here
[06:30] <unewbie1> i dont remember installing postfix
[06:31] <unewbie1> looks like postfix is coming with the aptitude
[06:31] <unewbie1> should i cancel it?
[06:32] <MTecknology> no
[06:32] <MTecknology> dependencies
[06:32] <unewbie1> ok
[06:32] <unewbie1> and?
[06:33] <unewbie1> how can i configure this postfix?
[06:35] <unewbie1> are you sure we need this postfix?
[06:35] <MTecknology> just let it install with defaults
[06:36] <unewbie1> no configuration?
[06:36] <MTecknology> no
[06:36] <unewbie1> or internet site?
[06:36] <unewbie1> or local only?
[06:37] <MTecknology> whatever was selected by default
[06:37] <unewbie1> i don't remember :D
[06:37] <unewbie1> i played around with the cursor
[06:38] <MTecknology> usually the default selected is the first in the list
[06:38] <unewbie1> that means no configuration
[06:38] <unewbie1> ok
[06:38] <unewbie1> is that ok?
[06:38] <MTecknology> press enter..
[06:39] <unewbie1> ok
[06:41] <unewbie1> ok it's done
[06:41] <unewbie1> now it is the guest installation
[06:45] <unewbie1> another error message
[06:45] <unewbie1> http://pastebin.ubuntu.com/445927/
[06:45] <unewbie1> it's already exist
[06:45] <MTecknology> !enter
[06:45] <MTecknology> delete the existing disk image
[06:46] <unewbie1> ok
[06:46] <unewbie1> it's quiet now
[06:47] <unewbie1> is it going to take a long time?
[06:47] <MTecknology> yup
[06:47] <unewbie1> ok
[06:47] <MTecknology> when it finishes it will start and you can ssh into it because openssh-server will be installed
[06:48] <MTecknology> the timezone should be setup - and I set what I perceived to be the fastest package mirror for you
[06:48] <unewbie1> is it installing the os now?
[06:48] <MTecknology> yes
[06:48] <Yosi> hi All, I've been using Ubuntu Server as an Apache for the server for the last week, and was thing about using it for a production machine.  all my friends at work are telling me to only use RHEL or CentOS in prodcution, is ubuntu server no stabl;e enough?
[06:49] <MTecknology> and that command i gave you - i hate vim-tiny so i removed it and installed the full vim for ya
[06:49] <MTecknology> Yosi: lol... it's a matter of flavor - I use Ubuntu on all of my production servers
[06:49] <unewbie1> i only understand nano :D
[06:50] <unewbie1> yeah me too
[06:51] <unewbie1> and yes my friend said the same thing to me
[06:51] <Yosi> do i need a firewall in front of the default install?  or are all the major security ports locked by default
[06:51] <MTecknology> I know the centos community and they're great- so the OS choice really comes down to what you like better - for me i like the amazing usability of ubuntu
[06:52] <Yosi> performance wise, I'm assuming ubuntu is as fast as rhel or cent for apache?
[06:52] <Yosi> yes, new to linux, and ubuntu seems easier than cent
[06:52] <MTecknology> Yosi: in most linux distros there's nothing being listened to by default - I always use ufw which is amazingly easy - just makes sure nothing else is sneaking in
[06:53] <unewbie1> Yosi: me too
[06:53] <Yosi> fair enough..
[06:54] <MTecknology> I would always put a firewall in front (pfsense by choice) - that just makes routing much easier
[06:54] <unewbie1> it's my first distro actually
[06:54] <Yosi> my raid tools are made for cent/rhel etc.. from LSI (9260-8i card), but I can just configure it by the RAID bios, and apparently kernel 2.6 has the lsi drivers build in...  i just lose the management tools, but once the raid is established, i dont really need them
[06:55] <naiad2> I'm wondering if it's possible to have apparmor restrict the output of ps like the grsecurity kernel module does.  limit ps to only pids that the user owns?
[06:56] <qman__> Yosi, ubuntu server is most definitely secure by default
[06:56] <qman__> there is nothing listening on any ports on a default install
[06:56] <MTecknology> naiad2: I highly doubt it
[06:56] <Yosi> qman__: thanks...
[06:57] <Yosi> what i don't understand is why, big sites like ebay etc.. have problems with DOS attacks... can't most firewalls block a DOS attack?
[06:57] <MTecknology> no
[06:58] <Yosi> can't they just block that IP from spamming them?
[06:58] <MTecknology> nope
[06:58] <MTecknology> if you have 10GB traffic running through a 1GB pipe; you're plugged up - even if you block all that traffic - you still have to process that other 9GB
[06:59] <qman__> yes
[07:00] <qman__> the problem is DDoS
[07:00] <qman__> it comes from many IPs, all over the world
[07:00] <qman__> they run a public site, they can't just shut down all connections
[07:01] <MTecknology> the only way to deal with it is to have more resources than your attacker
[07:01] <qman__> you can configure your firewall to dynamically slow single-poing DoS attacks, but DDoS is nearly impossible to stop that way
[07:01] <qman__> single-point*
[07:01] <X-Sleepy-X> Easy solution: Compose a fake letter where it says it's terrorists behind the attack and send it to Pentagon. They will have those hackers nuked sooner than you can say "Wake up it's a beautiful morning!"
[07:02] <qman__> the usual approach is just that, get more bandwidth, and try to find the source of the attacks
[07:02] <X-Sleepy-X> :P
[07:02] <qman__> the person running the botnet
[07:02] <qman__> and get the botnet shut down that way
[07:02] <MTecknology> X-Sleepy-X: little offtopic for this channel
[07:02] <Yosi> back, got some water...
[07:02] <MTecknology> I wonder if that's why it's so hard to purchase upload speed in the states...
[07:02] <X-Sleepy-X> MTecknology: Kind of, yeah. But it was my way of saying good morning. ;)
[07:03] <MTecknology> I can get 30mbit download for cheap enough - getting >1mbit upload gets to be insane
[07:03] <qman__> the problem stateside is that the telcoms have government-backed monopolies
[07:03] <qman__> no accountability, so they are way behind on upgrading their lines and tech
[07:03] <Yosi> in Canada 20mbs fiber provisioned over E100 service is about $1400 a month
[07:04] <qman__> when my connection doesn't work the way it's supposed to, I don't have any resolution
[07:04] <qman__> there is only one ISP
[07:04] <MTecknology> ouch
[07:04] <qman__> if they don't perform, I'm SOL
[07:05] <MTecknology> burn
[07:05] <X-Sleepy-X> I pay about 300 euros for 10/10 Mbit/s but I get about 15/30 Mbit/s.
[07:05] <qman__> US connections are the opposite
[07:05] <qman__> you pay for 10/2, and get 6/1.5 in off-hours
[07:05] <MTeck-ricer> pretty much
[07:05] <MTeck-ricer> Actually.. I get what i pay for
[07:06] <Yosi> i think the US is the cheapest I've seen for bw, cheaper than canada
[07:06] <MTeck-ricer> I get 15/1
[07:06] <unewbie1> MTeck-ricer: installation is finish
[07:06] <rayvtirx> are these ordinary home connections
[07:06] <MTeck-ricer> $30/mo
[07:06] <unewbie1> now how to ssh?
[07:06] <qman__> business connections too
[07:06] <rayvtirx> im from uk and my upload is disproportinately slower
[07:06] <qman__> DSL and cable
[07:07] <qman__> unless you pay $300+ for a T1, you won't get your advertised speeds
[07:07] <MTeck-ricer> unewbie1: 'nmap -sP 172.16.88.0/24' - figure out which new node popped up
[07:07] <qman__> (per month)
[07:07] <qman__> at least, you can't count on getting your advertised speeds
[07:07] <rayvtirx> i get around 8 Mbit down and 512 kbits up
[07:07] <qman__> some areas are better than others
[07:07] <MTeck-ricer> I pay ~$45/mo for 15/1 on cable
[07:07] <MTeck-ricer> I get what i paid for and that makes me happy
[07:07] <X-Sleepy-X> I pay 38,58 for my connection. I get 15/30.
[07:08] <X-Sleepy-X> :)
[07:08] <unewbie1> another package to be installed
[07:08] <Yosi> x-sleepy-x: where are u from?
[07:08] <X-Sleepy-X> Sweden
[07:08] <Yosi> nice
[07:08] <qman__> my cable is $70/mo, for 8/1.5
[07:08] <X-Sleepy-X> I could order 100/100
[07:08] <Yosi> I pay $60 for 8/1 at home
[07:08] <X-Sleepy-X> But that's a bit more expensive
[07:09] <qman__> but it's only gotten that fast in the last month
[07:09] <X-Sleepy-X> I only pay for 10/10 but I get a bit more... ;)
[07:09] <qman__> I was paying the same rate for 5/512k
[07:09] <unewbie1> what am i looking for here?
[07:09] <MTeck-ricer> I switched ISP's - I was getting 1/256k
[07:09] <MTeck-ricer> same into price
[07:10] <X-Sleepy-X> That's not good.
[07:10] <qman__> only one option here, the cable company
[07:10] <MTeck-ricer> this new intro price lasts about 8mo longer - and after the change this new one is cheaper
[07:10] <qman__> no DSL available, satellite and cell is a joke
[07:10] <X-Sleepy-X> Start a ISP company...
[07:10] <X-Sleepy-X> an*
[07:10] <Yosi> i hope ubuntu sees my raid array..   i have 4 intel ssd's x25-m 160gb comming tomorrow... pluggin them into my lsi 9260 6 gb/s raid card and going to load ubuntu on that as my apache server...
[07:11] <Yosi> would be fun to start a datacenter...  lol
[07:11] <MTeck-ricer> qman__: gah... theyr website is horrible too
[07:11] <qman__> Yosi, you might look into the different versions of apache available
[07:11] <qman__> the default one is a bit slower in favor of the biggest feature set
[07:11] <qman__> mpm-prefork
[07:12] <MTeck-ricer> or just use nginx
[07:12] <qman__> with a system like that, you're obviously concerned with performance
[07:12] <MTeck-ricer> nginx > apache | stop abusing resources
[07:12] <qman__> so try the different options out
[07:12] <Yosi> thanks qman, i will check into that
[07:12] <qman__> see what works
[07:12] <twb> prefork isn't the default
[07:12] <twb> It's only pulled in if you ask for PHP
[07:13] <Yosi> qman: exactly, I want hi IOPs.....    its apache, php and mysql
[07:13] <qman__> which you ask for if you use LAMP
[07:13] <MTeck-ricer> qman__: wow.... their website is BAD
[07:13] <qman__> charter? yeah
[07:13] <qman__> you should try using their webmail
[07:13] <qman__> it's an event
[07:13] <qman__> their DNS servers are useless too, I run my own
[07:14] <twb> bah, tasksel blows
[07:14] <MTeck-ricer> qman__: if i buy now i can get phone/email/phone for only $130/mo :P
[07:14] <MTeck-ricer> why not opendns?
[07:15] <qman__> opendns didn't exist at the time
[07:15] <qman__> when I set it up
[07:15] <MTeck-ricer> twb: I used to like it - but i no long er use it - not tasksel fullfills what i want anymore :P
[07:16] <qman__> so I learned BIND, and set up local DNS to address windows being completely retarded about network names
[07:16] <qman__> and just stuck with it
[07:16] <MTeck-ricer> qman__: mine - http://www.midco.net/
[07:16] <twb> I wouldn't use BIND to resolve hosts on a private network.
[07:16] <twb> dnsmasq is prettier for that
[07:16] <unewbie1> MTeck-ricer is there another way to find where is my guest os?
[07:16] <Yosi> is it faster to resolve dns on the same box as the webserver or better performance have dns handled by a seperate company?
[07:17] <MTeck-ricer> unewbie1: twb how can you figure out the ip that a guest virt instance has?
[07:17] <qman__> well, if performance is an issue, I'd just disable name resolution on the web server itself
[07:17] <MTeck-ricer> s/unewbie1: twb/twb:/
[07:18] <twb> MTeck-ricer: erm, ask your virtualization infrastructure?
[07:18] <unewbie1> :D
[07:18] <unewbie1> that's me
[07:19] <MTeck-ricer> twb: if it were me i would have checked the mac against dhcp leases in the firewall :P
[07:19] <MTeck-ricer> hm..
[07:19] <MTeck-ricer> unewbie1: do you have access to your dhcp server so you can s
[07:20] <MTeck-ricer> unewbie1: do you have access to your dhcp server so you can see the dhcp leases? *
[07:21] <Yosi> qman, so u wouldn't run bind on the LAMP server...   my upstream provider that does our server colocation, has dns we've been using for now...  just wasn't sure if it would be faster to have godaddy pointed to BIND on my LAMP server or to my upstream provdider DNS then to my LAMP server without BIND
[07:21] <qman__> handling DNS requests for clients, let another server do the job
[07:22] <qman__> whether it's yours or the ISP's doesn't matter, but not on the web server
[07:22] <qman__> I thought you meant looking up names for firewalling or log viewing
[07:22] <Yosi> ahhh
[07:22] <Yosi> no just for the clients
[07:22] <qman__> which should be disabled on the web server, and handled in the firewall or your log viewer
[07:22] <Yosi> fair enough, makes sense
[07:23] <MTeck-ricer> unewbie1: virsh dumpxml barracuda   that'll give you teh mac addy the system is using
[07:25] <Yosi> got disconnected
[07:25] <unewbie1> MTeck-ricer:http://paste.ubuntu.com/445936/
[07:26] <Yosi> i need a simple backup solution for ubuntu server LAMP config.. tht if it crashes its fast and simple to restore..  like an image backup
[07:27] <MTeck-ricer> unewbie1: now you have the mac addy - find the ip that is assigned to it
[07:27] <unewbie1> i don't see such mac address after nmap
[07:27] <qman__> the config is all in /etc, apache2, php, mysql
[07:27] <MTeck-ricer> unewbie1: use either your dhcp server of cli tools
[07:27] <MTeck-ricer> unewbie1: I need to head to sleep now
[07:27] <qman__> the tough part to back up is the mySQL database
[07:27] <MTeck-ricer> unewbie1: you should be able to pretty easily get the rest on your own
[07:27] <Yosi> qman - will it backup live?
[07:27] <unewbie1> ok thanks
[07:27] <Yosi> mysql
[07:27] <qman__> the config, yes
[07:27] <qman__> and you can set up a slave server for mysql if yo uwant
[07:28] <qman__> which would effectively be a live backup
[07:28] <Yosi> it would be easier if i just ran ESXi on the server, then i could snapshot the drive...  but then i lose performance
[07:30] <qman__> a mysql slave server is a one-way relationship, the master updates the slave, the slave cannot update the master
[07:30] <qman__> so it would make an effective realtime backup of your database
[07:31] <Yosi> qman - interesting
[07:31] <qman__> but if the master starts spewing corrupt data, that would forward to the slave
[07:31] <Yosi> but then i need to setup anothe mysql server
[07:31] <Yosi> true..
[07:31] <qman__> so it would only be effective against failure
[07:31] <Yosi> there must be some backup software with plugins to support mysql live backup
[07:31] <qman__> however
[07:31] <qman__> you could have the slave run periodic mysqldumps
[07:31] <qman__> there probably is
[07:31] <Yosi> oy, too much work
[07:32] <qman__> but I don't know much about it
[07:34] <twb> Maybe google for "mysql replication"
[07:34] <qman__> using a mysql slave with whatever backup solution you choose might also be a good idea, performance wise
[07:34] <qman__> but only if the master slave relationship takes less resources than the backup
[07:34] <qman__> something you'd have to test
[07:35] <Yosi> have u heard of this
[07:35] <Yosi> http://www.zmanda.com/
[07:36] <qman__> mysql isn't exactly known for its scalability
[07:37] <Yosi> how does it compare against mssql?
[07:37] <qman__> in my experience, MSSQL is a joke
[07:37] <qman__> I've only ever seen it used in conjunction with other microsoft software, which requires it
[07:38] <Yosi> fair enough
[07:38] <Yosi> have u had much experience with vmware?
[07:38] <qman__> only in the 1.x versions
[07:38] <qman__> I didn't care for 2.0 and haven't paid much attention to it since
[07:39] <Yosi> ilike how simple it is to backup stuff in vmware with the snapshots, just don't liek the performance loss of virtualization
[07:39] <twb> Presumably you consider "simple" to mean "can be done with a mouse"
[07:40] <Yosi> one file backup restorre
[07:40] <Yosi> i don't mind cli, if i can backup the server live including mysql, and 1 file to restore the server
[07:41] <qman__> the problem I had with snapshots in 1.x was the fact that it had to pause operation to take them
[07:41] <qman__> they may have addressed that since, but when I was using it, that meant your server was missing for anywhere from a few seconds to a few minutes
[07:42] <Yosi> yeah in ESX 4.0, snapshots are  and can be scheduled as backups...
[07:42] <qman__> not an issue with one person tinkering on their desktop, but in a production environment, that matters
[07:42] <Yosi> live*
[08:14] <unewbie> test
[08:23] <twb> I have an 8.04 server and a bunch (say, 100) 8.04 workstations, running something along the lines of LTSP5.
[08:24] <twb> In the last six months, I've seen seven errors of the form:
[08:24] <twb> [7254]:  WARNING: gdm_slave_session_start: User passed auth but getpwnam (p97661) failed!
[08:24] <twb> Which is because the client can't see the LDAP server (nscd: nss_ldap: could not search LDAP server - Server is unavailable)
[08:25] <twb> Why would the server very occasionally by inaccessible to freshly-booted workstations?  AFAICT the server isn't rebooting or restarting slapd at the time.
[08:47] <ScottK> How's network loading?
[08:49] <twb> Well, the workstations PXE boot, then run casper, which runs a second dhclient
[08:49] <twb> I'm *pretty* sure I've scrapped NM on the workstations
[08:50] <azteech1> for some, this may be a dumb ? but, if I forget to select a server package to install, one can still access the base server, and then set up the packages after install; at least I am hoping that is the case ...
[09:51] <kaushal> hi
[09:51] <kaushal> I have issues on compiling ipp2p on Ubuntu server 10.04, can i seek help here ?
[09:53] <twb> kaushal: don't ask the same question simultaneously in multiple channels.
[09:53] <kaushal> twb: apologies
[09:55] <twb> ipp2p appears to be part of the xtables-addons-common package.
[09:58] <kaushal> twb: Thanks
[09:58] <kaushal> twb: when i run iptables -m ipp2p --help
[09:59] <kaushal> i get iptables v1.4.4: Couldn't load match `ipp2p':/usr/local/libexec/xtables/libipt_ipp2p.so: cannot open shared object file: No such file or directory
[10:01] <ScottK> azteech: Yes.  sudo tasksel should do it.
[10:01] <twb> kaushal: did you install xtables-addons-common?
[10:01] <kaushal> yes
[10:02] <twb> Then at least one of xtables-addons-common, iptables, or the kernel is too old.
[10:02] <twb> Consider migrating to a newer release of Ubuntu.
[10:04] <kaushal> twb: I am on 10.04
[10:04] <twb> Then file a bug against xtables-addons-common
[10:05] <twb> You're doing something wrong, because lucid has ipp2p in that package:
[10:05] <twb> http://packages.ubuntu.com/lucid/amd64/xtables-addons-common/filelist
[10:06] <twb> Wait, /usr/local?
[10:06] <twb> Did you compile iptables by hand?  Don't do that.
[10:08] <kaushal> twb: ok
[10:13] <mase_home> hey guys, there is a package which has given me the option of manual configuration later. ...which i chose..however now it has marked that package as not fully installed
[10:13] <mase_home> how can i mark it as such ?
[10:25] <nijaba> Hello. Anyone seen ivoks?
[10:35] <\sh> ScottK, do you think it would be a good idea to ask Christoph Haas of http://workaround.org/ispmail/lenny <- to adjust his howto for ubuntu or should we work on a similar documentation for ISP Mail setup ?
[10:37] <ScottK> \sh: I'd talk to sommer about it.  See if he wants to support it in the server guide.
[10:45] <kaushal> hi again
[10:45] <\sh> ScottK, will do :)
[11:03] <RoyK> how can I change text console resolution in lucid?
[11:10] <kaushal> ScottK: hi
[11:11] <kaushal> Thanks for your reply
[11:11] <kaushal> Still i have issues
[11:11] <kaushal> I did removed /usr/local/libexec/xtables
[11:12] <kaushal> when i ran iptables -m ipp2p --help it worked
[11:12] <kaushal> I mean it didnot worked
[11:15] <cloakable> ...
[11:16] <kaushal> cloakable: are you referring to me ?
[11:16] <cloakable> no
[11:16] <kaushal> ScottK: you around ?
[11:16] <cloakable> for 10.04, kernel devs have compiled in EVERY cpufreq driver.
[11:17] <cloakable> Thus resulting in my centrino laptop using acpi-cpufreq and not speedstep-centrino
[11:37] <sanderj_> Anyone have a document describing what the diffrence on VPS and IaaS is.. and the downsides of using both?
[11:38] <kaushal> checking in again for my query ?
[12:04] <pthsWork> Anyone know of a page that really explains the new Samba idmap system? Haven't found any that explains in a "For Dummies" style :p
[12:28] <proxee> hi... where should I go with a question about apache config on ubuntu 9.10?
[12:28] <soren> here, probably
[12:30] <proxee> thanks... I've got a server with a virtualhost that uses ProxyPass and ProxyPassReverse. I'm seeing lots of random traffic and it seems I've got an open proxy.
[12:30] <proxee> I've tried following all the documentation I can find, but the problem persists.
[12:32] <proxee> So, I guess two questions: first, how can I check that the proxy is open/closed. And second, how to close it?
[12:34] <proxee> Am I in the right place for these questions?
[12:44] <spronk> hi guys - any ideas on what log files I should be inspecting to find out why fsck gives me a clean report on boot, then booting stalls?
[12:46] <sanderj_> Anyone know what kinda functionality ubuntu elastic cloud have?
[12:46] <proxee> i guess i'm in the wrong place for apache/unbuntu questions... where should I go?
[12:47] <Jeeves_> proxee: Sometimes there aren't that many people online :)
[12:48] <Jeeves_> The sun doesn't light the entire earth at the same time :)
[12:48] <proxee> i see.. ok... I saw a long list... should I wait and ask my Q again later? (noob here)
[12:49] <Jeeves_> http://httpd.apache.org/docs/2.2/mod/mod_proxy.html#access
[12:49] <proxee> I've read the documenation... problem persists
[12:49] <Jeeves_> Then you've misinterpreted the documentation :)
[12:50] <Jeeves_> Can you paste your config on pastebin?
[12:50] <Jeeves_> !pastebin
[12:50] <proxee> probably, yeah
[12:50] <proxee> ok.. let me try pastebin
[12:51] <RoyK> hi all. trying to setup authentication against AD fails - seems bindpw is ignored in libnss-ldap.conf
[12:52] <RoyK> anyone here using ldap auth?
[12:55] <proxee> Jeeves_: thanks, here's the part of my apache config: http://paste.ubuntu.com/446061/
[12:56] <proxee> Jeeves_: I have a server running on a different port, and I'm using a proxy to route the traffic there
[12:58] <proxee> I've tried a bunch of different variations, I still see tons of random requests in my other_vhosts_access.log file
[12:58] <sanderj_> I wonder what the key diffrences between a virtual private server and a IaaS, infrastructure as a service..
[12:58] <sanderj_> VPS vs. IaaS.
[13:01] <zul> morning
[13:02] <cloakable> sanderj_: the differences between a computer and infrastructure? :)
[13:02] <cloakable> You could probably define VPS as CaaS :)
[13:03] <sanderj_> cloakable, And the C means?
[13:03] <Jeeves_> proxee: I can't browse via your proxy
[13:03] <cloakable> sanderj_: Computer :P
[13:03] <Jeeves_> I just get the site metraspace
[13:03] <proxee> so it's closed for you?
[13:04] <Jeeves_> as far as I can see...
[13:04] <Jeeves_> But i'm no expert on this :)
[13:04] <failover> proxee, there is not wrong with your conf, probably you got bots trying to look for a openproxy !
[13:04] <sanderj_> cloakable, can you explain what kinda infrastructure IaaS gives me, compeard to CaaS/VPS ?
[13:05] <proxee> yes, I see lots of junk in my log files
[13:05] <RoyK> hi all. no idea if this is the right place to ask, but trying... I'm trying to configure linux to authenticate to an MS AD server, but it seems libnss ignores binddn/bindpw and just does a 'simple' bind. this is what I see when sniffing the traffic. any idea how to force it to authenticate?
[13:05] <cloakable> sanderj_: no, because I run a home server :P
[13:05] <proxee> Jeeves_: more expert than me, for sure... how can I test whether it's open or not?
[13:05] <spronk> how can i enter like.. an extra verbose boot mode?
[13:05] <cloakable> I provide all my infrastructure myself :)
[13:05] <Jeeves_> proxee: I'm installing scanssh now
[13:10] <failover> Roky, i think you don't need a dn/pw at ldap conf to authenticate.
[13:10] <RoyK> failover: you do if anonymous browsing isn't allowed, which is the case with M$ AD
[13:11] <failover> makes sense...
[13:11] <Jeeves_> proxee: 95.142.161.84:80 <error: response code>
[13:11] <proxee> Jeeves_: ...means it's okay, or not okay? :?
[13:12] <Jeeves_> What do you see in your logs from 213.154.238.23 ?
[13:12] <Jeeves_> proxee: It looks ok
[13:13] <proxee> hard to say what's in my log from your address... the log file is growing by 10K bytes every few seconds :)
[13:15] <proxee> i have a suspicion something is still wrong... and here's why...
[13:16] <proxee> if I telnet to the http server and enter "GET http://www.yahoo.com/ HTTP/1.1", then hit return twice, I get some HTML back from yahoo.com. That seems wrong.
[13:18] <Jeeves_> No, you get
Your browser sent a request that this server could not understand.<br />
[13:18] <Jeeves_> And if you do:
[13:19] <Jeeves_> GET / HTTP/1.1
[13:19] <Jeeves_> host: www.yahoo.com
[13:19] <proxee> true, but I also get: <address>Apache/2.2.12 (Ubuntu) Server at www.yahoo.com Port 80</address>
[13:19] <Jeeves_> You get a redirect from location: http://www.yahoo.com/main.ssp
[13:19] <Jeeves_> Yes, that's your server, using the hostname you enter in the request
[13:19] <Jeeves_> (usecanonical, or something, in httpd.conf)
[13:20] <proxee> oh... that's not very obvious :D
[13:21] <proxee> ok, so that's progress... is there something else I can do to keep my log files from growing like crazy as bots are hammering on the door?
[13:22] <incorrect> I need to sign up for a bunch of ubuntu VPS providers around the world, can people suggest their fav?
[13:22] <Jeeves_> proxee: Can you pastebin some loglines?
[13:23] <proxee> incorrect: I tried gandi.net in France. They have ubuntu and a one-month refundable trial offer.
[13:23] <proxee> Jeeves_: yes, just a sec
[13:23] <Jeeves_> incorrect: You mean people supplying Ubuntu VM's/VPS's?
[13:23] <incorrect> i just wanted to know of the provider and if they used it personally
[13:24] <incorrect> thanks proxee
[13:24] <Jeeves_> incorrect: xlshosting.nl, works fine
[13:25] <proxee> incorrect: it was real easy to set up and they answered quickly to support/billing questions
[13:25] <proxee> Jeeves_: here's some log file http://paste.ubuntu.com/446079/
[13:25] <incorrect> thanks :D i have to sign up for a ton around the world
[13:26] <Jeeves_> proxee: Those are all failing requests
[13:27] <proxee> Jeeves_: i see... how do I read that they fail?
[13:27] <proxee> oh.. duh.. 404
[13:27] <Jeeves_> proxee: You can tell by the http response code .. :)
[13:28] <proxee> and what's the garbage url with "\xcf\xf2" after the response code and length?
[13:34] <proxee> I think what happened is that I had "ProxyRequests On", and some bots found my site. I figured out that was a boo boo, switched it to "Off", but the bots are still trying my site.
[13:34] <proxee> The log files are still growing like crazy, eventually overflowing the disk
[13:36] <proxee> If many requests are coming from a few addresses, is there some way I can block them and keep that from being logged?
[13:41] <Jeeves_> proxee: You might try fail2ban
[13:41] <Jeeves_> but that's kinda ugly
[13:41] <sommer> mornings
[13:42] <proxee> ok... less ugly approaches?
[13:42] <proxee> or... if I just wait, maybe they'll go away?
[13:43] <proxee> ...since I no longer have an open proxy? :D
[13:47] <ItalicBold> in a shell script, is there a way to format a command and its arguments over multiple lines?
[13:47] <Jeeves_> proxee: Probably, yes
[13:47] <Jeeves_> ItalicBold: \
[13:47] <ItalicBold> ta
[13:47] <Jeeves_> \<newline> that is :)
[13:48] <proxee> ok... I'll keep wiping the log files and hope they give up. Thanks a million for your help!!
[14:07] <ccheney> good morning
[14:11] <zul> hey ccheney
[14:15] <zul> Daviey: can you check out #590201 for me
[14:16] <lumaphyte> hi.  i have a computer that, after a reboot, suddenly is broken.  all that is displayed is  a normal fsck message for the root filesystem, after which point it simply hangs with a blinking cursor.  how can i troubleshoot this further?
[14:21] <jdstrand> naiad2: re apparmor and pids> no, but apparmor does disallow things like ptrace and access to files in /proc, as well as any tools that use them
[14:21] <jdstrand> naiad2: more specifically, if you have an apparmor profile for an application, you have to grant it permissions to do those things
[14:47] <nealmcb> lumaphyte: fscks can be very slow in lucid with ext4 - see Bug #571707 - so it could be the next fsck taking a LONG time
[14:48] <lumaphyte> oh, hmm.
[14:49] <lumaphyte> nealmcb: thanks.  i'll let it sit and see what happens.
[14:54] <lumaphyte> nealmcb: oh, it's not running fsck - the filesystem is marked as clean.
[15:32] <spineau> ttx: ping
[15:33] <ttx> spineau: pong
[15:34] <spineau> ttx: I need your help regarding a kvm potential issue with my cloud
[15:35] <spineau> ttx: instance fails to launch with the following msg in the nc.log :
[15:35] <ttx> spineau: I'll help if I can. Daviey and hallyn may also answer ^
[15:35] <hggdh> spineau: heh. exactly about that -- what machines are running what there?
[15:35] <hggdh> spineau: i.e., who runs CLC, Walrus, CC, SC, and NCs?
[15:36] <spineau> hggdh: a frontend with CLC,W,CC,SC and a separate NC
[15:37] <hggdh> tamarillos is the frontend, then?
[15:37] <spineau> ttx: libvirt: Failed to add tap interface to bridge '': No such device (code=38)
[15:37] <spineau> hggdh: yes
[15:37] <ttx> hm
[15:38] <spineau> hggdh: pricklypear the NC
[15:39] <spineau> ttx: I 've looked at the eucalyptus.conf on the NC and the VNET_BRIDGE is still empty even if I write "br0"
[15:39] <ttx> spineau: is the bridge set up on the NC ?
[15:39] <spineau> ttx: I wonder who is erasing this parameter
[15:39] <spineau> ttx: How can I check that ?
[15:40] <ttx> ifconfig should show a br0 bridge
[15:40] <spineau> ttx: yes, It shows br0
[15:40] <ttx> hm
[15:43] <ttx> spineau: so you change /etc/eucalyptus/eucalyptus.conf... stop eucalyptus CLEAN=1
[15:43] <ttx> is the setting still there ?
[15:43] <ttx> start eucalyptus CLEAN=1
[15:43] <ttx> is the setting still there ?
[15:44]  * spineau is doing the steps
[15:46] <spineau> ttx: staring eucalyptus erases the value
[15:46] <spineau> ttx: starting
[15:46] <ttx> spineau: in /etc/eucalyptus/eucalytpus.conf ? hmmm....
[15:47] <ttx> that's bad, considering we created  /etc/eucalyptus/eucalytpus.conf.local so that euca_conf avoids changing that conffile
[15:47]  * ttx grumbles
[15:48] <ttx> spineau: which mode are you using ? SYSTEM ?
[15:48] <spineau> ttx: I talk about the eucalyptus.conf on the node
[15:48] <spineau> ttx: managed-nvlan
[15:48] <spineau> ttx: novlan
[15:51]  * ttx greps in the code
[15:52] <ttx> spineau: what happens if you just do stop eucalyptus-nc / start eucalyptus-nc ?
[15:54] <spineau> ttx: I've got the same result, br0 values deleted from eucalyptus.conf
[15:55] <ttx> spineau: looks like something is running euca_conf -bridge
[15:56] <ttx> but that should not even modify /that/ file
[15:56] <hggdh> Daviey: on bug 588861 -- you reverted the UEC kernel back to lucid, and it works, or you reverted the Maverick kernel?
[15:57] <ttx> spineau: just to be sure: comment out the euca_test_nc line in /etc/init/eucalyptus-nc.conf
[15:57] <spineau> ttx: this NC has been rebooted 2 times and nothing (I mean a test) is running since start
[15:58] <ttx> and try restarting again
[15:58] <spineau> ttx: ok
[15:59] <ttx> if that still deletes it, I suspect something would show in the logs, so please pastebin the NC logs
[16:02] <spineau> ttx: value deleted
[16:02]  * spineau 's doing several scp to pastebin
[16:04] <ccheney> Daviey, i resubmitted the merge proposal with the fixes mathiaz mentioned
[16:10] <spineau> ttx: Does the whole nc.log is needed ?
[16:10] <spineau> ttx: 11 AM
[16:10] <spineau> ttx: 11 mb
[16:11] <ttx> spineau: I need the startup, until you see the VNET_BRIDGE being rewritten
[16:11] <ttx> I /hope/ that in th emiddle of those 11Mb it would mention that it rewrites a configuration file
[16:12] <ttx> spineau: but I'm puzzled as I see no reference in the code doing that. It's 1.6.2 / Lucid, right ?
[16:13] <spineau> ttx: yes
[16:16] <spineau> ttx: https://pastebin.canonical.com/33073/ for the part you wanted
[16:18] <ttx> spineau: Is this eucalyptus-nc.log on the NC ? you have the SC running on the same host ?
[16:18] <spineau> ttx: log comes from the NC. the SC is part of the frontend, another machine
[16:20] <ttx> spineau: strange line mentioning the SC there, but I don't have a running local cloud to compare
[16:21] <spineau> ttx: I have my own local cloud at home, what do you want to compare ?
[16:22] <ttx> why the logs mentions the SC
[16:23] <Daviey> ccheney: Great.. only changing debian/changelog ?
[16:23] <ttx> but the source should answer that question
[16:24] <Daviey> hggdh: Maverick install, installed the lucid kernel.. otherwise Maverick vanilla alpha 1 with updates
[16:25] <spineau> ttx: I find the same SC sentence in my nc logs
[16:25] <ttx> looks like node goes through storage.c after all
[16:27] <d1b> does anyone maintain xen packages?
[16:27] <d1b> like ppa
[16:27] <ttx> spineau: I'd suggest filing a bug... I never reproduced that and I don't get why the NC startup would rewrite that config file
[16:28] <spineau> ttx: ok, what do I need to join with such a bug ?
[16:28] <ZenMasta> I tried to isntall some packages but got an error about not locating the files or generate a list f packages http://pastie.org/995051
[16:30] <ttx> spineau: symptoms
[16:30] <ttx> topology
[16:30] <ttx> installed versions
[16:31] <ttx> Just concentrate on the startup part. You should get bridge=br0 in that vnetInit line at startup
[16:31] <ttx> something rewrites it before that... and I can't find the code in the NC that would do that.
[16:31] <ttx> spineau: I mean, you should get bridgedev=br0
[16:31] <spineau> ttx: ok, thanks a lot for your help
[16:33] <ccheney> Daviey, yea and adding the sru info to the bugs but that wasn't part of the bzr
[16:40] <Daviey> ccheney: Oh cool, you've done the SRU part aswell?
[16:40] <ccheney> Daviey, yea, still need to do the test for the SC loopback part
[16:40] <ccheney> Daviey, but already written up
[16:41] <Daviey> ccheney: Ok, i'm just about to do a quick test of your branch here.. if it turns out dandy, i'll ack the merge..  If you do need to go afk in a hurry, we'll sort out what is left.
[16:41] <ccheney> Daviey, do i just do a bunch of euca-create-volume for that?
[16:41] <ccheney> Daviey, ok
[16:41] <Daviey> ccheney: which bug #?
[16:42] <ccheney> 586134
[16:42] <Daviey> bug #586134
[16:42] <Daviey> thanks uvirtbot
[16:42] <Daviey> ccheney: yeah, that should test that
[16:42] <ccheney> looks like euca-create-volume is what i need but wasn't certain if that is 'SC'
[16:42] <ccheney> ok
[16:44]  * zul lunches
[16:45] <ccheney> apparently this is invalid syntax, but i am not sure what i am missing
[16:45] <ccheney> euca-create-volume --config -z cluster1 -s 1
[16:46] <ccheney> ah --config not needed
[16:47] <ccheney> ok seems to work fine for me
[16:51] <ccheney> Daviey, once i got to 50 i started getting errors on the SC test
[16:51] <ccheney> "1. Your proposed upload exceeds the maximum allowed object size. (edu.ucsb.eucalyptus.cloud.EntityTooLargeException)"
[16:52] <Daviey> hmmmpf
[17:00] <ccheney> i'll copy the relevant bits out of my logs
[17:08] <bogeyd6> this ubuntu desktop could do with some huge bug fixes
[17:10] <binBASH> Hi, I tried to migrate a KVM machine however it doesn't work.
[17:10] <binBASH> virsh # migrate db2 qemu+ssh://foo.bar/system
[17:10] <binBASH> Error: monitor socket did not show up.: Connection refused
[17:10] <binBASH> someone knows what could be wrong?
[17:11] <binBASH> I'm on lucid 10.04lts
[17:12] <bogeyd6> binBASH, same
[17:12] <bogeyd6> x64 of cos
[17:12] <binBASH> yup ;)
[17:14] <ccheney> Daviey, etienne mentioned there is a config option in web ui for reserved space for volumes
[17:14] <ccheney> Daviey, its set to 50GB
[17:27] <ccheney> Daviey, it seems to fail later, tracking down the reason now, it failed at 194 for me, it also looks like it might eat loopback files on failure
[17:28] <Daviey> ccheney: ahh.. interesting
[17:28] <Daviey> oh suck.
[17:28] <ccheney> "Your proposed upload exceeds the maximum allowed object size.
[17:28] <ccheney> com.eucalyptus.util.EucalyptusCloudException: Your proposed upload exceeds the maximum allowed object size"
[17:28] <ccheney> wtf
[17:29] <Daviey> ccheney: Okay.. i'm going to try this here also.. but personally.. blocking out at 194 doesn't sound too promising for a medium or large cloud
[17:29] <ccheney> Daviey, i'm going to nuke all volumes and try again after verifying the sc is empty
[17:29] <linxeh> Hi there - how do I go about setting up SSL client certificate authentication with LDAP authorisation in Apache? I've done the certificate but, but struggling to see how to combine it with LDAP
[17:30] <ccheney> Daviey, well on my test box i had it set to 200GB due to laptop disk limit, but i expected to hit at least 199
[17:30] <Daviey> linxeh: client side certificates?
[17:30] <linxeh> Daviey: yeah; I've got full mututal certificate authentication set up at the moment
[17:31] <linxeh> ie, all valid client certificates can access resources, but I want to restrict it to group information stored in an LDAP server
[17:32] <Daviey> linxeh: using mod_authz_ldap?
[17:33] <linxeh> thats what I was hoping to use yeah
[17:33] <axisys> after upgrading ubuntu from 9.04 to 9.10 proftpd is not starting anymore.. it is setup to start standalone.. is it something due to upstart ?
[17:33] <ccheney> oh no
[17:33] <axisys> need help with starting proftpd
[17:34] <Daviey> linxeh: read http://authzldap.othello.ch/mod_authz_ldap.HOWTO ?
[17:34] <ccheney> Daviey, i think, not 100% certain yet, but that when you create an unavailable/failed volume it does not get cleaned on euca-delete-volume
[17:34] <Daviey> ccheney: *sigh*
[17:35] <ccheney> Daviey, after deleting all volumes i still have 42 loops and 34 files in the volume dir
[17:35] <Daviey> ccheney: perhaps leave it 20 mins to see if it cleans itself up
[17:35] <linxeh> Daviey: that sounds like exactly what I want - thanks
[17:35] <ccheney> Daviey, ok
[17:36] <axisys> /usr/sbin/proftpd is missing
[17:36] <Daviey> linxeh: This is not something that is documented for Ubuntu Server.. it would be awesome if you could do a write up of how you did it for the docs and or ubuntu wiki.
[17:36] <Daviey> linxeh: Would you be able to do that, as you go?
[17:38] <Daviey> ccheney: I haven't read the cleanup code.. i don't know if it will do anything.. but as you've encountered, sometimes euca' has a delay on other cleanup things
[17:38] <ccheney> Daviey, ok, yea it cleaned up most of them immediately probably the ones that were working properly
[17:38] <bogeyd6> axisys, isnt proftpd an xinetd service?
[17:39] <axisys> /usr/sbin/proftpd is missing after the upgrade.. i needed to install proftpd-basic
[17:39] <Daviey> axisys: Would you mind raising a bug.. something hasn't happend that should have happend
[17:39] <linxeh> Daviey: yeah, I'll write it up (I need to do this anyway for the company wiki)
[17:39] <axisys> bogeyd6: so installing proftpd-basic was the fix
[17:40] <linxeh> Daviey: I'm pretty sure the server guide instructions for openldap are broken too
[17:40] <axisys> Daviey: asking me or linxeh ?
[17:40] <Daviey> linxeh: That is GREAT!  Really appreciated.. If you ping me, i'll certainly help point you in the direction of where to put it.
[17:40] <linxeh> Daviey: some bits work, others dont
[17:40] <Daviey> linxeh: And you'll be left with the warm feeling, knowing others will be able to find their way - based on your work \o/
[17:41] <Daviey> axisys: You, if you upgraded and proftpd isn't working as it should.. it sounds like a potential bug
[17:41] <linxeh> Daviey: yeah, I've contributed things in the past (though not always as myself, if that makes sense)
[17:41] <linxeh> :)
[17:41] <Daviey> linxeh: oh aye :)
[17:42] <dassouki> although this is not a strictly server question, but if you have chronlogical data, how can i find where queues occured?
[17:46] <ccheney> Daviey, i think the timeout isn't going to happen
[17:47] <ccheney> Daviey, at least not after ~ 10m or so anyway
[17:47] <ccheney> i'm going to reboot and see if they just go away
[17:47] <ccheney> it shouldn't be hard to reproduce the issue if it is what i think it is
[17:48] <Daviey> ccheney: agreed
[17:50] <ccheney> ok so rebooting didn't help the volumes are still all attached to loops
[17:52] <ccheney> so it doesn't know its a volume but it registers it as one anyway
[17:52] <ccheney> gar
[17:53] <ccheney> hmm i just noticed euca-describe-addresses that will be helpful to look at when i try to replicate the ip exhaustion issue again
[18:02] <ccheney> Daviey, should volumes created show up under /dev/mapper ? i see where it appears to use lvm on them
[18:03] <ccheney> it seems nothing is using them losetup lets me delete them manually
[18:07] <sommer> mathiaz: had a chance to review the openldap-dit branch?
[18:07] <mathiaz> sommer: not yet
[18:07] <sommer> mathiaz: okay no problem
[18:08] <axisys> Daviey: ok
[18:19] <ccheney> how do i properly delete a volume that is not showing up in euca?
[18:26] <ccheney> isn't it something like /etc/init.d/eucalyptus start CLEAN=1 ?
[18:33] <kaushal> what does ipp2p mean
[18:33] <hggdh> ccheney: euca-describe-volumes does not show it?
[18:33] <kaushal> I know p2p means peer to peer
[18:33] <kaushal> what does ip mean in ipp2p ?
[18:33] <ccheney> hggdh, yea
[18:33] <ccheney> hggdh, i think i found a bug
[18:33] <ball> p2p over ip?
[18:33] <hggdh> ccheney: there is bug on it already opened...
[18:33] <ccheney> hggdh, i ended up removing the /var/lib/eucalyptus/db/storage.script
[18:33] <ccheney> hggdh, oh do you happen to know the number?
[18:33] <kaushal> ball, what does that exactly mean ?
[18:34] <hggdh> ccheney: just a sec
[18:34] <ball> kaushal: It does what it says on the box.
[18:34] <ccheney> hggdh, thanks
[18:36] <hggdh> ccheney: bug 517086
[18:36] <kaushal> ball, Thanks
[18:36] <panfist> does ubuntu-server have any kind of network manager like the desktop edition?
[18:38] <ccheney> hggdh, yea i bet that is at least part of the problem
[18:38] <malchias> I am trying to setup basic email sending (not an smtp server, or relay).  I want scripts to be able to use something to send email.... can someone direct me to a tutorial on something simple, and secure?
[18:39] <ccheney> hggdh, i don't see any error communicating when creating the volumes but that might be why they don't get deleted properly
[18:39]  * ccheney looking into why they don't get created
[18:39] <hggdh> ccheney: I think this is part of the issue -- I had a quick test some days ago on the Dell rig -- created 512 volumes, and then deleted them all
[18:40] <hggdh> ccheney: I was left with 70 VGs hanging
[18:40] <ccheney> hggdh, yea, i created 200 and only 179 were created
[18:40] <ccheney> hggdh, before when i then removed them a lot were left behind, which may be the error communicating issue
[18:40] <ccheney> hggdh, something else appears to be going wrong with it not letting me create the volumes initially
[18:40] <hggdh> ccheney: what euca version? One of the issues I found there was OOM
[18:41] <ccheney> hggdh, i seem to be getting an exception on vgcreate
[18:41] <hggdh> ccheney: Java OutOfMemory, the Dell rig was not running 30.1, so it did not have a larger -xmms at start
[18:41] <hggdh> ccheney: this is different
[18:41] <ccheney> hggdh, i'm running the lucid update plus fixes i made for the new sru
[18:42] <hggdh> ccheney: what will be 30.2?
[18:42] <hggdh> or whatever
[18:42] <ccheney> yea
[18:42] <ccheney> hggdh, http://pastebin.ubuntu.com/446228/
[18:42] <ccheney> hggdh, thats an example of a failed creation
[18:43] <ccheney> see the pool-10-thread-4
[18:43] <ccheney> hggdh, is there some way to make it show what vgcreate returned?
[18:44] <hggdh> ccheney: not to my knowledge. But this is indeed a different error
[18:44] <ccheney> hggdh, ok
[18:47] <hggdh> ccheney: if you still have the allocation... what happens if you run the vgcreate by hand?
[18:47] <ccheney> it looked like it ellipsed a bit but i can try to see what happens
[18:48] <ccheney>   Volume group "vg-6E6tMQ.." successfully created
[18:48] <ccheney> apparently the ellipse really is part of it
[18:49] <ccheney> i didn't just run vgcreate i ran exactly the same code it tried
[18:49] <ccheney> /usr/lib/eucalyptus/euca_rootwrap vgcreate vg-6E6tMQ.. /dev/loop93
[18:50] <ccheney> which was what returned the above
[18:52] <hggdh> time to see exactly what rootwrap does with stderr/stdout, I guess...
[18:54] <ccheney> i'm deleting all volumes now to see if i can get the other bug to show up
[18:54] <ccheney> it left 60
[18:55] <ccheney> so far no results for grep -i "commun"
[18:55] <ccheney> so maybe another failure case as well
[18:58] <ccheney> hmm and it doesn't even log it didn't work
[18:58] <ccheney> there are 40 vg's relating to the volumes still active
[18:58] <hggdh> heh. rootwrap fires off a execv
[18:59] <sourcode> จาเอา iphoneeeeee
[19:00] <ccheney> hggdh, it let me run losetup -d on all but a few
[19:00] <ccheney> hggdh, which then made the vg's disappear
[19:00] <ccheney> left 5
[19:00]  * ccheney looking at those 5 in the log now
[19:04] <ccheney> hggdh, ah it also failed to lvremove later
[19:04] <ccheney> hggdh, when i ran the same command again it works fine
[19:05] <hggdh> ??
[19:05] <hggdh> this is weird
[19:05] <ccheney> eg: /usr/lib/eucalyptus/euca_rootwrap lvremove -f /dev/vg-qJp0gg../lv-XL682g..
[19:05] <ccheney> maybe i should run memtest on my boxes to make sure they aren't flaky, unless you also see this problem?
[19:05] <hggdh> ccheney: I do not have enough space to test :-(
[19:06] <hggdh> and the Dell rig is busy now
[19:06] <ccheney> ok
[19:10]  * ccheney lunch
[19:12] <RoyK> does ext[234] have anything like zfs in the means of SSD buffering?
[19:17] <ccheney> hggdh, it seems to setup them up sparse, so doesn't use any space until you actually use it
[19:18] <hggdh> ccheney: yes, they are set sparse... let me try
[19:18] <ccheney> it seemed to not take any space for 200 1GB files for me :)
[19:20] <hggdh> ccheney: did you just loop creating the volumes?
[19:20] <ccheney> hggdh, yea
[19:20]  * hggdh wants to replay as much as possible
[19:20] <hggdh> ccheney: roj
[19:20] <ccheney> for i in `seq 1 200` ; do euca-create-volume $i ; done
[19:21] <ccheney> er not exactly that
[19:21] <ccheney> for i in `seq 1 200` ; do euca-create-volume -z cluster1 -s 1 ; done
[19:24] <hggdh> ccheney: yeap, pretty much the same I am running. Oh lack of imagination... ;-)
[19:24] <ccheney> i don't put any sleep in there, not sure if it is a timing issue if so it still needs fixing :)
[19:25] <hggdh> mine is for i in $(seq 1 200); do euca-create-volume --config=/home/ubuntu/eucarc -s 1 -z UEC-TEST1; done
[19:27] <hggdh> ccheney: do you notice that euca-describe-volumes takes a long time to process?
[19:27] <ccheney> hggdh, i don't remember if it did or not
[19:27] <ccheney> hggdh, probably was a little slow
[19:28] <hggdh> ccheney: mine time-ed at 31 secs to show 200 volumes (in a mix of created/creating/failed)
[19:29] <ccheney> ah ok, i didn't actually time it, it might have been a while i was looking in another window while running it
[19:29] <hggdh> sorry, creating/available/failed
[19:30] <hggdh> 52 secs second run
[19:31] <hggdh> ccheney: 70 failures, so far
[19:31] <ccheney> ok
[19:31] <ccheney> i'll file a bug once we can reproduce it properly on your box
[19:32] <ccheney> i just noted the issues in the sc loop bug up til now
[19:34] <Yosi> hi all
[19:34] <Yosi> whats the comman to install the latest webadmin in a fresh copy of ubuntuserver
[19:34] <Yosi> commdn*
[19:34] <Yosi> command*
[19:35] <Pici> !webmin | Yosi
[19:35] <Pici> (there isn't one)
[19:35] <hggdh> ccheney: weird. I am getting thousands of cloud-debug.log:14:32:20 ERROR [SystemUtil:pool-10-thread-5] com.eucalyptus.util.ExecutionException: ///usr/lib/eucalyptus/euca_rootwrap losetup -f
[19:35] <Yosi> ahhh, darn, i tried it before and it seemed to work fine....
[19:35] <ccheney> hggdh, do you have the 30.2 patches to allow over 32?
[19:36] <Yosi> how does ebox compare?
[19:36] <ccheney> hggdh, https://code.launchpad.net/~ccheney/ubuntu/lucid/eucalyptus/lucid-sru/
[19:36] <hggdh> ccheney: ccheney oh crap. I reinstalled, and forgot to manually set it
[19:36] <ccheney> hggdh, ah ok :)
[19:37] <ccheney> hggdh, the loopback part is easy to manually do
[19:37] <hggdh> ccheney: yeah, will get it done now, and bounce the SC
[19:38] <hggdh> ccheney: OK, done & restarted, will test again
[19:39] <ccheney> ok
[19:39] <ccheney> you might want to make sure the vg's are gone, etc
[19:39] <ccheney> otherwise they seem to stick around, which i think is part of the bug :-\
[19:40] <Pici> Yosi: I've never really used either of them, so I can't compare, sorry.
[19:41] <mathiaz> Daviey: https://ubuntumathiaz.wordpress.com/2008/09/18/automate-ubuntu-server-iso-testing/
[19:41] <mathiaz> Daviey: ^^ this is one of my blog post about my workflow
[19:43] <zul> mathiaz: hi https://code.edge.launchpad.net/~zulcss/+junk/server-sru-tracker
[19:43] <mathiaz> zul: cool - thanks
[19:43] <zul> np
[19:45] <hggdh> ccheney: for reference: no volumes -- time euca-describe-volumes clocks at .15s
[19:46] <ccheney> ok
[19:47] <hggdh> I am now creating 200 vols of 1G
[19:48] <hggdh> ccheney: 200 vols created, 11 failures. Looking at them now
[19:48] <ccheney> ok
[19:51] <hggdh> ccheney: http://pastebin.ubuntu.com/446252/
[19:52] <ccheney> hggdh, looks the same as mine
[19:52] <hggdh> yup
[19:52] <ccheney> vgcreate fails for unknown reasons (or euca thinks it did anyway)
[19:52] <hggdh> yes. Now... let me see if the VG was (or was not) actually created
[19:53] <hggdh> no, it does not exist (at least as far as vgdisplay is concerned
[19:54] <ccheney> maybe we actually found some sort of race in vgcreate
[19:56] <hggdh> yes
[19:57] <hggdh> now, the second test -- removing all volumes
[19:57] <Yosi> just tried ebox and it sucks!  is there anyway to make webadmin work with server 10.04
[19:57] <Yosi> ?
[19:58] <ccheney> hggdh, i wonder if lvm fails if you are currently running another lvm command (not sure if that is what we are seeing though)
[19:59] <hggdh> a race in lvm?
[20:00] <hggdh> ccheney: I just tried to remove all volumes, was left with 8 vgs
[20:00] <ccheney> yea
[20:00] <ccheney> hggdh, and you can probably manually remove all of them using the same command that euca tried to do
[20:01] <hggdh> ccheney: oh yes, nothing that a losetup -d, etc could not resolve
[20:02] <hggdh> ccheney: but the point is euca-describe-volumes shows *NO* volumes
[20:02] <hggdh> and there are still 8 VG/loop/files there
[20:04] <ccheney> hggdh, yes, it appears it knows that it failed to remove but then forgets about it in euca-describe-volume
[20:05] <ccheney> btw if you did not remove them manually then euca would set them back up in the backend somehow after restarting but still not show them in euca-describe-volume
[20:05] <ccheney> i noticed that when doing testing earlier
[20:05] <ccheney> it notes that it is doing it in the log files
[20:06] <hggdh> yes. This, I think, is bug 517086 (or realted to it)
[20:08] <ccheney> hggdh, did you see any of the 'Error communicating' messages? i don't recall seeing them on my box
[20:09] <hggdh> let me search
[20:11] <hggdh> no, I do not see it
[20:11] <hggdh> hum
[20:11] <hggdh> Yet Another Bug?
[20:13] <ccheney> probably :-\
[20:13] <ccheney> so i'll file the bug and you can add whatever you want to it, i'll note it seems similar to the other bug but doesn't appear to be the same issue
[20:14] <ccheney> i'll file it against eucalyptus until we can figure out why lvm is causing problems
[20:18] <ccheney> bug 590929
[20:20] <ccheney> added part about it not showing up when delete fails
[20:22] <hggdh> ccheney: and I just confirmed/Medium it
[20:22] <ccheney> ok
[20:22] <ccheney> added it for lucid also
[20:22] <hggdh> (and subscribed mesself)
[20:24] <ccheney> ok
[20:27] <TuxIce> Anyone got an idea why this is happening? "Permission denied (publickey).". I'm thinking its a public key permissions issue, server side. My public key is copied to the server under files authorized_keys and id_rsa.pub, permissions 0600 owned by my server user. id_rsa.pub is not in my home folder on my client.
[20:28] <TuxIce> ssh adam.delvecchio@server.tld -p 8022 -v produces: http://pastebin.com/SBmbycLQ
[20:29] <Yosi> is there a simple Hard Disk/SSD benchmark took that i can run from the cli in ubuntu server
[20:29] <Yosi> tool*
[20:42] <sjm> TuxIce: not sure right now, but using "server.tld" doesn't hide much when the information is in the pastebin :)
[20:43] <Hypnoz> this doesnt seem to work huh :(  tar -C /home/ -xvf http://10.1.1.1/pxe/net.tar
[20:56] <qman__> Yosi, hdparm -tT
[21:02] <amstan> hey guys, i'm trying to get suexec/suphp working, none of them seem to work
[21:02] <amstan> do you guys know of any good guides? besides that perfect server one
[21:07] <ccheney> seems my wife's contractions just stopped happening, they were happening regularly for over 12 hours and vanished, very weird
[21:11] <lifeless> ccheney: good luck
[21:12] <ccheney> lifeless, thanks :) i now have no idea when we will be going to the hospital i was sure it was going to be sometime this afternoon, heh
[21:30] <sentabi> how to fix this error : http://pastebin.com/3azNXCWD
[21:32] <amstan> sentabi: what are you trying to do? on what version of ubuntu?
[21:34] <SpamapS> amstan: on your suexec/suphp question .. i'd say (and I know this sucks) that its one of the worst ideas ever.. and you're better off with virtualization. :)
[21:36] <SpamapS> sentabi: I just installed bin9, same version, worked fine (on maverick)
[21:36] <amstan> SpamapS: virtualization?? are you crazy?
[21:37] <amstan> what i'm trying to do here is making sure none of my clients can use php to read/modify other's files
[21:37] <amstan> for example, right now i can just use php to read the config.inc.php from another site and find out the passwords
[21:37] <amstan> i have like 40 clients, with very simple sites
[21:37] <amstan> no way i'm getting 40 virtual machines
[21:41] <sentabi> amstan : ubuntu server 10.04
[21:41] <sentabi> installing bind9
[21:41] <amstan> weird..
[21:41] <amstan> i just installed the same thing today, do you have any special setups that might make that not work?
[21:41] <sentabi> no, i dont
[21:42] <amstan> sentabi: try redownloading it
[21:42] <sentabi> let me try to reinstall the OS :) thanks
[21:42] <amstan> it seems like the the init script is missing when the installer wants to execute it
[21:43] <amstan> sentabi: check the cd first then
[21:45] <SpamapS> amstan: Nah its not that crazy these days. Virtualization is easy and lets you not care so much about the little things. ;)
[21:45] <amstan> SpamapS: still.. 40 VMs is not a pretty sight
[21:45] <SpamapS> amstan: maybe if you're hand rolling Xen. ;)
[21:46] <amstan> SpamapS: how much time would it take to setup such a thing?
[21:46] <amstan> i only have like 2 days
[21:47] <SpamapS> amstan: should be very quick if you go with UEC .. but you'll have to devote a large swath of IPs, which you may not have. ;)
[21:47] <amstan> oh yeah.. this all needs to be done using virtual hosts
[21:47] <amstan> so... 1ip/vm is not cool
[21:47] <SpamapS> amstan: you can still do that, with a simple reverse proxy.
[21:48] <SpamapS> and that makes them portable, which is nice.. you can spawn a new instance on another machine and move it without the client knowing. ;)
[21:48] <amstan> if suphp/suexec would of been working i would of been done in 10 minutes
[21:48] <amstan> with my current setup
[21:48] <SpamapS> yeah, I've never seen suphp/suexec work right
[21:48] <SpamapS> I've seen it "work"
[21:48] <SpamapS> but not without a lot of duct tape
[21:48] <amstan> there must be a way, there's tons of webhosts out there, and i really don't believe they have 1 ip per custommer
[21:53] <SpamapS> no, it works, but its a constant headache because it breaks some assumptions
[21:53] <amstan> SpamapS: oh, of course, i noticed this first hand
[21:53] <SpamapS> and you can always, very easily, run a reverse proxy to unify them all under 1 ip.
[21:53] <amstan> but those assumptions are way less important than getting security done
[22:00] <ccheney> hggdh, i'm not completely sure but it looks like euca_rootwrap should be causing the error if any from lvm to be displayed, it calls perror() at the end after running the command
[22:01] <hggdh> ccheney: I thought perror() would just return the errno
[22:01] <ccheney> hggdh, actually hmm it seems it just prints the output of errno, which wouldn't be sufficient for this i think
[22:01] <hggdh> if any
[22:01] <ccheney> hggdh, yea i just realized that :)
[22:02] <ccheney> i guess i could wrap the calls with a script that logs itself
[22:02] <hggdh> ccheney: I did not see anywhere where stdout/err is saved -- and the thing is execv-ed off :-(
[22:02] <ccheney> yea
[22:02] <ccheney> so i could move lvremove to lvremove.real and have a script log it all i think?
[22:02] <hggdh> er
[22:03] <hggdh> yeah, it might work
[22:03] <ccheney> do i need to do something to have a script return the code from the thing i am calling?
[22:03]  * hggdh was worried on being root, but permissions should be maintained there
[22:03] <ccheney> exit $var ?
[22:03] <ccheney> and save $? to var while doing any processing in between?
[22:04] <hggdh> probably easiest would be to write a small shell hack to save it to a log
[22:04] <hggdh> append to the log
[22:05] <ccheney> hmm its return $var for shell, exit is c
[22:05] <ccheney> ok
[22:05] <ccheney> i'll set that up and see how it does
[22:05] <hggdh> cool. I am curious
[22:05]  * ccheney goes to find a snack before running the test
[22:06] <hggdh> oh. Good idea. This is what is probably causing this strange feeling of emptyness on self
[22:07] <ccheney> i didn't end up eating anything for lunch and forgot to actually do so due to my med side effect
[22:11] <ccheney> hggdh, as its all one real binary i can catch all lvm calls with the same test
[22:11] <ccheney> :)
[22:12] <hggdh> yes!
[22:21] <SpamapS> anybody have a good suggestion for a really low power low footprint piece of hardware I could put in my home office to plug a couple of giant USB drives into for the purpose of running a local archive mirror?
[22:22] <qman__> SpamapS, probably a netbook
[22:22] <SpamapS> interesting.. but.. so uninspired. ;)
[22:23] <qman__> you can build an atom machine, too, but they're a lot bigger
[22:23] <SpamapS> I was thinking more like converting a wrt54g or something
[22:23] <SpamapS> you know, something *crazy*
[22:23] <qman__> routers with USB are quite expensive
[22:23] <qman__> and finding one that works with dd-wrt is a crapshoot
[22:24] <SpamapS> I wonder if one can still find xbox's cheaply? I bet that would owrk.
[22:24] <SpamapS> but probably not low power
[22:24] <qman__> not exactly low power
[22:24] <qman__> it's a pentium 3 system
[22:25] <qman__> but it would work
[22:25] <qman__> but they're not exactly small either
[22:26] <qman__> might as well just dumpster dive for a P3 rig and put the drives inside it
[22:28] <ccheney> grr i messed up euca-desribe-volumes somehow
[22:28] <ccheney> my script was a bit buggy and then i fixed it
[22:29] <ccheney> but now euca-describe-volumes just hangs :-\
[22:30]  * ccheney hopes rebooting will fix whatever went wrong
[22:35] <ccheney> hggdh, ugh replacing the file is a bit more complicated than i thought
[22:35] <ccheney> hggdh, because when you do that it messes up lvm's idea of what $0 is
[22:36] <ccheney> hggdh, i may have to change all the symlinks to be the script
[22:36] <hggdh> ccheney: ugh
[22:37] <hggdh> ccheney: OTOH, this will catch all errors, and right now I am not sure where else we are failing
[22:37] <ccheney> i don't know of way to convince lvm.real that its $0 is something else inside a script at least
[22:37] <hggdh> shift?
[22:37] <hggdh> then call lvm.real with the rest of the arguments
[22:39] <hggdh> no
[22:39] <ccheney> yea not sure if that will work
[22:40] <ccheney> hmm actually it may be slightly more complicated than i thought
[22:42] <ccheney> gah
[22:42] <ccheney> stupid thing
[22:42] <ccheney> lvm vgscan won't work unless the vgscan symlink is pointing at the right binary
[22:43] <incorrect> is ebox any good?
[22:43] <ccheney> it probably reexecs itself
[22:45]  * ccheney thinks to see if there is any way to hook into the calls
[22:46] <ccheney> erg, i can't think of a way to make this work due to the crazy lvm thing :-\
[22:46] <mjmac> has anyone had much luck with installing 10.04 from a USB stick?  i've done it, but it required a lot of ugliness and beatings with a wrench.  wondering if i'm failing to see an easier way to do it.
[22:46] <mjmac> 10.04 server, that is
[22:46]  * ccheney tries one file to see if it does anything
[22:47] <funkyHat> Hi, this is my smb.conf http://pastebin.com/cfELhvNe -- I've set it up to allow access to home dirs (I thought after supplying a password), but it appears I only have to have the right username on the host and I can access the corresponding ~... anyone able to spot what I've got wrong?
[22:48] <funkyHat> My config for the "homes" share starts at line 236
[22:49] <mjmac> i did it before by extracting the ISO contents onto the USB stick and using grub1 to boot the installer...  thought i'd try grub2 and a loopback mount of the ISO.  i can boot into the installer and get things going after manually loopback-mounting the ISO in the initrd, but that's ugly
[22:49] <hggdh> ccheney: all hail security ;-)
[22:49] <hggdh> there must be a way...
[22:49] <NG_> sup guys?! how can i change an ubuntu server kernel to the latest version w/o restarting the whole system?
[22:49] <hggdh> NG_: you cannot.
[22:49] <lifeless> you can't
[22:49] <NG_> hggdh, i'm fucked :>
[22:49] <NG_> :<
[22:49] <lifeless> we don't generate live-patches
[22:49] <hggdh> NG_: you might be, but please mind your language
[22:50] <ng_> soz
[22:50] <ng_> :/
[22:50] <kees> ng_: if you want, see www.ksplice.com
[22:51] <mjmac> ng_: there's ksplice, but it ain't free
[22:51] <ng_> read about it on google
[22:51] <ng_> running a server w/o x
[22:51] <ng_> so not what i'm looking for, as far as i saw
[22:51] <mjmac> why do you need x to update the kernel?
[22:51] <kees> ng_: it doesn't need x.
[22:51] <mathiaz> zul: looking that the sru-report branch
[22:51] <mathiaz> zul: I've generated a report:
[22:52] <ccheney> kees: do you know of any way to make a binary think its running as a different program name when running it, besides the symlink trick?
[22:52] <mathiaz> zul: http://people.canonical.com/~mathiaz/report.html
[22:52] <mathiaz> zul: so all the rows are in red - which is not really useful
[22:52] <ng_> kees, looking deeper into it, thanks =)
[22:52] <kees> ccheney: depends on the method it looks at itself.  what is the specific issue?
[22:52] <mathiaz> zul: I've updated the Spec page to outline what the output should look like
[22:52] <kees> ng_: http://www.ksplice.com/pricing  ubuntu is free
[22:52] <mathiaz> zul: https://wiki.ubuntu.com/ServerMaverickSruProcess#preview
[22:53] <kees> ng_: (well, desktop is free, but you can try it)
[22:53] <ccheney> kees, i need to log lvm output and its being called with execv in eucalyptus, i tried using a wrapper script and moving lvm to lvm.real but that obviously wouldn't work
[22:53] <mathiaz> zul: could you update the scripts to generate a page using the format outlined in the specification (under the SRU Tracking page section)?
[22:54] <ccheney> kees, lvm seems to randomly not work when i am doing volume manipulation so i need to see what exactly it says for each lvm command run
[22:54] <kees> ccheney: why doesn't the lvm -> lvm.real thing work?
[22:54] <ccheney> kees, maybe i just don't know how to run lvm and have it work right?
[22:55] <kees> oh, lvm re-writes its stuff based on the calling name.
[22:55]  * kees ponders
[22:55] <ccheney> pastebin's his script
[22:55] <ccheney> kees, http://pastebin.ubuntu.com/446318/
[22:55] <kees> ccheney: doesn't euca use a wrapper to call lvm?  you could hook that?
[22:55] <ccheney> the wrapper is a c program that calls execv directly to lvm
[22:56] <ccheney> when i call lvm.real in my script it loses its $0 identity as eg vgscan
[22:56] <mjmac> so, i'm guessing by the deafening silence that no one watching this channel has done server installs from USB...  :)  kinda surprised, it's so handy to install from USB when messing with stuff.  i guess i'll set up netboot.
[22:56] <kees> ccheney: but it reads commands to run from /etc/euca/something.conf
[22:56] <ccheney> and calling lvm.real with eg vgscan doesn't help
[22:56] <kees> ccheney: you could change the "lvm" target to be lvm-logger and then call lvm from lvm-logger
[22:58] <kees> ccheney: /etc/eucalyptus/wrappers.conf and change "/sbin/lvm" to point to your script, and have your script call /sbin/lvm
[22:59] <ccheney> kees, yea i think that will work, just need a symlink to the script for each lvm related binary call (i think)
[22:59] <hggdh> ccheney: what if you mv /sbin/lvm /bin/lvm?
[22:59] <kees> ccheney: (you'll have to change all them, though)
[22:59] <hggdh> forget, kees has the answer
[22:59] <ccheney> hggdh, yea
[22:59] <ccheney> kees, yea ok, will do that seems easier than finding and implementing any alternative to that :)
[23:00] <ascheel> Quick question.  Upgraded Karmic server to Lucid server and I now can no longer boot.  No questions were ever asked regarding my boot order or fstab, so it should have ran unmodified.  Any ideas on recovery?
[23:00] <SpamapS> I just love rebuilding php5 .. 45 minutes now.. :-P
[23:01] <ZenMasta> my computer is not getting an ip, when I view ifconfig it just shows 127.0.0.1. If I try to edit /et/network/interfaces and restart it says error...no such device failed to bring up eth0
[23:01] <SpamapS> ZenMasta: maybe your network interface wasn't detected.
[23:02] <SpamapS> ZenMasta: is this a hardwired connection?
[23:03] <ZenMasta> SpamapS well that's what I'm assuming, it is a wired connection. The thing is though, I have used live cd's and other distros on this box before without problems
[23:04] <ZenMasta> btw this is a new install of 10.04
[23:04] <ZenMasta> well maybe I shouldn't go as far and say problems in general, just not this problem
[23:06] <funkyHat> huh. browsable is spelt wrong in my smb.conf
[23:07] <SpamapS> ZenMasta: how about 'grep eth0 /var/log/dmesg' ?
[23:07] <funkyHat> Or at least differently to how it is in the samba docs
[23:07] <SpamapS> ZenMasta: should produce something like [    2.947822] e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
[23:08] <ZenMasta> lemme try, gotta reconnect monitor to try
[23:08] <SpamapS> ZenMasta: also while you're at it 'lspci | grep Ethernet'
[23:10] <ZenMasta> its running... is there supposed to be an apostrpohe after dmesg
[23:10] <ccheney> kees: cjwatson pointed me at exec -a NAME which works also :)
[23:10] <ccheney> hggdh, see ^
[23:10] <SpamapS> ZenMasta: no
[23:10] <ZenMasta> oh okay hehe
[23:11] <ccheney> apparently its a bashism but is good enough for my hack
[23:11] <hggdh> heh
[23:11] <SpamapS> ZenMasta: if it didn't show anything, then there's no driver most likely.
[23:11] <SpamapS> ZenMasta: lspci should show your network card though
[23:12] <SpamapS> 02:01.0 Ethernet controller: Intel Corporation 82545EM Gigabit Ethernet Controller (Copper) (rev 01)
[23:12] <ZenMasta> they both say stuff.. looks kinda standardish
[23:12] <SpamapS> something like that
[23:12] <ZenMasta> the first one though, grep dmesg that one says renamed eth0 to eth1
[23:13] <SpamapS> ZenMasta: ding ding ding
[23:13] <SpamapS> ZenMasta: maybe try eth1? ;)
[23:14] <ZenMasta> I don't understand... shouldn't it just auto detect? I just installed so the default interfaces file says auto eth0
[23:14] <SpamapS> well does it give a reason for renaming eth0 to eth1?
[23:14] <ZenMasta> lemme check again
[23:15] <hggdh> zul: on the samba apport hook all we need is *not* to open a bug if /etc/samba/smb.conf is missing, correct?
[23:16] <mathiaz> hggdh: well - if /etc/samba/smb.conf is missing *and* the package has been removed (but not purged)
[23:16] <mathiaz> hggdh: if the package is removed, configuration files are *not* removed
[23:16] <mathiaz> hggdh: if the smb.conf file is deleted by hand, upon *reinstallation* of the package smb.conf will *not* be reinstalled
[23:17] <ccheney> still having some weird issue but maybe reboot will fix it this time
[23:17] <mathiaz> hggdh: which leads to smb not starting, which means that the package fails to install correclty and thus opens a bug
[23:17] <ZenMasta> SpamapS no reason but I edited the interfaces file to auto eth1 instead of eth0 and it has now obtained an ip
[23:18] <ZenMasta> its not showing up as an attached device on my router and I can't ssh into it yet though :/
[23:18] <mathiaz> ZenMasta: has the MAC address of the interface changed in between?
[23:19] <mathiaz> ZenMasta: if eth0 doesn't work but eth1 works then it usually means that the mac address has changed somehow
[23:19] <ZenMasta> mathiaz possibly i installed it on one computer and then moved the hd to another (exactly same processor mobo ram etc
[23:19] <hggdh> mathiaz: please let me read it slowly ;-)
[23:19] <mathiaz> ZenMasta: see /etc/udev/rules.d/70-persistent-net.rules
[23:19] <mathiaz> ZenMasta: for the mapping of mac address <-> ethX
[23:20] <mathiaz> ZenMasta: right - the mac address is different then
[23:20] <mathiaz> ZenMasta: which is why the new network interface is showing up as eth1
[23:20] <ZenMasta> I see. Well I've never encountered that before. I've done the hard drive switcheroo thing a couple of times and has not presented this problem before
[23:22] <SpamapS> mathiaz: I didn't know we did that.. bleh.
[23:23] <mathiaz> SpamapS: yeah - I've run into that issue when cloning vms
[23:23] <hggdh> mathiaz: so, pretty much, smb.conf will *only* not exist in this scenario?
[23:24] <kees> ccheney: ah!
[23:24] <SpamapS> RHEL 5 introduced it to the RH world and by 5.1 it was no longer the default behavior.
[23:24] <SpamapS> Bites too many people when they do exactly as ZenMasta has done.. move the drives to another machine.
[23:25] <SpamapS> In fact, thats a pretty standard thing to do when you've lost something on the motherboard.. move the disks to identical cold hardware and boot.
[23:29] <hggdh> mathiaz: and the solution is, also pretty much, apt-get purge samba && apt-get install samba?
[23:29] <ZenMasta> I'm still not sure it is working though. ifconfig shows an ip now, but I still can't ssh into it, nor do I see it in my router backend
[23:31] <Yosi> basic question:  how do i shutdown ubuntu server, if i use the "shutdown now" command it brings me to a recovery menu
[23:34] <SpamapS> Yosi: right, you need to use shutdown -h now, or 'halt'
[23:34] <Yosi> ahh... why halt?
[23:35]  * Ng eyes NG_ suspiciously
[23:35] <mathiaz> hggdh: yes
[23:35] <hggdh> mathiaz: thank you
[23:35] <ZenMasta> SpamapS is it possible that I have to configure the firewall to allow ssh? I still can't connect or see the computer on my router
[23:35] <mathiaz> hggdh: make sure to get your apport hook reviewed by zul or ttx
[23:35] <hggdh> mathiaz: roger wilco
[23:36] <SpamapS> Yosi: because thats how 'shutdown' works.. its shutting down the system for maintenance/whatever .. not necessarily halting it.
[23:36] <SpamapS> ZenMasta: quit possible yes
[23:36] <ZenMasta> SpamapS how can I check and if its not enabled how do I do so?
[23:36] <Yosi> ahh thanks!
[23:41] <ccheney> this script was way too complicated, needed several bash specific features to make it work :-\
[23:44] <ZenMasta> hmm, ssh wasn't installe.d I coulda sworn I marked it during install
[23:47] <ccheney> hggdh, my logging is a bit messed up (i think i can fix it) but i think i see why it fails
[23:47] <ccheney>   /var/lock/lvm/P_orphans: flock failed: Resource temporarily unavailable
[23:47] <ccheney>   Can't get lock for orphan PVs
[23:49]  * ccheney wonders if this is somehow related to the reason uec-testing doesn't hit 100%
[23:51] <hggdh> ccheney: I do not believe it is why we cannot get 100% on UEC, since we were not testing volumes, and the lock -- by the name -- relates to LVM
[23:51] <hggdh> ccheney: but it sounds like resource starvation
[23:52] <ccheney> ok
[23:52] <ccheney> i think i fixed my logging i am going to run it again and see how it goes
[23:52] <hggdh> cool. If you need, I can play with it also
[23:54] <hggdh> ccheney: it may be that the man page is out of date, but man 2 flock does not list this errno
[23:54] <IPNixon> hey, i've got 10.04 server installed, and i'd like to run a command at startup.  what are the steps i have to take to get a shell script to run automatically at startup?
[23:55] <ccheney> hggdh, yea my logging might be corrupt atm
[23:55] <ccheney> hggdh, something is returning err 5
[23:55] <panfist> does anyone know where the dhcp server logs activity?
[23:55] <hggdh> ccheney: input/output err??
[23:56] <ccheney> hggdh, multi threaded logging into one file :)
[23:57] <hggdh> ccheney: oooohhh, fun...
[23:58] <ccheney> i think i know how to make it work or at least less likely to mix but running into an issue