[00:01] <SpamapS> MTecknology: disks are cheap
[00:01] <SpamapS> MTecknology: why would you send stuff over the wire?
[00:01] <maek> anyone using ubuntu in a PCI compliant environment?
[00:02] <SpamapS> MTecknology: If you can't afford proper tape backup, just get yourself an external enclosure with hot swap drives of some kind.. backup to them, take one offsite every once in a while.. you'll be fine. But don't waste your time trying to exclude crap. It will bite you.
[00:02] <MTecknology> SpamapS: my backup server is one place. my production servers are elsewhere
[00:03] <MTecknology> SpamapS: when we get more money our dev system will be the backup system
[00:03] <MTecknology> for now - I'm stuck on a crappy budget
[00:03] <SpamapS> MTecknology: understood. When you're starting out especially, you have to take those risks. ;)
[00:04] <SpamapS> MTecknology: its good to actually sit down and think about what the risks are though.
[00:16] <zul> mathiaz: do you want to merge puppet or do you want me to?
[00:17] <MTecknology> zul: how much different is puppet? I hjave yet to see it
[00:21] <zul> MTecknology: dunno i havent quite looked at it yet
[00:22] <chewbranca> I've got UEC running eucalyptus and a handful of running nodes, I can only ping those nodes from my master server, but not any other servers on my network, even though I can ssh into those boxes from various locations, what do I need to do to enable ping on those boxes?
[00:22] <nginxx> hi MTecknology
[00:33] <MTecknology> nginxx: hi
[00:37] <SpamapS> mathiaz: ping
[01:21] <hggdh> do any of you know of any backup pacakge for clouds (storage backups, I mean)?
[01:30] <zul> hggdh: bacula? ;)
[01:31] <hggdh> zul: perhaps could be used, yes, thank you. We should look at it.
[01:37] <chewbranca> anyone monitoring eucalyptus vms with zenoss?
[01:44] <hazmat> does cloud init install recommends?
[02:32] <X-Sleepy-X> hi
[02:32] <X-Sleepy-X> is it worth reporting abusive IP's to the ISP that holds the IP?
[02:33] <X-Sleepy-X> has anyone here done it?
[02:33] <X-Sleepy-X> and if so, did you get a respons?
[02:33]  * X-Sleepy-X is not sure if it's worth the hassle
[02:34] <Todd> My infosec friends say it's useless.
[02:34] <Todd> As a matter of fact I've heard of them logging into vulnerable dd-wrt routers and fixing bugs themselves so that attackers can't get back in and continue harassing their network.
[02:35] <Todd> But that's typically a bored Saturday night.
[02:37] <mac9416> Cloud noob question: if I were to set up a UEC cloud on a few of my older machines, could I install a GUI on the cloud and use it just like a desktop machine?
[02:45] <qman__> mac9416, in short, no
[02:45] <qman__> that's not the purpose of a cloud
[02:51] <mac9416> qman__, I see.
[02:52] <mac9416> I'm just looking for a purpose for my pile of old computers.
[02:52] <mac9416> Distributed computing is what I have in mind.
[02:52] <MTecknology> cluster into a compiling system
[02:53] <qman__> yeah, a beowulf style cluster would be more effective
[02:53] <qman__> cloud is aimed at high performance hardware
[02:53] <mac9416> Ah
[02:53] <qman__> say you have a dozen quad xeon servers, and you need to run 100 VMs
[02:53] <qman__> that's what a cloud is for
[02:53] <mac9416> I see.
[02:53] <mac9416> That's something I don't have.  :-)
[02:53] <mac9416> beowulf then. I'll look it up.
[02:54] <qman__> a beowulf cluster can spread threads across multiple physical machines
[02:54] <qman__> so if you have a CPU intensive application that supports multithreading, just turn the threads up and it'll share the load
[02:55] <mac9416> Nice.
[02:55] <mac9416> I guess one application I had in mind was Blender.
[02:55] <mac9416> It supports multiple threads for rendering.
[02:55] <qman__> of course, it has to be relatively big, otherwise the network performance will negate the advantage
[02:55] <mac9416> Right.
[02:57] <mac9416> Well, even if it does a half-decent job, I'll learn something in the process, so it won't all be for nothing.
[03:11] <orudie> how can I check which mail protocols are enabled with dovecot/postfix
[03:20] <dragondon> can someone help me manage my logs.  They keep recording dropped/aborted/limited events despite havig removed them and something keeps adding those rules back in....getting very frustrating to troubleshoot real software issues with I have 80,000 lines of nothing but those 3 iptables rules....
[03:42] <CppIsWeird> every time i try to mount anything i get mount: /dev/sr0: unknown device. I've never even attempted to mount /dev/sr0.
[05:32] <Hilikus> hey guys
[05:32] <Hilikus> how do i do a remote upgrade to 10.4? when i try do-release-upgrade it says it's not recommended to do remote upgrades
[05:33] <Hilikus> but i remember i used something else to go from 9.04 to 9.10
[05:39] <ScottK> Hilikus: do-release-upgrade is the best way.  It does offer you an alternate ssh connection in case of problems.  I recommend you check you can access it before you start the actual upgrade.
[05:49] <Hilikus> thanks
[06:00] <unewbie> can i install ubuntu 8.04 as guest os with kvm in my ubuntu 10.04 host?
[06:09] <twb> unewbie: I don't see why not.
[06:11] <unewbie> ok, i'll try
[06:14] <twb> I suppose you might get better *performance* with guests that have recent kernels and thus support virtio, but kvm *should* handle *any* guest OS, albeit slowly.
[06:51] <prestonc> what's the footprint like (memory wise) of server compared to desktop? Rough guestimate. I need to install a LAMP server and I'm deciding between loading my MacBook up with RAM (I bought a basic MacBook with 2GB) or just installing the server version under VMWare.
[06:56] <jmarsden> server is smaller by far, if you don't run any services :)  RUn a huge database there and you'll need plenty of RAM... what are you doing with it?
[06:59] <prestonc> I'm doing development work on an iPhone app against a RESTful server
[06:59] <prestonc> I have a staging server and live server in the cloud to do testing, etc. Load testing.
[07:00] <prestonc> However, I need to be able to do development of the app on my VMWare instance.
[07:01] <jmarsden> You can run server happily in 512MB with apache and a small MySQL db.  Probably in 256MB if you must.
[07:03] <prestonc> ok
[07:03] <prestonc> thanks
[07:03] <jmarsden> You're welcome.
[07:11] <twb> jmarsden: 128 MB is not sufficient to run the normal installer; it silently fails near the end and the result isn't bootable
[07:12] <twb> When installing I allocated 512 to my VM; afterwards I *think* was booting happily with 128MB.
[07:12] <twb> Oh, that was with 8.04 and without the "lowmem=N" boot options.
[07:13] <jmarsden> twb: Makes sense.  I only suggested going as low as 256MB, which should be OK for install, as well as running thereafter.
[07:23] <SpamapS> I have a lucid VM running w/ 192M .. but I installed w/ 384
[07:23] <SpamapS> should be able to go even lower depending on the services you want to run
[07:24] <prestonc> basically a LAMP server. Nothing fancy.
[07:24]  * twb grumbles
[07:25] <twb> "nothing elegant" is more like it.  Stupid mysql.  Stupid PHP.
[07:25] <prestonc> haha
[07:25] <prestonc> I agree
[07:25] <prestonc> I didn't pick the architecture
[07:26] <prestonc> If I were starting it from scratch it would probably be Python on AppEngine or somewhere where I could take advantage of something like BigTable
[07:50] <AndrewIII> Hi
[07:56] <twb> When generating a ramdisk inside a 10.04 chroot, I get a lot of grep: /proc/modules: No such file or directory
[07:56] <twb> In what way is the 10.04 chroot's initramfs-tools trying to be cleverer than the 8.04 chroot, and how do I tell it NOT to be clever?
[08:03] <_ruben> just mount /proc ?
[08:04] <twb> Except that I suspect it's doing stuff along the lines of MODULES=dep, which will generate a totally wrong ramdisk because the host I'm building on and the host I'm booting on are totally different.
[08:13] <_ruben> i always kinda assumed a ramdisk to be specific to a host its built on, so building for another host would only work if hardware would be identical/similar
[08:13] <_ruben> well .. an initrd/initramfs/etc
[08:14] <corpse> when using "get" in ftp, is there a needed option to download a full directory?
[08:15] <_ruben> depending on the client you could use get -R, or mget, or ...
[08:15] <corpse> sorry, ubuntu 10.04 using lftp
[08:21] <twb> _ruben: that wouldn't work very well for e.g. the ramdisks that d-i and the desktop CD user
[08:21] <twb> s/user/use/
[08:22] <twb> corpse: that's up to lftp.  The FTP protocol itself has no facility for recursive downloads.
[08:30] <jmarsden> corpse: I think you may be looking for the "mirror" command in lftp?
[08:31] <corpse> I am trying to download a directory that has multiple files in it from a seedbox to my fileserver though a ssh window
[08:31] <corpse> so instead of just moving one file at a time i would like to copy over the hole directory
[08:32] <jmarsden> So in lftp, type help mirror and use it to move the stuff you need around... right?
[08:34] <jmarsden> for just a "a bunch of files from one directory" you could use mget instead.  mirror will recurse if your directory to copy contains subdirectories you also want to copy...
[08:35] <corpse> jmarsden: When i tryed to use mget i would get a file access error. the mirror command seems to be working though. Thanks much
[08:35] <jmarsden> You're welcome.
[09:12] <trapmax> after upgrading our server, noticed this: http://pastebin.com/cF1TVjx1
[10:00] <kaushal> hi
[10:01] <kaushal> I accidentally deleted the sarg index.html file located under /var/www/squid-reports/ the index.html file says about Daily, Monthly and weekly, is there a way to regenerate it again ?
[10:02] <kaushal> sarg --> squid analysis report generator
[10:02] <kaushal> I am on ubuntu 9.04 server
[10:04] <kaushal> checking in again for my query ?
[10:14] <huats> morning
[10:22] <RoyK> http://cache.gawkerassets.com/assets/images/4/2010/03/for_traffic_cameras.jpg
[10:30] <soren> kaushal: Dude. Two minutes after posting your question is /not/ an appropriate time to pester the channel again.
[11:13] <remix_tj> Does anyone has installed ubuntu server on IBM xServer 225 ? I get kernel panic while installing...
[11:14] <twb> I installed 8.04 on some kind of IBM rackmount box
[11:15] <remix_tj> i'm talking of 10.04...
[11:15] <remix_tj> now i'm changing cd reader, maybe is broken
[11:15] <twb> Looking at the IBM product poster, I think it was an x3950 or so
[11:16] <remix_tj> ah, modern one :-)
[11:16] <twb> It was new at the time, yes
[11:16] <twb> It would've been about 3RU, not a blade server, with three SATA disks in a hardware raid5
[11:17] <twb> It worked, anyway, apart from $boss forgetting to buy the RAID5 ROM
[11:18] <twb> Does the panic indicate what failed?
[11:19] <Error404NotFound> are there any nice web management tools for iptables beside webmin and bitfrost?
[11:19] <twb> Error404NotFound: that suggests webmin is nice.  It isn't.
[11:20] <twb> ebox is what Ubuntu prefers for web management.  Personally I think it's a bloody waste of time and it cripples both the system and the sysadmin who learns it.
[11:20] <Error404NotFound> twb, correct :P
[11:20] <twb> For iptables I'd be using the iptables-persistent package and writing a simple, elegant ruleset.
[11:21] <twb> If you prefer to have a wrapper make a mess of things, ufw is simple and shorewall is powerful.
[11:21] <Error404NotFound> twb, shorewall is great...
[11:22] <Error404NotFound> twb, i think if i get some time i might write a basic PHP web ui with mysql support for managing shorewall...
[11:22] <twb> Shorewall adds complexity, and I'm not convinced that you get much in return.
[11:22] <twb> Error404NotFound: I expect someone has already done that, but then of course you'll be running PHP and MySQL...
[11:22] <Error404NotFound> twb, hmmm, i also expect, but unable to find one :P, there is a module for webmin and shorewall though
[11:23] <kaushal> soren: apologies
[11:28] <twb> I guess most cowboys simply use webmin
[12:17] <mco1> Hello. We ware currently experiencing problems with samba on our ubuntu 9.04 server. Access to the shares is very slow. How can I best analyze the problem and have their been recent reports on problems with samba?
[12:20] <twb> mco1: have you looked at the logs?
[12:20] <twb> Have you checked /proc/mdstat, SMART, dmesg, etc, for an underlying error?
[12:20] <_ruben> i'd start at the bottom and work yourself up .. check diskload ... network load
[12:21] <twb> Yeah, that too.
[12:23] <mco1>  /proc/mdstats reports no errors
[12:23] <mco1> The only error I can find in dmesg is "[   33.109609] svc: failed to register lockdv1 RPC service (errno 97)."
[12:23] <mco1> processor and memory load seems fine
[12:24] <mco1> But there seems to be a problem with disk load:
[12:24] <mco1> $ df -h
[12:24] <mco1> Dateisystem            Größe Benut  Verf Ben% Eingehängt auf
[12:24] <mco1> /dev/md0              277G  277G  266M 100% /
[12:24] <mco1> udev                  501M  264K  501M   1% /dev
[12:24] <mco1> none                  501M     0  501M   0% /dev/shm
[12:24] <mco1> none                  501M  568K  501M   1% /var/run
[12:24] <mco1> none                  501M     0  501M   0% /var/lock
[12:24] <mco1> none                  501M     0  501M   0% /lib/init/rw
[12:24] <mco1> none                  277G  277G  266M 100% /var/lib/ureadahead/debugfs
[12:24] <mco1> /dev/md1              466G  447G   20G  96% /daten/groups
[12:25] <_ruben> df doesnt show diskload, it shows diskusage
[12:25] <mco1> my bad, thanks.
[12:25] <_ruben> but full disks tend to be a bad sign anyway
[12:26] <mco1> I don't exactly know the purpose of debugfs?
[12:26] <_ruben> dont worry about that one, / is more important (and the same disk)
[12:27] <_ruben> as for diskload, use iostat (eg: iostat -m -x 3)
[12:30] <mco1> I will try to make some space on the disks, meanwhile I will read the manpage of iostat. :)
[12:33] <Felixb> hello everyone
[12:34] <Felixb> can someone help me? anyone here?
[12:35] <hggdh> !ask
[12:38] <Felixb> well, i have the ubuntu 10 server cd in my 2nd pc now, and want to install it. but it seems that ubuntu doesnt recognize my keyboard, i cant choose the language from the list. but in the "normal" ubuntu version and also in the bios the keyboard works.. anyone got an idea?
[12:48] <Felixb> ok found the solution... by enabling the usb keyboard support in the bios..
[13:18] <bogeyd6> truecrypt manage to lose a whole volume i had on a usb drive
[13:20] <e-DIO-t> does anyone have to work with Ubuntu hosts on ESXi4 Server and deal with vlans?!
[13:22] <tydeas> what firewall settings does the ufw must have so it can have dns server bind9?
[13:23] <e-DIO-t> tydeas, guess only needing is that it let pass 53UDP
[13:23] <twb> Perhaps "ufw allow dns"?
[13:23] <twb> Make that "ufw allow domain", per /etc/services
[13:23] <tydeas> Could not find a profile matching 'dns'
[13:24] <tydeas> what do you mean per /etc/services
[13:24] <jdstrand> yes, it can use /etc/services, port number[/protocol] or application profile (see ufw app list)
[13:24] <bogeyd6> twb, you mean make that "sudo ufw allow 53"
[13:24] <bogeyd6> UDP/TCP will be open on 53 with ufw in that configuration allowing full DNS
[13:25] <bogeyd6> e-DIO-t, i used to have to do it but then we did away with VLANS and went to true switching environment
[13:25] <jdstrand> any of these would work:
[13:25] <jdstrand> sudo ufw allow 53
[13:25] <jdstrand> sudo ufw allow domain
[13:25] <jdstrand> sudo ufw allow Bind9
[13:25] <bogeyd6> check that Bind9
[13:25] <mco1> It seems like our samba problem was mainly caused due to our mac clients. Thanks to _ruben and twb for helping me find the problem faster.
[13:25] <tydeas> There is an app Bind9 :)
[13:26] <e-DIO-t> sigh bodgey: my boss won't buy other switches guess :P
[13:26] <e-DIO-t> [the fact is that tha idiot bought 2 48port switchs :°°°°]
[13:26] <bogeyd6> sounds like you dont even need VLAN
[13:27] <bogeyd6> but anyways
[13:27] <bogeyd6> whats the prob e-DIO-t
[13:27] <e-DIO-t> well: i'll "show" you all the way
[13:27] <e-DIO-t> 1 - installed vlan package on the ubuntu server on ESXi
[13:28] <e-DIO-t> 2 - configured a iface eth0.10 [raw_device etcetc] on network 10.0.0.0/24
[13:28] <e-DIO-t> 3 - built up a portgroup on Vswitch to Truck all vlan [vlanID 4095]
[13:29] <e-DIO-t> 4 - set as "tagged member" of VLAN10 the output port of the Vswitch
[13:29] <e-DIO-t> 5 - setup a eth0.10 on my own laptop, and setup my own switchport as tagged member of vlan10
[13:30] <e-DIO-t> ...no ping :°°° but the same configuration with 2 phisical hosts [i mean: no vswitch in between] works
[13:30] <e-DIO-t> [ps: ping fails, arping too]
[13:30] <bogeyd6> i assume you setup up the actual switch with VLAN
[13:31] <bogeyd6> e-DIO-t, also we will need a pastebin.ubuntu.com of your "sudo ifconfig"
[13:32] <bogeyd6> sorry
[13:32] <bogeyd6> e-DIO-t, also we will need a pastebin.ubuntu.com of your "sudo ifconfig -a"
[13:32] <e-DIO-t> well, i'm cisco "born" so those 3com are a bit "strange" to me. Anyway since the same config with 2 ubuntu phisical hosts works, guess the phisical switch config it's ok, and even the ubuntu one
[13:32] <e-DIO-t> ...ok, i'm pasting..would you like a /etc/network/interfaces too?
[13:33] <e-DIO-t>  http://pastebin.ubuntu.com/450075/ <== heres Ubuntu server "in" EsxI" conf
[13:34] <bogeyd6> e-DIO-t, is the 8021q module enabled?
[13:35] <e-DIO-t> not "staticly": but i modprobed it during proofs
[13:35] <bogeyd6> k
[13:35] <bogeyd6> have you rebooted since
[13:36] <e-DIO-t> guess...i should have not.
[13:36] <bogeyd6> this will load 8021q on boot sudo su -c 'echo "8021q" >> /etc/modules'
[13:36] <bogeyd6> then we need a pastebin of your /etc/network/interfaces
[13:37] <e-DIO-t> isical switch config it's ok, and even the ubuntu one
 e-DIO-t, also we will need a pasteù
[13:37] <e-DIO-t> damn...ghgh wait
[13:38] <e-DIO-t> here's the server one
[13:38] <e-DIO-t> http://pastebin.ubuntu.com/450077/
[13:39] <bogeyd6> well there is your problem
[13:39] <bogeyd6> configured a iface eth0.10 [raw_device etcetc] on network 10.0.0.0/24
[13:40] <e-DIO-t> mmmh
[13:40] <bogeyd6> but you clearly have it set as eth1.10
[13:40] <e-DIO-t> ...
[13:40] <e-DIO-t> ..well..wait
[13:40] <bogeyd6> this brings serious doubt of credibility
[13:40] <nodeadmin> hi, I have eucalyptus (UEC) running on two machines (cc+clc+walrus+sc) and (nc). I can ping created virtual machines and they receive ips from my real DNS in system mode. The problem is that I cannot ssh into them, even though they are running a regular Ubuntu Lucid image. Any ideas?
[13:40] <bogeyd6> nodeadmin, is the firewall enabled?
[13:41] <e-DIO-t> no wait
[13:41] <e-DIO-t> i can't follow u:
[13:41] <nodeadmin> ufw says inactive on the controller and node.
[13:41] <e-DIO-t> i got: eth0 as "native", eth1 as auto, and eth1.10 on /etc/network/interface
[13:42] <bogeyd6> e-DIO-t, well the pastebins look fine and with your apparent knowledge of the system it looks like everything but your switch is setup. You do realize you have to configure the switch for that VLAN and then it should start working just fine.
[13:42] <bogeyd6> nodeadmin, that was my best guess
[13:43] <e-DIO-t> mmh...damn: but it won't go :°°°°
[13:43] <nodeadmin> lol, thanks
[13:43] <bogeyd6> e-DIO-t, https://wiki.ubuntu.com/vlan
[13:44] <e-DIO-t> well boogey :P right, problem's not the ubuntu one
[13:44] <nodeadmin> its kind of funny, either I can ssh into them from only the controller, and they can access the outside world, but nobody can access them. Or, in system mode, everybody can ping them but not connect, not even with hybridfox.
[13:44] <e-DIO-t> guess there's something between switch and vswitch
[13:45] <tydeas> Is there any gui for bind9"?
[13:45] <bogeyd6> e-DIO-t, which model of 3com switch do you have?
[13:45] <e-DIO-t> Baseline Switch 2948-SFP Plus
[13:45] <bogeyd6> tydeas, not at the moment but here is the best docu i have ever seen on bind9 setup in ubuntu, http://www.ubuntugeek.com/dns-server-setup-using-bind-in-ubuntu.html
[13:45] <e-DIO-t> and i'm hating them: they got also some "strange issue" on MAC/IP based acl
[13:45] <e-DIO-t> tydeas: for sure it's called "gnome terminal" ;)
[13:46] <bogeyd6> !dns | tydeas
[13:46] <tydeas> e-DIO-t: i love it.
[13:46] <e-DIO-t> ^_^
[13:46] <tydeas> my boss asked for a gui.
[13:46] <e-DIO-t> ahaha mine too
[13:46] <tydeas> using irssi
[13:47] <e-DIO-t> is still asking for gnome-desktop over a group of servers ;)
[13:47] <tydeas> to talk here ;)
[13:47] <e-DIO-t> damned windows-grown geehehe
[13:47] <tydeas> xoxox
[13:47] <tydeas> even starting using w3m when want to read documentations such as how to set bind9
[13:47] <sommer> morning
[13:48] <tydeas> i am a fedora user though
[13:48] <bogeyd6> tydeas, http://www.debianadmin.com/bind-dns-server-web-interfacefrontend-or-gui-tools.html
[13:48] <tydeas> using ubuntu-server
[13:48] <bogeyd6> e-DIO-t, http://support.3com.com/infodeli/tools/switches/baseline/3Com_Baseline-Switch-2948-SFP-Plus_User-Guide.pdf  suggest 802.1q is not even supported
[13:48] <e-DIO-t> NOOOOOOOOOOOOOOOOOOOOOOOOOO
[13:48] <e-DIO-t> THAT WAS MY NIGHTMARE
[13:49] <e-DIO-t> :°°°°
[13:49] <e-DIO-t> ....
[13:49] <tydeas> what?
[13:49] <e-DIO-t> i'm going to kill my boss :°°°
[13:49] <e-DIO-t> ok...i'm going to smoke and seriously offende some random-god. BRB :(
[13:50] <bogeyd6> lol
[13:50] <bogeyd6> read the user guide first
[13:50] <tydeas> how can i check if the webmin is in the repo of ubuntu?
[13:51] <bogeyd6> i did a cntrl+f for 802.1q and couldnt find it
[13:51] <bogeyd6> tydeas, sudo apt-cache search webmin
[13:51] <Pici> !webmin | tydeas
[13:51] <bogeyd6> even if it isnt, there is a .deb from webmin themselves
[13:52] <tydeas> any one form the ones bogeyd6 posted my before that are supported?
[13:52] <bogeyd6> !ebox tydeas
[13:52] <bogeyd6> lol
[13:52] <bogeyd6> !ebox | tydeas
[13:52] <tydeas> is it for BIND9 config too?
[13:53] <bogeyd6> the whole shebang a bang
[13:53] <tydeas> did not understood
[13:54] <bogeyd6> tydeas, http://packages.ubuntu.com/intrepid/ebox-dns
[13:54] <bogeyd6> depending on server version of course
[13:55] <tydeas> used apt-get to take it :)
[13:59] <tydeas> what's all about dynamic dns?
[13:59] <tydeas> i have static ip?
[13:59] <tydeas> do i need something from this?
[13:59] <bogeyd6> e-DIO-t, IEEE 802.1Q defines an architecture for virtual bridged LANs, the Tagging
[13:59] <bogeyd6> Compliance services provided in VLANs, and the protocols and algorithms
[13:59] <bogeyd6> involved in the provision of these services. An important requirement
[13:59] <bogeyd6> included in this standard is the ability to mark frames with a desired Class of
[13:59] <bogeyd6> Service (CoS) tag value.
[13:59] <bogeyd6> so maybe it is considered a feature of the switch
[14:00] <bogeyd6> !ebox | tydeas
[14:00] <tydeas> what this suppose to mean?
[14:06] <bogeyd6> !english | tydeas
[14:16] <hggdh> ttx: the server meeting conflicts with Brazil's game today :-(
[14:17] <ttx> hggdh: use your right brain to cheer and your left brain to type ?
[14:17] <hggdh> ttx: I guess so :-)
[14:17] <hggdh> ttx: when does France play?
[14:18] <ttx> France is still on track ?
[14:18] <ttx> j/k
[14:19] <EtienneG> Guys, any way we can actually turn off / disable the Image Store entirely in UEC?
[14:19] <ttx> hggdh: Thursday 1830 UTC
[14:19] <EtienneG> I am thinking that it would be best to remove the tab entirely from the web ui where there is no Internet connectrivity
[14:19] <hggdh> ttx: thanks. Should be a nice game to watch
[14:20] <Ryan1> I'm setting up key based auth for SSH. Am I able to copy my public key to any other machine I use (eg. by USB stick) or should I generate a new key for each machine?
[14:20] <EtienneG> Ryan1, check ssh-copy-id
[14:21] <EtienneG> copying your public key to other machine is basically what it does
[14:21] <Ryan1> My understanding is that it copies it to the SSH server and adds it to the authorized_keys file
[14:21] <EtienneG> Ryan1, just fyi, id == key pair, in ssh parlance
[14:21] <EtienneG> Ryan1, yep
[14:22] <Ryan1> Basically, my home desktop is my SSH server
[14:22] <Ryan1> And I'm using my laptop at the moment, which I've set up correctly
[14:22] <Ryan1> And used ssh-copy-id to move the public key to the desktop/server
[14:22] <Ryan1> But I also SSH from work
[14:23] <Ryan1> So I need to know if I can just put my public key on a USB drive and take it to work
[14:23] <Ryan1> Or whether I need to generate a new key from my work PC
[14:23] <EtienneG> Ryan1, ok then.  You will need to carry your private key with you, to be able to ssh into the machine where you have deployed your public key in authorized_keys
[14:23] <EtienneG> Ryan1, *or* generating a second key and adding the public part to the authorized_keys file is apossibility
[14:23] <Ryan1> ok, thanks
[14:24] <EtienneG> both would work; I prefer single key
[14:25] <tydeas> i can not understand how to start ebox yet.
[14:27] <hggdh> EtienneG: IIRC, it is all javascript
[14:27] <e-DIO-t> bogeyd6, so in the end: those 3com are vlan useless :°°°
[14:27] <EtienneG> hggdh, erm, ok.  Any easy way to tweak that?
[14:28] <hggdh> EtienneG: and, unless you drive a connection to (say) the Image Store, you will not know if it is reachable or not
[14:28] <hggdh> EtienneG: no, I do not know of any easy way
[14:28] <EtienneG> hggdh, too bad, I will forget it then
[14:28]  * hggdh goes have a look
[14:31] <bogeyd6> e-DIO-t, i dunno
[14:32] <bogeyd6> e-DIO-t, read the user guide and see if i missed something
[14:32] <bogeyd6> bbiab, gonna do my backports update and see if this fixes my laptop problems
[14:40] <e-DIO-t> "4.2 802.1x Auto-VLAN Assignment This release allows VLANS to be assigned via the Radius server. Note that this feature if enabled can assign untagged VLANs to ports where users are already connected. If they
[14:40] <e-DIO-t> are assigned to other untagged VLANs then they will be removed from the port."
[14:40] <e-DIO-t> ...so guess it should work with 802.1q trunks :°°°
[14:41] <e-DIO-t> and i'm in the same ole deep sh!t :°°°
[15:02] <_ruben> e-DIO-t: why do you need vlan support within your vm? you need more than 4?
[15:05] <e-DIO-t> nope _ruben: i'd be glad to separate broadcast domains even within the same phisical blade
[15:07] <e-DIO-t> anyway the fact is: why shouldn't work since vswitch supports 802.1q trunks, phisical switch supports 802.1q trunks and linux supports 802.1q trunks?!
[15:07] <e-DIO-t> that's sending me nuts :°°°
[15:07] <_ruben> it works for me as well .. but gotta run now
[15:08] <e-DIO-t> :°°°°
[15:08] <e-DIO-t> ok...hope to  got you any other day. to get a "chat" bout this
[15:08] <Pici> e-DIO-t: This conversation is a bit over my head, but perhaps you'd find better support asking in ##networking ? If not, feel free to ignore me ;)
[15:09] <e-DIO-t> ops: pici neither tought it could exist :P
[15:09] <e-DIO-t> [and neither searched for it... i'm going to take a look]
[15:13] <e-DIO-t> mmmh
[15:19] <smoser> hggdh, ping
[15:20] <hggdh> smoser: pong
[15:21] <smoser> eucalyptus bug... mail from them suggesting fix
[15:21] <smoser> i'll do a build if you'd like and get you some binaries if you can then attempt a fix
[15:21] <hggdh> yes, small fix, two modules. I tried patching it, but don't know how :-(
[15:21] <hggdh> smoser: yes, thank you very much
[15:54] <kirkland> hggdh: ping
[15:55] <kirkland> smoser: hggdh: actually, i just sent a build to ~ubuntu-virt ppa
[15:55] <kirkland> smoser: hggdh: get that tested whenever you can
[15:55] <kirkland> hggdh: maybe you can start prepping the rig for that test
[15:55] <smoser> kirkland, cool. thanks.
[15:55] <smoser> my build failed, don't know why.
[15:58] <smoser> i basically did the same thing you did
[15:59] <smoser> ah. i know what i did differently, i tried to build on maverick
[16:01] <hggdh> kirkland: the rig is ready. As soon as your build is done, I will install and run tests on it
[16:10] <smoser> kirkland, so, just an fyi, lucid-proposed eucalyptus fails to build from source on maverick.
[16:10] <smoser> probably a "oh well", but just fyi
[16:11] <kirkland> smoser: yeah, i think Daviey is on top of that one ...  something about libgroovy, i think
[16:13] <Daviey> smoser: yep.. that is fixed in a pending upload
[16:17] <Daviey> kirkland: Are you able to upload that?  (Note, i've pushed to the branch today)
[16:17] <kirkland> Daviey: the 1.6.2 maverick?
[16:17] <pmatulis> in top i see the process 'apache2'.  is there any way to get more info than that such as what virtualhost this process is serving?
[16:17] <kirkland> Daviey: i'll sponsor that now
[16:18] <kirkland> Daviey: wait, i couldn't build a source package
[16:18] <kirkland> Daviey: where is the orig tarball?
[16:19] <Daviey> kirkland: from upstream branch
[16:19] <Daviey> kirkland: wait 1, one more commit, just noticed something
[16:19] <kirkland> Daviey: okay, i can mumble now
[16:30] <musictoto> hi all! i'm using ubuntu 10.04 LTS server edition and i need an option like the 'lock screen' option you have in the desktop edition, without logging out. is that possible ? and how ?
[16:33] <smoser> musictoto, you're talking about the console ? ie, you've logged in and you want to lock it ?
[16:34] <smoser> if thats the case, I'd suggest using gnu screen (or, better, byobu) and then hitting 'ctrl-a x'
[16:39] <kirkland> smoser: musictoto: actually you can run "byobu" and then hit F12
[16:39] <ruben23> hi guys i installed mysql-server 5.1 and accidentally replaced my.cnf with my.cnf version 5, not its not starting, this is a fresh install of 5.1 are there ways i can download fresh default config of mysql5.1-
[16:40] <musictoto> smoser: kirkland: thanks guys
[16:53] <SpamapS> hah.. I just hit Cmd-Q instead of Cmd-Tab in my Chrome window.. about 25 tabs closed.. at first I was mad, but then I felt *FREE*
[16:53] <ruben23> hi guys how do i remove mysql-server 5.1  package
[16:54] <ruben23> completely
[16:54] <SpamapS> ruben23: apt-get remove mysql-server-5.1 should do it
[16:54] <webPragmatist> any of you know if it is possible to use pacemaker+heartbeat instead of corosync
[16:54] <webPragmatist> (openais)
[16:55] <ruben23>  SpamapS: when i do that it does conompletly uninstall- it just freeze..
[16:55] <ruben23> no activity at all
[16:55] <SpamapS> ruben23: what do you mean it just freeze?
[16:56] <SpamapS> ruben23: the uninstall process?
[16:56] <SpamapS> ruben23: maybe mysqld is stuck
[16:56] <ruben23>  SpamapS:yes..., what happen really is i accidentally replaced the my.cnf of 5.1 with 5.0 my.cnf version--after that mysql-server is not responding at all.
[16:58] <webPragmatist> ooh nm
[16:58] <webPragmatist> i just installed heartbeat and it installed pacemaker
[16:58] <webPragmatist> genius
[16:59] <ruben23> guys help, im stock
[17:12] <webPragmatist> anyone ever run "logfile"
[17:13] <SpamapS> logfile: command not found
[17:13] <ruben23> how to check if a package is completely remove
[17:15]  * JPP is away: Work
[17:15] <webPragmatist> yea it's like a logd maybe something old in this confi
[17:15] <webPragmatist> g
[17:19] <mathiaz> kirkland: Daviey: hallyn: bug 594179 - does it make sense?
[17:19]  * Daviey looks
[17:20] <Daviey> mathiaz: hmm.. the reporter hasn't stated where the patched derived from
[17:20] <Daviey> is it a cherry pick?
[17:21] <mathiaz> Daviey: hm - I think wrote it up]
[17:21] <mathiaz> Daviey: I'm more asking whether what he tries to do *actually* makes sense
[17:21] <mathiaz> Daviey: ie can you change a format of a snapshot?
[17:22] <mathiaz> Daviey: it seems to me that the use case he tries to solve is to be able to use qcow2 as the format of the snapshot for a raw file (for example)
[17:23] <Daviey> mathiaz: Ahh, i see.. I can see that being useful.. not sure it's SRU stuff :/
[17:23] <Daviey> if it's for maverick, would be nice to see in upstream and then we get it for free
[17:23] <mathiaz> Daviey: well - does qemu *actually* suport doing that
[17:23] <Daviey> mathiaz: not a feature set i would be comfortable us carrying, personally.
[17:23] <Daviey> mathiaz: pass, never done it
[17:24] <mathiaz> Daviey: I don't think you can just add the format option to the command line and off you go it works magically
[17:24] <Daviey> mathiaz: brb, on a call
[17:42] <kirkland> hggdh: did you notice that that eucalyptus ppa built?
[17:43] <hggdh> kirkland: yes, was on an interview call, but just downloaded them
[17:59] <jits1998> hi guys... we run a training institute and need a nfs + ldap authentication setup for ubuntu .. we have around 200 users .. please point me to a documentation i can follow as a newbie ... and what kind of hardware server will be needed .. Thanks!!
[18:00] <RoyK^> jits1998: the ubuntu server guide is quite usable
[18:00] <RoyK^> 200 users can be anything, really, it just depends what sort of traffic they make
[18:01] <jits1998> RoyK^: we run a development training institute .. so frequent updates will be there..
[18:01] <RoyK^> homedirs on nfs?
[18:01] <jits1998> RoyK^: we want to share home dir frmo nfs
[18:01] <RoyK^> or local homes?
[18:01] <RoyK^> that'll create some traffic
[18:01] <RoyK^> some as in quite a lot
[18:01] <jits1998> RoyK^: yeah i wd think .
[18:02] <RoyK^> are we talking about 200 concurrent users?
[18:02] <jits1998> RoyK^: 120 concurrent via nfs clients..others will likely be using mail and other ldap based stuff..
[18:03] <RoyK^> how much data space do you plan to support?
[18:04] <jits1998> RoyK^: data space as in hard disk quota ? .. or total hard disk size ?
[18:04] <RoyK^> also, is there a fat budget, or do you want things to be cheap?
[18:04] <RoyK^> total
[18:04] <jits1998> RoyK^: cheap only .. we are just starting up .. :-) .. we will be renting server ...
[18:05] <RoyK^> ok, first, how much space do you need?
[18:05] <RoyK^> 1TB? 10?
[18:05] <jits1998> RoyK^: may be max 1 Gb per user.. don't know if we would do much more than that per user ..
[18:05] <RoyK^> how much do they charge for the server rent?
[18:06] <RoyK^> would it be possible to get something cheap instead?
[18:06] <RoyK^> cheap or even used
[18:06] <jits1998> RoyK^: we have a policy to not invest in hardware . as yet......
[18:06] <RoyK^> doesn't seem too smart to me, but anyway, not my choice
[18:07] <jits1998> RoyK^: the rentals are planned in  a way to get profit after 2 years i think ..
[18:07] <RoyK^> so, how much per month?
[18:07] <RoyK^> btw, where are you located?
[18:07] <jits1998> RoyK^: I am in india ..
[18:07] <RoyK^> ok
[18:07] <jits1998> jits1998: the laptops are rented at around 50 USD per month i think ..
[18:08] <jits1998> RoyK^: sorry .. replied to myself :P
[18:08] <RoyK^> hehe
[18:09] <RoyK^> I just bought this test box at work for storage - 8GB RAM, 8x2TB disk and some SSDs for root and cache, and it cost me less than NOK 25k
[18:10] <RoyK^> or $3500
[18:10] <RoyK^> you can get decent hardware quite cheap these days
[18:10] <RoyK^> and you won't need that amount of storage
[18:10] <jits1998> we can look at that.. there are some tax issues i guess.. but not sure..
[18:11] <RoyK^> but still - what you will need is probably good random I/O speeds, meaning lots of drives
[18:11] <RoyK^> they don't have to be big for your current needs, but you will be wanting many
[18:12] <RoyK^> say a couple of small drives for the root mirror and 4+ drives for the data
[18:12] <hggdh> kirkland: just started a test run, 2,000 instances
[18:13] <kirkland> hggdh: ETA?
[18:13] <RoyK^> jits1998: placing the storage on a dedicated box will help things a bit, especially if you use something like Nexenta or OpenSolaris on that - that allows you to to dynamically grow the storage without downtime
[18:13] <hggdh> kirkland: 3 hours to finish, but I will be monitoring. We will have a very good guess in about 30 min
[18:13] <kirkland> hggdh: great
[18:13] <RoyK^> hggdh: what sort of test run is this?
[18:14] <hggdh> RoyK^: UEC runs, mix of all instance types, for both i386 and AMD64 images
[18:16] <RoyK^> hggdh: any idea about this one ? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/579276
[18:20]  * ccheney thought he would have rest and just deal with the baby while on leave but has a todo list a mile long :-\
[18:20] <RoyK^> jiboumans: ?
[18:21] <jiboumans> RoyK: yes?
[18:22] <hggdh> ccheney: hey welcome ;-)
[18:22] <RoyK^> jiboumans: given a ballpark figure of your budget, I could recommend something, but recommending something without any idea of what you may spend on the setup is quite impossible
[18:23] <jiboumans> RoyK: sorry, i'm lacking context i think
[18:23] <RoyK^> erm - bloody autoexpand
[18:23] <jits> RoyK^: sorry i got d/c after you said i wd need lots of drives with good i/a
[18:23] <jits> i/o*
[18:23] <RoyK^> :)
[18:23] <RoyK^> jiboumans: sorry
[18:24] <jiboumans> RoyK: clearly i need to have everyone with an irc nick beginning with 'ji' banned ;)
[18:24]  * zul is getting high off the new asphalt in his driveway
[18:24] <jiboumans> zul: asphalt; the new glue?
[18:24] <RoyK^> jits: for 120-200 clients with nfs-mounted homedirs, you'll need good I/O. processing power is not likely to be an issue
[18:24] <jits> RoyK^: okay.. how do i get that :-s
[18:24] <zul> jiboumans: i dunno...all i know is that its not sitting well with me
[18:25] <RoyK^> jits: lots of drives
[18:25] <RoyK^> jits: not necessarily big ones, but more drives
[18:26] <jits> RoyK^: umm.. as raid or something ?
[18:26] <jits> RoyK^: i have no clue :-(
[18:26] <RoyK^> jits: current 2TB drives are very fast linearly, but all 7200rpm drives do around 120 IOPS, meaning not a lot unless you do LARGE reads, which you won't
[18:26] <jits> RoyK^: yeah it will be very small read and writes .. but frequently..
[18:27] <RoyK^> jits: yes, RAID-5 or RAID-6. I'd recommend using an external box with nexenta/opensolaris for the storage, since that's very nice, and can use SSDs for caching, can grow existing volumes without downtime and do blocklevel checksumming of everything - you won't get that on Linux until btrfs stabilises and catches up with zfs, which is likely to take a year or more
[18:28] <jits> RoyK^: that sounds very very expensive :|
[18:29] <RoyK^> why?
[18:29] <RoyK^> you can use cheap, even used hardware and make it run very fast
[18:29] <vraa> opensolaris/nexenta is free, your bulk costs are harddrives
[18:30] <jits> opensolaris can run on non sun systems ?
[18:30] <RoyK^> yes
[18:30] <RoyK^> with zfs (the filesystem in nexenta/opensolaris) you can also replace smaller drives in a RAID, one by one, and when you're done, the zpool has grown
[18:31] <RoyK^> I don't think linux has anything like that
[18:31] <hggdh> kirkland: does not look good...
[18:31] <jits> okay ..
[18:31] <RoyK^> jits: we have a dual 50TB setup built on supermicro hardware and cheap drives.....
[18:32] <kirkland> hggdh: alrighty
[18:32] <jits> RoyK^: 50 TB :P .. we will be rich when we need that kind of disk :P
[18:32] <vraa> RoyK you still have to resliver after each drive right?
[18:32] <RoyK^> vraa: yes
[18:33] <vraa> :) i love zfs
[18:33] <RoyK^> but if you have an extra slot, you just replace it, no degradation
[18:34] <RoyK^> jits: http://hub.opensolaris.org/bin/download/Community+Group+zfs/docs/zfslast.pdf
[18:34] <RoyK^> that's good reading about zfs
[18:34] <jits> RoyK^: thanks.. let me quickly go thru it and see :-)
[18:42] <FunnyLookinHat> Looking for some quick advice: what's the most reliable way to transfer large files between servers ?
[18:42] <FunnyLookinHat> (Not on the same network)
[18:42] <RoyK^> rsync?
[18:43] <FunnyLookinHat> Not mirrored directories....
[18:43] <FunnyLookinHat> One time pushes of large files
[18:43] <FunnyLookinHat> i.e File shows up in directory - cronjob sees it there and sends it off
[18:43] <FunnyLookinHat> I was thinking SCP ?
[18:44] <RoyK^> sure
[18:44] <RoyK^> rsync --partial can be better if the job is likely to be interrupted, though
[18:46] <RoyK^> FunnyLookinHat: just make sure the transfer script locks, so you don't get multiple transfers going in parallel :þ
[18:46] <sjm> FunnyLookinHat: you can use rsync over ssh, too.  look at the man page.
[18:47] <RoyK^> sjm: rsync runs over ssh by default these days
[18:47] <RoyK^> if you rsync somefile somehost:
[18:47] <sjm> RoyK^: Really?  When did that happen?
[18:47] <RoyK^> years ago :)
[18:47] <Delemas> On other distributions there are init script functions (/etc/rc.d/init.d/functions) which can be included so init scripts can easily find the pid of a forked child etc. Where can I find these on Ubuntu?
[18:47] <sjm> ok.  good to know.
[18:48] <RoyK^> Delemas: see /etc/init.d/skeleton
[18:49] <Delemas> RoyK^: excellent! Thanks :)
[18:50] <sjm> RoyK^: since version 2.6.0 (Jan 1, 2004) to answer my own question.  I now feel out of date. :)
[18:50] <RoyK^> hehe
[18:50] <Delemas> The initscript for bmc-watchdog is totally busticated so I'm trying to fix it...
[18:56] <RoyK^> jits: get the point?
[19:03] <FunnyLookinHat> sjm / RoyK^  - Cool thanks!
[19:04] <jits> RoyK^: yet to go thru it .. its 11:30 pm here :-) .. time to sleep.. will do it tom morning..
[19:04] <RoyK^> jits: ok, nite
[19:12] <webPragmatist> where's the documentation on different pacemaker resources that are supported
[19:13] <webPragmatist> i'm trying to figure out if i can make it switch a dns record
[19:23] <toxic> Hello, I'm willing to install ubuntu-server on a 4096MB flash-IDE drive. Is 4096MB enought to get all services for a homeserver up'n'runnin' ? (I intend to install zfs-fuse, lamp, mt-daapd, squid, webmin... the usual ;) )
[19:24] <webPragmatist> yes… probably 3gb would be okay
[19:24] <Pici> toxic: You may want to note that webmin is not in the repositories.
[19:24] <Pici> !webmin
[19:25] <toxic> !ebox
[19:25] <toxic> thanks Pici !
[19:25] <toxic> and Pici, what about the minimal requirement for HDD-usage of ubuntu-server ?
[19:26] <RoyK^> toxic: just make sure zfs gets a gig of ram if you're planning something useful
[19:26] <RoyK^> toxic: how's the performance of zfs-fuse, btw?
[19:27] <toxic> RoyK, RAM is not going to be a problem, but as I bought a small SilverStone  Sugo case, I need a very small HDD for the system if I want my 5 drives in my case ;)
[19:27] <RoyK^> ok
[19:28] <MTecknology> is it possible to do something like ssh user:pass@server-ip ?
[19:28] <RoyK^> why not nexenta if you need zfs?
[19:28] <Pici> toxic: The documentation claims 500mb for base, 1gb for all tasks, see https://help.ubuntu.com/10.04/serverguide/C/preparing-to-install.html
[19:28] <RoyK^> MTecknology: nope, but look up key auth
[19:28] <toxic> RoyK, right now I'm using ZFS on a very cheap Motherboard on the integrated SATA II Controller, and Achieve only 40-50 MiO/s reading from zfs and writing to zfs at the same time...
 hallyn: so on your stack of verify-kvm-feature-foo work items ...
 hallyn: let's start a document in a wiki page, or in google docs, that shows the command line (and any additional instructions) we're using to verify these
 hallyn: actually, let's take this to #ubuntu-server (no need to be secret here)
[19:29] <kirkland> hallyn: i just verified -vnc :1, and -serial stdio
[19:29] <kirkland> hallyn: but both of those have lots more values than ":1" and "stdio"
[19:29] <toxic> RoyK, I'm "pretty good" with ubuntu, but have yet no idea of any other linux distrib... And Nexenta is solaris-like I belive ?
[19:29] <kirkland> hallyn: i think it would be good to note what we've verified
[19:29] <MTecknology> RoyK^: I'm trying to make a script to automate configuring servers - one of the parts is setting up those shared keys :P
[19:29] <kirkland> hallyn: and I suspect we'll need to re-test all of these again, after your upload, and closer to release
[19:29] <RoyK> toxic: nexenta is opensolaris kernel with debian userspace
[19:30] <RoyK> toxic: same thing, really, for most use
[19:30] <RoyK> MTecknology: $HOME/.ssh/authorized_keys holds the keys
[19:30] <hallyn> kirkland: yeah, that's something i was trying ot ask last week.  especially once 12.4 goes to maverick *and*then* gets backported to lucid...
[19:30] <toxic> RoyK, would you advise me to swich to Nexenta for more stable zfs (still experimental with fuse, but had no issue with it in the last year...)
[19:30] <RoyK> toxic: yes, without doubt
[19:31] <kirkland> hallyn: let's do a wiki page for this
[19:31] <hallyn> kirkland: but i'm logging what i've verified under https://blueprints.launchpad.net/ubuntu/+spec/server-maverick-hypervisor
[19:31] <panfist> how could i find what user is running the apache service on my server?
[19:31] <hallyn> new wiki page woudl be good
[19:31] <hallyn> then we can be more precise/verbose
[19:31] <RoyK> toxic: tuning, debugging etc will be different, but zfs is zfs all way through
[19:31] <kirkland> hallyn: right, mark them DONE there --- but let's talk about what we've actually done in a wiki page
[19:31] <hallyn> ok.
[19:32] <toxic> RoyK, I'll try that out, I need to find time to test it on a VM, could you just answer this one question : as a ubuntu-geek, how long do you belive I need to adapt to nexenta and have a fully functional homeserver ?
[19:32] <kirkland> hallyn: how does https://wiki.ubuntu.com/KVMFeatureVerification sound to you?
[19:32] <RoyK> toxic: not very long
[19:32] <toxic> (I wanna know if I need 2 days or an hour ;) )
[19:32] <RoyK> toxic: just try to install nexenta on that wee box before filling it up
[19:33] <toxic> RoyK, I'll do that ! Thx for the advise !
[19:33] <toxic> Nexenta runs zfs in the kernel ? It should improve throughput :p no ?
[19:33] <hallyn> kirkland: sounds good.  i'll create that after mtg if you haven't already
[19:35] <hggdh> kirkland: I stopped the first run (about 400 instances) and will restart another 2,000 instances. Meanwhile I will look at the errors
[19:36] <RoyK> toxic: yes, nexenta runs zfs in kernel - that helps a wee bit :þ
[19:37] <toxic> Cool RoyK thanks ! I'm now asking on #nexenta to leave this chan for ubuntu ;)
[19:37] <RoyK> !webmin
[19:38] <kirkland> hggdh: so no silver bullet, eh?
[19:39] <hggdh> kirkland: does not look like it is silver, or even silver-coated
[19:42] <kirkland> hggdh: dammit
[19:43] <kirkland> hggdh: you can verify that you did install the new packages
[19:43] <kirkland> hggdh: restarted the various services?
[19:43] <kirkland> hggdh: dpkg -l | grep euca
[19:43] <kirkland> hggdh: i know, i'm sure you did, just checking ....
[19:44] <hggdh> kirkland: I did verify I installed all new packages -- dpkg -l euca\* uec\* before, then dpkg -i the ones needed on each box, then anotehr dpkg -l euca\* uec\* to verify all had the correct version string
[19:44] <hggdh> I wish I had missed some :-( then there would still be hope.
[19:45] <hggdh> kirkland: additionally, we are still running with VNET_ADDRSPERNET = 64, so we will not have spurious errors due to the 32-3 limit
[19:45] <kirkland> hggdh: okay
[19:46] <kirkland> hggdh: would you please reply to Chris' note on the cano-euca list with your findings?
[19:46] <kirkland> hggdh: we need him to get back to work on this
[19:47] <hggdh> kirkland: will do. I am now finding out what errors we had. Metadata related are certainly more than 10 so far
[19:49] <kirkland> hggdh: regression then?
[19:49] <kirkland> hggdh: what a mess ... :-/
[19:49] <kirkland> hallyn: https://wiki.ubuntu.com/VirtFeatureVerification
[19:49] <kirkland> hallyn: i put it under "Virt" feature verification, b/c some stuff is kvm, some is qemu, and some is libvirt
[19:50] <hggdh> kirkland: no, not really regression, just no improvement. Chris stated he could not repeat it...
[19:51] <kirkland> hggdh: okay;  please continue the dialog on that list
[19:56] <jits> RoyK: should i be looking at nexenta stor or nexenta :-s
[19:56] <RoyK> jits: nexenta core platform
[19:57] <jits> RoyK: okay thanks .. will download it ..
[19:57] <jits> RoyK: it works fine with ldap ?
[19:57] <RoyK> I'd guess so
[19:58] <RoyK> even Sun is replacing NIS with LDAP
[19:58] <jits> RoyK: cool ... thanks..
[19:58] <jits> RoyK: just could not sleep .. will put it to download :P
[19:58] <RoyK> we have opensolaris boxes as CIFS servers in an AD
[19:58] <RoyK> jits: where are you located?
[19:59] <jits> RoyK: India :)
[19:59] <RoyK> ah
[19:59] <jits> RoyK: its 12;30 now.. have to be in office at 9 :P
[19:59] <RoyK> still....
[19:59] <RoyK> I forgot
[19:59]  * RoyK is in .no
[20:01] <kirkland> hallyn: feel free to toy with the template or display of the info in that page
[20:01] <kirkland> hallyn: i just threw up the basics
[20:05] <hggdh> kirkland: this is different -- on the first run (430 instances) I had 39 failures. BUT this time there were no other failures but metadata
[20:05] <kirkland> hggdh: hrm, that's interesting
[20:06] <hggdh> kirkland: before that we would have a mix of metadata and SSH failures (probably due to bad IP addresses)
[20:06] <hggdh> kirkland: so, it seems there is a net gain
[20:07] <bogeyd6> page preview is broken in openoffice for x64 10.04
[20:08] <mathiaz> SpamapS: https://bugs.launchpad.net/~ubuntu-server/+packagebugs
[20:11] <jits> RoyK: got that...
[20:11] <jits> RoyK: still searching for a document to setup nexenta as nfs
[20:12] <RoyK> jits: that's zfs basics - zfs sharenfs
[20:12] <bogeyd6> !nfs jits
[20:12] <bogeyd6> !nfs | jits
[20:12] <bogeyd6> mebbe some guidelines?
[20:12] <RoyK> jits: /j #nexenta
[20:12] <RoyK> bogeyd6: not in nexenta
[20:13] <jits> RoyK: googling..
[20:14] <RoyK> jits: http://docs.sun.com/app/docs/doc/819-5461
[20:35] <kirkland> Daviey: so you're saying i should be pushing to: lp:ubuntu/maverick/eucalyptus ?
[20:35] <kirkland> Daviey: can you confirm that?
[20:36] <Daviey> one mo
[20:36] <Daviey> kirkland: lp:~ubuntu-core-dev/eucalyptus/ubuntu/
[20:36] <Daviey> lp:~ubuntu-core-dev/ubuntu/maverick/eucalyptus
[20:37] <cdubya> are there any good references on how to setup an odbc connection to ms sql on ubuntu server 10.04?
[20:37] <kirkland> Daviey: kirkland@x200:/tmp/euca$ bzr branch lp:~ubuntu-core-dev/ubuntu/maverick/eucalyptus
[20:37] <kirkland> bzr: ERROR: Invalid url supplied to transport: "lp:~ubuntu-core-dev/ubuntu/maverick/eucalyptus": ~ubuntu-core-dev/ubuntu/maverick/eucalyptus is too short to be a branch name. Try '~<owner>/+junk/<branch>', '~<owner>/<product>/<branch> or '~<owner>/<distribution>/<series>/<sourcepackage>/<branch>'.
[20:39] <Daviey> kirkland: bzr branch lp:~davewalker/ubuntu/maverick/eucalyptus/maverick_to_quilt/ ; bzr push lp:~ubuntu-core-dev/ubuntu/maverick/eucalyptus/devel
[20:41] <Guest24216> no matter what package i try to install, i get "Could't find package". But if I try to install an installed package I get "already the newest version" what am i doing wrong?
[20:42] <guntbert> Guest24216: what package are you trying to install? what command did you use?
[20:42] <Guest24216> and I can only find installed packages with apt-cache, I cant find anything else; example apt-cache search distcc
[20:43] <Guest24216> guntbert: anything example ... apt-cache install distcc
[20:43] <Guest24216> whoops apt-get install distcc
[20:43] <Guest24216> "couldnt fiond package
[20:44] <guntbert> Guest24216: did you enble the "universe" repo?
[20:44] <Guest24216> no... how do i do that?
[20:44] <cdubya> Guest24216, yeah, have you checked your sources?
[20:45] <Guest24216> yea.. source.list is full
[20:45] <Guest24216> of http: sites
[20:45] <cdubya> Guest24216, yes, but are the universe lines uncommented
[20:46] <guntbert> Guest24216: you must edit /etc/apt/sources.list and uncomment the relevant lines, then sudo apt-get update --- see https://help.ubuntu.com/10.04/serverguide/C/package-management.html
[20:46] <Guest24216> yes they are
[20:46] <Guest24216> all unicverse lines are uncommented... this is a fresh install
[20:47] <guntbert> Guest24216: if there is a # in front the line doesn't work, remove the #
[20:47] <Guest24216> yes ... i know comments
[20:47] <cdubya> Guest24216, and you've tried sudo apt-get update?
[20:48] <Guest24216> apt-get update "Cannot initiate the connection to 8080:80
[20:48] <Guest24216> whoops must be a proxy fault
[20:48] <Guest24216> damn
[20:50] <cdubya> !odbc
[20:51] <cdubya> !freetds
[20:52] <guntbert> !askthebot > cdubya
[21:03] <bogeyd6> dang neckbeards
[21:24] <a3ist> Someone want to help me find out a delay in ssh authentication?  It occurs after the client checks blacklist files.  Here's the client side verbose log: http://pastebin.com/aRQ8pqEy.   IPv6 is disabled/blacklisted
[21:28] <kirkland> hallyn: mathiaz: qemu-kvm uploaded to maverick
[21:28] <kirkland> hallyn: mathiaz: thanks a lot
[21:28] <kirkland> hallyn: how about that memory leak SRU for lucid?
[21:29] <hallyn> kirkland: https://code.launchpad.net/~serge-hallyn/ubuntu/lucid/qemu-kvm/memleak-fix2   proposed merge/review
[21:29] <hallyn> kirkland: mathiaz: thanks! (for mav upload)
[21:30]  * kirkland gets on that one
[21:32] <kirkland> hallyn: the mem leak one should be fix-released in maverick, now that uploaded 0.12.4, correct?
[21:32] <kirkland> hallyn: ie, the git commit that you cherrypicked was in the 0.12.4 release, correct?
[21:33] <hallyn> yup it's in there
[21:33] <smoser> hggdh, chris misunderstood your email
[21:33] <hallyn> (just checked 0.12.4 source)
[21:33] <smoser> please respond
[21:45] <ccheney> hggdh: hi :)
[21:45]  * ccheney just got back from taking his wife to the doctor
[21:46] <smoser> ccheney, all things good ?
[21:48] <hggdh> smoser, ccheney, what did I do?
[21:49] <smoser> hggdh, see msg
[21:49] <hggdh> of course, it is probably my fault ;-)
[21:53] <LowValueTarget> is installing from usb faster than installing from a dvd
[21:55] <kirkland> hallyn: fwiw, i subscribed to https://wiki.ubuntu.com/VirtFeatureVerification?action=subscribe
[21:58] <hallyn> kirkland: me too :)
[22:01] <kirkland> hallyn: oooh, -redir, neat
[22:01] <kirkland> hallyn: i just added that to testdrive's default kvm args :-)
[22:02] <kirkland> hallyn: also, for virtio network and disk, we should find some way of semi-consistently benchmarking this
[22:02] <kirkland> hallyn: such that we might be able to detect severe regresssions
[22:06] <hallyn> kirkland: jdstrand has the libvirt qa test script.  we shoudl be able to add some simple things to that maybe
[22:06] <kirkland> hallyn: definitely
[22:07] <kirkland> hallyn: would be nice to automate as many of these as possible
[22:07] <hallyn> we'll want to use a very fast, minimal os.  can you think of one?  :)
[22:10] <SpamapS> kirkland: btw thanks for the link to your UEC-on-a-key ... I won't get too ambitious w/ the demo.. just want to show it to people so they're aware that it exists and can be played with. :)
[22:10] <kirkland> SpamapS: sure thing
[22:10] <kirkland> SpamapS: thanks for reminding me about that
[22:10] <kirkland> SpamapS: did you catch my blog post on the subject?
[22:11] <SpamapS> kirkland: no I am eons behind in my RSS feed reading.
[22:11] <SpamapS> kirkland: but reading now. :)
[22:11] <kirkland> SpamapS: http://blog.dustinkirkland.com/2010/06/cloud-in-your-pocket-uec-liveiso.html is your guide to your demo ;-)
[22:11]  * SpamapS makes a steeple with his fingers
[22:12] <SpamapS> exxxxxccceelllent smithers
[22:13] <SpamapS> mmmmm dpkg source format 3.0 just saved my bee-hind.
[22:15] <SpamapS> mathiaz: you were asking how to get jars in the debian dir.. you just have to put '3.0 (quilt|native)' in debian/source/format and add all the binaries to debian/source/include-binaries  :)
[22:18] <webPragmatist> uhg
[22:18] <webPragmatist> why wouldn't shutdown -r now work?
[22:19] <SpamapS> webPragmatist: weird BIOS doesn't like rebooting?
[22:19] <webPragmatist> no clue >.<
[22:21]  * SpamapS decides he needs a burrito.. now
[22:22] <Delemas> cool I sent in a patch for freeipmi.
[22:30] <Phlogistique> Hi everyone
[22:30] <Phlogistique> I have a dedicated server I rebooted for a silly reason
[22:30] <Phlogistique> and now I can't get it to ping anymore
[22:31] <Phlogistique> my hosting has investigated
[22:31] <Phlogistique> and told me that the boot worked, it is possible to login locally, but
[22:31] <Phlogistique> there is a message "no network configured"
[22:32] <Phlogistique> they then booted a rescue system on the machine for me to investigate and fix
[22:32] <Phlogistique> but I didn't change the configuration and what's in the /etc/network/interfaces file looks perfectly sane
[22:33] <Phlogistique> any idea of where I can find more info/what I can try?
[22:36] <Phlogistique> http://pastebin.com/fMnfhWaB content of /etc/network/interfaces
[22:36] <lifeless> kirkland: still up ?
[22:36] <kirkland> lifeless: yop
[22:37] <zash> Phlogistique: try poking with ifconfig
[22:37] <lifeless> kirkland: http://www.tenshu.net/archives/2010/06/08/the-lawnmower-man/
[22:37] <lifeless> in there he says
[22:37] <Phlogistique> zash: I can't, my system is not running
[22:37] <lifeless> The best driver for disk/networking is the paravirtualised “virtio” driver. I found that ubuntu-vm-builder had already configured the networking to use this, but not the disk, so I modifie
[22:37] <Phlogistique> another system is running on the machine
[22:37] <lifeless> ...
[22:37] <lifeless> kirkland: is that a bug or feature?
[22:37] <kirkland> lifeless: page slow to load ...
[22:38] <Phlogistique> zash: I can look at files but ifconfig won't be interesting
[22:38] <lifeless> kirkland: in fact, there are a bunch of tunin things there that perhaps should be bugs
[22:39] <kirkland> lifeless: what part are you specifically talking about as a bug or featureA?
[22:39] <kirkland> lifeless: for one thing, i only ever use qcow2 as my backing disk image (never lvm as whomever writes there)
[22:39] <zash> Phlogistique: i meant run stuff like ifup and see what happens
[22:40] <kirkland> lifeless: is that you, or ng's blog?
[22:40] <lifeless> kirkland: ng's
[22:40] <lifeless> Ng: ^
[22:40] <lifeless> why I am funnelling between you two. bah.
[22:40] <kirkland> lifeless: heh
[22:40] <kirkland> Ng: howdy
[22:40] <lifeless> Ng: meet kirkland. kirkland: meet ng. TALK DAMMIT
[22:41] <kirkland> Ng: lifeless: if the problem is with vmbuilder, i need to redirect you to soren
[22:41] <lifeless> kirkland: it seems to me that when one of our sysops looks at the result of ubuntu-vm-builder and notes that changes should be made, we should at least consider whether different defaults would be good.
[22:41] <kirkland> lifeless: ack, and i agree
[22:41] <lifeless> kirkland: and you're my go-to guy in server space :)
[22:41] <kirkland> lifeless: sane defaults should equal "ubuntu"
[22:41]  * kirkland takes a bow
[22:42] <Phlogistique> zash: oh you're right this might be interesting - maybe http://pastebin.com/yR5BmnsY
[22:43] <kirkland> Ng: i'm happy to chat about this tomorrow, when we're both online again
[22:43] <Phlogistique> or maybe not...
[22:44] <zash> I have something similar to this problem: http://ubuntuforums.org/showthread.php?t=1507407
[22:45] <zash> after dist-upgrade from hardy. mysql won't start from init related utils, works-ish if launched directly like sudo -b mysql mysqld
[22:46] <Phlogistique> zash: well, my problem's not about mysql
[22:47] <zash> Phlogistique: but mine is :P
[22:47] <Phlogistique> oh
[22:47] <Phlogistique> I thought you meant "somebody has a problem similar to yours" haha
[22:48] <ccheney> smoser: sorry got called away again, yea she seems to be doing ok they gave her some medicine to take
[22:54] <FFForever> what is the easiest way to setup a software raid on an already installed system?
[22:55] <FFForever> (with two new drives)
[22:58] <JPP> FFForever, mdadm is the tool for linux software RAID
[22:58] <JPP> in ubuntu server, to install type in apt-get install mdadm
[22:58] <JPP> (as root or sudo)
[22:58] <FFForever> also what would be the best way to copy ~500gb to the new raid?
[22:59] <FFForever> cp -R /oldFolder/* /newFolder/?
[22:59] <JPP> Yes, probably :)
[22:59] <JPP> you'll need to format your new harddisks as linux raid first
[23:03] <zash> What's up with upstart and mysql?
[23:07] <zash> and why is the load climbing when the server is othervise idle
[23:19] <jdstrand> hallyn, kirkland, smoser: fyi. it is my eod now, but I have a preliminary libvirt 0.8.1-2ubuntu1 package for maverick
[23:19] <hggdh> hi ccheney, how are you & SO?
[23:20] <jdstrand> hallyn, kirkland, smoser: I plan to test it tomorrow and upload maybe late tomorrow or Thursday
[23:48] <kirkland> jdstrand: awesome
[23:49] <kirkland> jdstrand: i sponsored hallyn's qemu-kvm today
[23:49] <kirkland> jdstrand: to maverick