[00:22] <beeman_nl> hi guys
[00:23] <beeman_nl> i'm working on a silent ubuntu install with kickstart, but i can't seem to get the right package selection
[00:23] <beeman_nl> i tried  @base , @core etc at the %packages stanza but it still seems to install the 'ubuntu-standard' system
[00:23] <beeman_nl> which includes stuff like openoffice which i dont' need on a server :)
[00:30] <centaur5> beeman_nl, try this out: tasksel tasksel/first multiselect standard
[00:30] <beeman_nl> centaur5: oke, will try. and does that need to go below %packages, instead of the others? :)
[00:32] <centaur5> beeman_nl, it's an entirely different string. The one for packages is: d-i pkgsel/include string \ samba \ emacs21 \ etc...
[00:33] <beeman_nl> centaur5: maybe i should note that i'm using a kickstart file, not a preseed file :)
[00:34] <centaur5> beeman_nl, Oh sorry, I didn't notice that.  kickstart didn't cut my needs so I learned about preseed.  I don't know what to tell ya then.  :)
[00:34] <beeman_nl> centaur5: thanks anyway :) maybe i'll switch to preseed anyway if this gives me too much hassle :)
[00:35] <centaur5> beeman_nl, I think that solution is probably inevitable in this case.  :)
[00:41] <jcastro> beeman_nl: try a whitespace after the @
[00:41] <jcastro> "@ desktop"
[00:41] <jcastro> or whatever
[00:45] <beeman_nl> jcastro: thanks, i'll try that :)
[01:20] <beeman_nl> it seems like it's still installing the same packages, even when i use @ ubuntu-minimal
[01:20] <beeman_nl> take too long to test this kinda stuff ;)
[01:39] <twb> This is vm-builder?
[01:47] <beeman_nl> twb: no it's just a custom pxe setup
[01:48] <beeman_nl> atcually i now use an iso that fetches my kickstart file from the network and then installs the system according to that file :)
[01:48] <beeman_nl> but obviously the same thing for my pxe setup :)
[02:07] <VSpike> what's the best way to configure a pc-card ethernet adapter? Should I just add it to /etc/network/interfaces at eth0 and hope for the best?
[02:09] <Ganymede> hey, anyone here have installed ubuntu 9.10 on a VMWare ESX host using paravirtual NIC (vmxnet3) and paravirtual SCSI? the installer keeps choking on both, even though i have selected, "install a minimal virtual machine"
[02:10] <Ganymede> or should i install the OS completely and then install vmware-tools and then install those paravirtualized devices?
[02:14] <twb> beeman_nl: ah, you're using kickstart instead of preseeding.
[02:14] <twb> I was trying to work out what that "@" crap was
[02:15] <twb> Personally I prefer di-netboot-assistant and a hand-written preseed file.
[02:16] <Ganymede> it seems...ubuntu cannot boot from paravirtualized SCSI?
[02:16] <twb> Ganymede: it can with qemu
[02:17] <Ganymede> twb, but not VMWare ESX?
[02:17] <twb> I'm not licensed to run ESX, sorry.
[02:17] <Ganymede> okay, thanks
[02:17] <twb> Where does it fail?
[02:18] <twb> Can you pastebin the output?
[02:19] <Ganymede> well...it'd be hard since it's a console that boots into ... busybox?
[02:20] <Ganymede> let me try typing it out
[02:20] <twb> If you were using qemu, you could just put it in stdio mode and copy the text from your xterm :-)
[02:21] <Ganymede> gave up waiting for root device, /dev/disk-by-uuid/blah-blah does not exist, dropping to a shell
[02:21] <Ganymede> actually, let me try another VM with a newer version of vmware-tools, hold on
[02:21] <twb> vmware-tools won't make a damn difference
[02:21] <twb> It's just shitty kernel taint for hgfs and friends
[02:23] <Ganymede> "hgfs"?
[02:24] <twb> It's the filesystem that VMware uses to export a directory tree on the hypervisor, to guests
[02:24] <twb> As opposed to using e.g. NFS between the host and guests
[02:25] <twb> When you're in the busybox rescue shell, cat /proc/partitions.  Does it list anything?  If so, what?
[02:25] <Ganymede> yep, newer vmware-tools still fails to boot
[02:26] <Ganymede> /proc/partitions is empty
[02:26] <Ganymede> looks like i'll have to settle with non-paravirtalized SCSI
[02:26] <Ganymede> at least for the root disk
[02:27] <twb> Oh, sorry, you're talking about *para*virtualized SCSI
[02:27] <twb> You mean that you're giving the guest OS direct access to a LUN?
[02:31] <Ganymede> i actually don't touch the "LUNs" or whatever
[02:31] <twb> Ok, let me rephrase.
[02:31] <Ganymede> i just have a 6 TB "datastore" where i put the disk images and i attach them to a SCSI controller in the VM configuration
[02:31] <Ganymede> now that SCSI controller can either be a LSI Parallel SCSI or something...forgot what it's called
[02:31] <twb> OK, so you have a disk image -- a file on the host's filesystem -- and you're presenting this to the VM as a virtual SCSI disk?
[02:31] <Ganymede> or a paravirtualized SCSI "pvscsi"?
[02:31] <beeman_nl> twb: yeah, i'll try it with preseeding lateron :)
[02:32] <Ganymede> the disk image is on a RAID in the SAN, but i don't have to worry about that, i just have a 6 TB "datastore" where i put the disk images
[02:32] <twb> beeman_nl: IMO preseeding is way less fugly than kickstart.
[02:32] <Ganymede> and that 6 TB datastore is somewhere on the SAN
[02:32] <twb> beeman_nl: admittedly, I've never kickstarted Ubuntu, only RHEL
[02:33] <twb> Ganymede: OK, then you aren't *para*virtualizing the SCSI disk -- it's fully virtualizaed
[02:33] <twb> Ganymede: paravirtualizing it would be if you gave your guest OS direct access to (parts of) a real, physical SCSI controller, and the LUNs (disks) behind it.
[02:34] <twb> Ganymede: i.e. you have a paravirtualized guest OS with fully virtual SCSI disk.
[02:34] <Ganymede> twb, i thought paravirtualizated SCSI controller simply doesn't emulate all the intracacies of a real SCSI controller, i do have the option of using a physical SCSI controller or iSCSI or something else from my FC HBA, but paravirtualiztion is a different option
[02:35] <Ganymede> i will try to use a paravirtual SCSI controller for the swap disk and i'll get back to you
[02:35] <twb> Ganymede: OK, so when you emulate a full SCSI controller it works, but not if you set ESX to "pvscsi"?
[02:39] <Ganymede> twb, well, i've only been testing the boot disk so far, not an extra disk, maybe that will work with pvscsi
[02:40] <Ganymede> and while i'm here, i'm thinking about using noop elevator since my disk image is stored on god knows what...some monster RAID array over fibre channel?
[02:40] <twb> Ganymede: I imagine that pvscsi is a proprietary VMware-specific thing, and it'll only work for the root disk if you recompile your ramdisk to include the drivers from vmware-tools.
[02:40] <Ganymede> does that sound advisable?
[02:40] <twb> Ganymede: but that's just a guess -- checking /proc/partitions will help
[02:41] <Ganymede> twb, yeah, you're probably right about that, i'm not terribly familiar with how the boot process works
[02:41] <Ganymede> actually...yeah, i didn't think of that, you're absolutely right
[02:41] <twb> It's OK: I *am* familiar with how boot works :-)
[02:41] <Ganymede> i thought that since i had installed vmware-tools, the pvscsi module is suddenly visible to the boot process
[02:41] <Ganymede> but i see why that would be incorrect
[02:42] <twb> it's unlikely that a vmware-tools .deb is rolled correctly enough to rebuild the ramdisks
[02:42] <Ganymede> maybe in upcoming versions
[02:42] <twb> Do a "update-initramfs -u -k all" to forcibly rebuild them
[02:42] <twb> You can also try talking to#vmware about it
[02:43] <Ganymede> twb, but would that update-initramfs command know to include pvscsi in particular?
[02:44] <Ganymede> (i'm not even sure pvscsi is the module name, i'll have to check once i get this working)
[02:44] <twb> Ganymede: the default behaviour for initramfs-tools is to include ALL kernel modules in the ramdisk
[02:45] <twb> (Well, all disk, network, etc. modules.)
[02:46] <Ganymede> twb, thanks, i'll be sure to give that a try in the near future
[02:48] <twb> Ganymede: let me know if you succeed
[02:48] <Ganymede> twb, that's automatically done on kernel upgrade, right?
[02:48] <twb> Ganymede: in theory, yes.  In your case, I'd do it manually just to be sure
[02:49] <twb> update-initramfs is idempotent if you rerun it without changing anything, so it should be safe
[02:50] <Ganymede> so i just installed a bare minimum VM to test this out...got to vmware tool installation...it asks to build kernel modules...just realized i don't have gcc...can't install it because i don't have drivers for my paravirtualized NIC yet...back to the drawing board
[02:50] <Ganymede> ah, the hoops and hurdles i go through to get paravirtualization working...and will probably never notice the speedup
[03:00] <twb> One more reason to go with LXC and/or kvm...
[03:02] <JuanMarquez> can filter pages with user login control?
[03:02] <JuanMarquez> using squid
[03:04] <twb> JuanMarquez: I don't understand the question.
[03:06] <JuanMarquez> twb, ok i need filter webs pages certain users
[03:06] <JuanMarquez> sorry mi inglish is bad
[03:06] <JuanMarquez> jeje
[03:07] <twb> sabé.
[03:09] <twb> se?  I forget how to conjugate it.
[03:10] <JuanMarquez> ok
[03:12] <JuanMarquez> I need to filter web pages, but using the method by user, using proxy
[03:12] <JuanMarquez> is possible?
[03:12] <twb> Do users need to authenticate to use your proxy?
[03:13] <JuanMarquez> yes, in mode transparente I do not think can
[03:14] <twb> If users authenticate to squid, then you can
[03:14] <JuanMarquez> ah ok
[03:15] <JuanMarquez> sample config place?
[03:15] <JuanMarquez> pleace
[03:15] <lifeless> http://wiki.squid-cache.org/CategoryConfigExample
[03:33] <JuanMarquez> thnk
[03:35] <Ganymede> twb, looks like i got paravirtualized scsi (module IS called pvscsi) working on non-boot disk, and according to hdparm -Tt, the paravirtualized is like 4 times faster...but is hdparm a good benchmark?
[03:36] <Ganymede> with LSI scsi controller, i got 120 MB/sec read, with pvscsi, i got between 390 and 520 MB/sec
[03:37] <zroysch> is there any way of saving an md raid partition if you get this when mounting mount: wrong fs type, bad option, bad superblock on /dev/md2,
[03:38] <Ganymede> wait...something is fishy...520 MB/sec on a 4 Gb fiber? host must be caching... =/
[03:41] <twb> zroysch: pastebin /proc/mdstat's contents
[03:48] <zroysch> md2 : active raid5 sdd[1] sde[2]
[03:48] <zroysch>       1953524992 blocks level 5, 64k chunk, algorithm 2 [3/2] [_UU]
[03:48] <zroysch> sorry wrong spot
[03:59] <zroysch> twb: does that help
[04:00] <cef> anyone done any old old old version upgrades using old-archives.ubuntu ?
[04:00] <cef> mainly: does it work?
[04:31] <twb> zroysch: that's saying your array is still OK, it's just that you can't lose ANOTHER disk.
[04:31] <zroysch> twb: right
[04:32] <twb> So mount shouldn't have a problem
[04:32] <zroysch> so when i go to mount it gives me the wrong fs type, bad option, bad superblock on /dev/md2,
[04:32] <twb> That shouldn't happen.
[04:32] <twb> file -s /dev/md2
[04:33] <zroysch>  /dev/md2: Linux rev 1.0 ext3 filesystem data, UUID=3b309b65-9d8c-40d8-9373-f74190958845 (needs journal recovery) (large files)
[04:33] <Roxyhart0> hi, there any way to detect who in my NAT is gettong virus (we are blacklisted). I got clamd but is doesn't detect which PC is.
[04:33] <twb> And what does dmesg say after a mount attempt?
[04:34] <zroysch> twb: http://pastebin.ca/1805398
[04:34] <zroysch> Roxyhart0: a virus scanner?
[04:35] <Roxyhart0> i need in my nat because i cant control the clients
[04:35] <Jordan_U> Roxyhart0, How do you know that someone in your NAT has a virus in the first place?
[04:35] <twb> zroysch: e2fsck -p /dev/md2
[04:36] <Roxyhart0> because the DNSs send to me a report about we are getting blacklisted because virus
[04:36] <Roxyhart0> reports
[04:37] <zroysch> twb: http://pastebin.ca/1805403
[04:37] <zroysch> its stopped there.
[04:37] <Roxyhart0> my nat server is email server and webserver as well (linux)
[04:38] <twb> OK, so somehow your journal got corrupted, and as a result it has been emptied out
[04:38] <twb> Now it should mount again
[04:39] <twb> Obviously you should also be replacing the dead node in your array at some point
[04:39] <zroysch> twb: http://pastebin.ca/1805405
[04:40] <twb> Is this an ext4 filesystem?
[04:40] <zroysch> was ext3 as far as i knew
[04:40] <twb> Well, the disk thinks it was ext4
[04:40] <zroysch> beautiful
[04:41] <twb> What mount command are you issuing?
[04:41] <zroysch> oh wow
[04:41] <zroysch> it mounts now
[04:43] <twb> Now unmount it and e2fsck again
[04:43] <zroysch> what
[04:43] <zroysch> dont i want to backup the data asap
[04:44] <twb> If you were concerned about making a backup, you shouldn't have tried to mount it
[04:44] <twb> You should have immediately taking a bitwise copy with dd.
[04:44] <zroysch> true
[04:44] <zroysch> im pretty sure ive had fsck lose data before
[04:45] <twb> fsck did just lose data
[04:45] <twb> it deleted your broken journal
[04:46] <zroysch> not sure what a journal is
[04:46] <zroysch> i cant umount now
[04:46] <twb> It's the thing that remembers what was going to happen when you kicked out the UPS
[04:46] <zroysch> saying its busy
[04:51] <Roxyhart0> hi i find one solution but i dont know how to solve it. I need to stop any package with port 25 which is not really an email. somebody have any idea how to do that?
[05:11] <AdamDV> 95% of this channel is joins and quits I tell ya.
[05:12] <twb> AdamDV: so drop those
[05:13] <AdamDV> I dont know how to in pidgin :/
[05:13] <Roxyhart0> AdamDv are you administrator for this forum?
[05:13] <twb> Then use Emacs
[05:14] <AdamDV> Roxyhart0: ?
[05:14] <Roxyhart0> yes?
[05:14] <AdamDV> Am I the administrator for what forum?
[05:14] <AdamDV> THis channel?
[05:14] <AdamDV> Oh god no.
[05:14] <Roxyhart0> this channel i mean
[05:14] <AdamDV> I wish, haha.
[05:14] <Roxyhart0> :)
[05:24] <zroysch> umount: /mnt/raid5: device is busy.
[05:24] <zroysch> lsof and fuser not showing me anything
[05:24] <twb> Were you using chroot?
[05:24] <[1]spike> Hi Everyone, Is anyone here knowledgable with setting a proxy server with ubuntu server? i tried http_proxy=192.168.1.2:6588 then export http_proxy but i still cant seem to get any internet
[05:24] <zroysch> twb: usingsudo
[05:25] <twb> [1]spike: what makes you think your http proxy is binding to 6588 instead of webcache (8080)?
[05:26] <[1]spike> because a file server is bound to 8080 so the proxy is set to 6588
[05:26] <twb> [1]spike: does ss (or netstat) agree with you?
[05:28] <[1]spike> umm huh? im real new to linux, i don't know what those are so ill google it and let you know in a sec :)
[05:30] <[1]spike> ok so i take it i need to run netstat on the proxy server machine?
[05:36] <[1]spike> ok it specifies listning TCP 6588 and TCP 1080
[05:36] <[1]spike> so ill try setting as 1080 and seeing what happens
[05:36] <[1]spike> but im guesing the file server would still show up as listning
[05:36] <twb> It says *squid* is listening on both those ports?
[05:37] <twb> Oh, 1080 is SOCKS
[05:37] <[1]spike> It has no process names
[05:37] <twb> You want -p
[05:37] <[1]spike> ok ill do that now thanks for all the help :)
[05:37] <twb> In any case, you should now try to talk to the service using netcat
[05:37] <cef> Roxyhart0: you could use something like snort on the NAT gateway to watch the traffic. it can actively check for bad stuff (from rulesets) in passing traffic
[05:38] <twb> First try on the proxy server itself, then try on your test machine
[05:38] <zroysch> is there a quick way to make a drive unbusy when it is too busy to unmount? or should i just reboot
[05:38] <cef> Roxyhart0: but you may also find simply monitoring all port 25 traffic with tcpdump will get you somewhere.
[05:39] <[1]spike> ok im going to google netcat then get back to you.  Thanks alot
[05:39] <twb> zroysch: that depends on why it's busy
[05:39] <zroysch> twb lsof and fuser dont seem to tell me anything
[05:41] <twb> zroysch: did you use chroot?
[05:42] <cef> Roxyhart0: lastly, if you dump with tcpdump to a file, you can load the packets in wireshark (gui, win/linux/mac versions out there) on another PC than the NAT box. That might make it easier to see what is going on
[05:42] <[1]spike> Hey i think i can bypass this step.  everyother pc on the network works fine using proxy 192.168.1.2:6588 the Ubuntu Server can ping 192.168.1.2 but im attemting "Ping Http://www.google.co.nz/" which tells me
[05:42] <[1]spike> "Unknowen Host"
[05:42] <twb> [1]spike: ping uses ICMP, not HTTP.
[05:43] <cef> drop the 'http://' and '/' from your ping and try again
[05:43] <[1]spike> ok ill try now
[05:43] <lifeless> twb: you can do http trace and pings :) though I don't think thats what [1]spike means.
[05:43] <twb> Granted
[05:44] <[1]spike> unsuccessful
[05:45] <zroysch> twb: im not familiar with chroot
[05:45] <lifeless> host www.google.co.nz
[05:45] <[1]spike> i just want the ubuntu server going thru the proxy till i can install the necessary hardware using apt-get so it can become the server
[05:45] <twb> Isn't host deprecated? ;-)
[05:46] <lifeless> twb: not at all
[05:46] <twb> Oh right, it just wants me to "upgrade" to bind9-host, with its bloaty deps
[05:47] <zroysch> rebooting. maybe one day i'll be able to find out what is using mounts
[05:48] <[1]spike> Could it be something to do with the fact i dont have a user name and password? The context is supposed to be http_proxy=http://username:password@host:port/
[05:48] <[1]spike> and im just using http_proxy=192.168.1.2:6588
[05:48] <lifeless> nothing t do with username and password
[05:49] <lifeless> however, you will want your variable to be http_proxy=http://192.168.1.2:6588/
[05:49] <[1]spike> and then use export http_proxy?
[05:49] <[1]spike> ill try now
[05:49] <lifeless> but 'ping' does not use http proxies
[05:50] <lifeless> so it won't help your problem.
[05:50] <twb> You should also be testing with curl -v, not some crappy GUI browser.
[05:50] <lifeless> if you want ping to work, you will need routing set up
[05:50] <[1]spike> i just want apt-get to work
[05:51] <[1]spike> so i can install my network card
[05:51] <[1]spike> then the ubuntu server doesn't need a proxy because it will be the server
[05:51] <cef> you want apt to work with a proxy?
[05:51] <cef> cli?
[05:51] <[1]spike> i use a 3g satalite dial up
[05:52] <[1]spike> i cant install it without being online with apt-get?
[05:52] <[1]spike> or is there another way?
[05:53] <[1]spike> so how can i test if the proxy is working?
[05:54] <cef> first up.. your proxy.. that same machine is the one you want to be able to apt-get, or is it a different machine?
[05:54] <[1]spike> different.  I have all windows machines.  I want to have all linux instead.  But my adsl is satalite dial up so i have a windows machine running squid so i can set up the ubuntu server
[05:55] <[1]spike> then ill set up squid on there and change all pc's to ubuntu and debian
[05:56] <cef> ok.. so you're installing a linux box, or updating an existing one (for apt?)
[05:56] <[1]spike> all new, never ever used linux before
[05:56] <[1]spike> installed last night
[05:57] <cef> but they're already installed, yes?
[05:57] <[1]spike> mistakingly i set a 128 digit pass phrase to.  Didn't realise i had to type it every startup.  Hope Ubuntu server doesn't need restarting as often as microsoft server
[05:57] <[1]spike> yes already installed
[05:58] <cef> pass phase for what, wireless?
[06:03] <cef> is it just me, or should all that needs to be done here is set up apt to use the squid cache by setting an option in /etc/apt/apt.d/ ?
[06:03] <cef> sorry <- work calls, got busy here
[06:04] <[1]spike> sorry work to
[06:04] <cef> btw: you did check to make sure that the windows box running squid has a hole in its firewall?
[06:04] <twb> IIRC apt-get can't speak digest auth
[06:05] <cef> oh crap it's http auth on the proxy? ewwww
[06:05] <[1]spike> the passphrase for the i dont know.  It wasn't an option it just said i had to have a passphrase.  It suggested at least 20 letters and numbers
[06:06] <[1]spike> as soon as the pc turns on i have to type in the passphrase then i have to login
[06:06] <[1]spike> i turned windows firewalls off cause they all suck
[06:06] <[1]spike> and as i said the other pc's browse the web fine
[06:07] <[1]spike> but they have gui's were i can enter the proxy for socks http ftp ect ect in a "Normal for me" windows format.  ie. Ip here port there
[06:07] <twb> 16:50 <twb> You should also be testing with curl -v, not some crappy GUI browser.
[06:08] <[1]spike> curl -v says not installed
[06:08] <twb> 16:37 <twb> In any case, you should now try to talk to the service using netcat
[06:08] <twb> 16:38 <twb> First try on the proxy server itself, then try on your test machine
[06:09] <twb> Did you do that?
[06:09] <[1]spike> need to use apt-get to install it. and apt get always says package not found
[06:10] <twb> [1]spike: you need to install netcat?
[06:10] <[1]spike> yup im trying
[06:11] <twb> How did you do the initial install?
[06:11] <twb> If it was from a CD, it should have those packages in its pool
[06:11] <[1]spike> i cant run netcat on windows? cant find a windows download
[06:12] <[1]spike> i installed by putting the iso on the usb drive
[06:12] <twb> You could try gnuwin32.sf.net, but I don't know about Windows.
[06:12] <twb> [1]spike: well, put that back into your sources.list
[06:13] <[1]spike> ok now your talking french again ill just google it :)
[06:13] <twb> Je ne parlez pas francais.
[06:15] <[1]spike> lol
[06:15] <[1]spike> this is making sence
[06:15] <[1]spike> i just need to find the linux answer for windows cd ..
[06:15] <[1]spike> because pwd says /user/home and i need to be in /ect/
[06:17] <[1]spike> lol jokes on me.
[06:17] <[1]spike> its the same command
[06:20] <[1]spike> ok ive found /etc/apt/sources.list but how do i open it?
[06:25] <[1]spike> am i like way to newb to be trying to set up an ubuntu server? lol
[06:27] <Ash-Fox> http://imgs.xkcd.com/comics/devotion_to_duty.png
[06:32] <[1]spike> sudo cp /etc/apt/sources.list /etc/apt/sources.list.backup
[06:45] <[1]spike> ok sorted sorta
[06:46] <[1]spike> i tried sudo apt-get install curl
[06:46] <[1]spike> but it still said package not found
[06:46] <[1]spike> However it is now attempting to get the packages from 192.168.1.2/ubuntu/
[06:46] <[1]spike> which is my network mirror of the ubuntu server iso
[06:50] <Roxyhart08> hi somebody have running wireshark on ubuntu?
[06:51] <twb> Roxyhart08: it works the same as anywhere else
[06:51] <[1]spike> ok more success sorta. i ran apt-get update now when i run sudo apt-get install curl, it cant find curl but says it found refrences to it so it has probably been removed from this ditribution
[06:51] <[1]spike> package curl has no installation candidate
[06:52] <[1]spike> so im figuring i need to download curl
[06:52] <[1]spike> so i will google ubuntu server curl package download
[06:52] <[1]spike> and put this in my 192.168.1.2/ubuntu/ directory
[06:54] <ChmEarl> looking in /usr/share/consolefonts/, it seems the smallest console font is VGA8 or Uni18?
[06:55] <ChmEarl> that should be Uni-VGA8
[06:55] <[1]spike> but do i need curl.dsc or curl.tar.gz or curl.diff.gz or all of the above?
[06:55] <twb> ChmEarl: here, the smallest font is Terminus 12x6
[06:56] <ChmEarl> I tried that one, but my system ignore it and went back to kernel font
[06:57] <Roxyhart08> hi someone have intalled wireshark?
[06:58] <twb> !anyone
[06:58] <ChmEarl> by kernel font I mean that set by init script
[06:58] <twb> ChmEarl: I know what you're talking about.
[06:59] <twb> aptitude install console-setup kbd+M console-terminus+M && sudo dpkg-reconfigure -plow console-setup
[06:59] <ChmEarl> the first font setting is here /etc/console-tools/config, then later it get it from /etc/default/console-setup
[07:00] <ChmEarl> twb, that mystical cmd will change all fonts to your setting?
[07:01] <twb> ChmEarl: console-setup does it.  It should be installed by default.
[07:01]  * Callum__ finds it stupid that Ubuntu Server doesn't include install-mbr >_>
[07:02] <twb> ChmEarl: you need to use kbd instead of console-tools to select 12x6
[07:02] <ChmEarl> twb, you recommend me to run that cmd , all on one line?
[07:02] <twb> ChmEarl: I don't really care what you do
[07:02] <ChmEarl> :)
[07:02] <Callum__> I have this new RAID array I want to use but since I can't install install-mbr ATM because the site the server is at doesn't have Internet yet, I'm stuck >_>
[07:03] <[1]spike> ok so ive hit a dead stop.  Does anyone know were the packages for apt-get are on the installation cd? so i can add new ones
[07:03]  * ChmEarl pastes cmd into editor to check formatting
[07:03] <twb> [1]spike: you can't "just add new ones"
[07:03] <[1]spike> dang it
[07:04] <[1]spike> i can download new ones lol
[07:04] <twb> Callum__: what do you expect install-mbr to do?
[07:05] <Callum__> twb: it isn't obvious in its name? >_> install an MBR to a RAID array so I can install partitions on it...
[07:06] <ChmEarl> twb, cmd ran fine and I changed to terminus 12x6 -- thanks
[07:06] <twb> Uh, installing an mbr is just a cat or dd.
[07:06] <twb> Unless you mean one of grub's retarded non-static MBRs, in which case you use install-grub or something
[07:08] <twb> Not that md arrays can have partitions, at least traditionally
[07:28] <[1]spike> so how do i install curl without apt-get?
[07:30] <twb> [1]spike: you don't
[07:31] <[1]spike> and i cant add packages to /main/pool/c/curl?
[07:31] <[1]spike> \pool\main\c\curl i mean sorry
[07:32] <twb> [1]spike: correct; you can't.
[07:33] <[1]spike> that makes no sence lol
[07:33] <[1]spike> i can download the necessary package but can't do anything with it :S
[07:34] <[1]spike> even thought the online mirrors have the same file structure but more data in the files
[07:34] <Speedy2> www.search2.net
[07:35] <[1]spike> i think i figured it
[07:36] <[1]spike> you are correct but wrong
[07:36] <[1]spike> i cant add packages but i can ad .deb files
[07:36] <[1]spike> which appear to be package files
[07:36] <[1]spike> very confusing
[07:36] <twb> Which will not be found by apt, because they are not listed in the signed Packages index.
[07:38] <[1]spike> isnt that why i use # dpkg-scanpackages debs file | gzip > debs/Packages.gz
[07:41] <[1]spike> or in my case # dkpg-scanpackages debs /dev/null
[07:41] <[1]spike> or in my case # dkpg-scanpackages debs /dev/nul | gzip > debs/packages.gz
[07:42] <twb> I thought you said your webserver was running Windows
[07:42] <Cappy> folks, anyone around?
[07:43] <[1]spike> it is but i can run that command on another linux box then copy directory with all contents to windows server then mirror them?
[07:43] <twb> !anyone
[07:43] <[1]spike> because wouldn't that now have all the information for apt-get
[07:43] <twb> [1]spike: if you can copy the files to the ubuntu server, just do that, and use file:/// in sources.list
[07:44] <[1]spike> or i could even (Appears more simple) put the .deb files on a usb stick insert it into the server and run those commands
[07:44] <[1]spike> cool im getting somewere :)
[07:44] <[1]spike> now to figure out how to mount my usb stick in linux :)
[07:47] <kaushal> hi
[07:47] <kaushal> I have openssh-server: Installed: 1:4.7p1-8ubuntu1 Candidate: 1:4.7p1-8ubuntu1.2, is there a way to install 1:4.7p1-8ubuntu1.2 without using sudo apt-get update ?
[07:48] <twb> kaushal: why don't you want to do "sudo apt-get update"?
[07:48] <kaushal> since my server is in production
[07:48] <[1]spike> twb can you please give me a hint on how to use my usb stick from CLI lol... google has 3 million+ results for "ubuntu server how to use usb stick" and they all seem to be about installing ubuntu from a usb stick
[07:49] <twb> [1]spike: mkdir /mnt/bugger; mount /dev/sdb1 /mnt/bugger
[07:49] <[1]spike> thanks
[07:50] <kaushal> twb: so the steps are sudo apt-get update and sudo apt-get install openssh-server ?
[07:50] <twb> No.
[07:50] <kaushal> that will update it to 1:4.7p1-8ubuntu1.2
[07:50] <twb> kaushal: you should be doing "aptitude update" and "aptitude safe-upgrade"
[07:51] <kaushal> what does aptitude safe-upgrade do ?
[07:52] <twb> kaushal: do you know how to check the manual?
[07:52] <[1]spike> didnt do much till i added sudo infront of it just said permission denied now it says can't find /dev/sdb1/mnt/bugger in /etc/fsab or /etc/mtab
[07:53] <kaushal> twb: Thanks
[07:53] <[1]spike> im guessing
[07:53] <[1]spike> i had to run the mkdir cmd from one of the listed directories?
[07:54] <kaushal> Installed packages will not be removed as per man aptitude
[07:54] <[1]spike> would i be correct? im in root btw
[07:54] <twb> [1]spike: spaces are important
[07:54] <kaushal> so does it mean I will have 1:4.7p1-8ubuntu1 and 1:4.7p1-8ubuntu1.2 ?
[07:54] <kaushal> I mean both versions
[07:54] <twb> kaushal: you cannot have multiple versions of the same package installed concurrently in Ubuntu
[07:54] <[1]spike> ohh ic
[07:54] <[1]spike> ta dude sorry
[07:55] <kaushal> twb: Thanks for clearing the confusion
[07:57] <kaushal> twb: a last question if i just need to upgrade only openssh-server and not the other packages ?
[07:59] <[1]spike> man your so helpful. but i just need to clarify 1 thing then i think im good for tonight assume = ass outta u and me!
[08:00] <twb> You're not making any friends with such gauche aphorisms.
[08:00] <[1]spike> since i now have the dir i dont need to use the mkdir cmd next time and can just use sudo mount /dev/sdb1 /mnt/bugger
[08:00] <[1]spike> sorry
[08:00] <twb> kaushal: then indeed "aptitude install openssh-server" would be the least ugly way to do it.
[08:00] <twb> kaushal: but I recommend you install all security patches.
[08:01] <kaushal> twb: Thanks :)
[08:01] <[1]spike> i was attempting to explain why i needed verification from a seemingly logical event.  Im just googling gauche aphorisms
[08:03] <twb> [1]spike: I think you already have enough of those.
[08:05] <[1]spike> no i ment i didn't know what gauche aphorisms ment. And again im sorry if my crude word association insulted you
[08:05] <twb> Don't worry about it
[08:06] <twb> I /ignore people that are truly annoying.
[08:06] <Cappy> anyone here able to give me some advice about install a desktop ubuntu with a little optimisation?
[08:06] <twb> Cappy: that would be #ubuntu
[08:17] <ChmEarl> twb, what is a normal initrd.img size for karmic 9.10 server with Lamp & VM Manager? mine is 7MB which seems too big
[08:20] <twb> I'd say anywhere between 5 and 15MB
[08:20] <ChmEarl> k, maybe its not worth it to tweak mkinitramfs
[08:22] <twb> ChmEarl: it's a great way to make your machine unbootable in order to save a few MB of RAM
[08:22] <ChmEarl> twb, my first try was a bust.. I used MODULES=dep in initramfs.conf
[08:23] <twb> Yep
[08:25] <ChmEarl> this box opens using 250 MB RAM, and with one VM its at 650MB
[08:25] <ChmEarl> plenty of headroom
[08:26] <[1]spike> ok how do i shutdown the server? (ie turn off) i tried sudo shutdown now and it takes me to a recovery screen
[08:27] <twb> [1]spike: shutdown -h now
[08:27] <[1]spike> thanks
[08:27] <[1]spike> wait darn thats when i should have used man shutdown?
[08:28] <twb> [1]spike: yes
[08:32] <[1]spike> how long have you been using linux?
[08:32] <twb> Since 2001 or so.
[08:32] <[1]spike> still things to learn?
[08:32] <twb> Sure.
[08:34] <twb> For example, I cannot split an inch-thick pine board with my head
[08:34] <[1]spike> i cant wait till more people in this country use linux.  Then linux will be more obtainable and local support / it's should appear id hope
[08:34] <[1]spike> lol
[08:34] <[1]spike> i ment still things to learn about linux lol
[08:35] <twb> [1]spike: what, New Zealand?
[08:35] <twb> The last LCA was there, FFS
[08:36] <[2]spike> wow. that was ubuntu server starting to serv.  It works. thanks for all your help
[08:45] <derknecht> hi there. i need more loopdevices than the default 8, how can i achive this? (using ubuntu-server 8.10) i have "loop max_loop=16" in /etc/modules, but there are still just 8 /dev/loop devices (i have rebooted). Thanks
[08:46] <twb> derknecht: what are you using loop for, that 8 isn't enough?
[08:58] <derknecht> twb: i use crypt container with losetup und cryptsetup on a local server. And i have 9 container files
[08:58] <twb> derknecht: does your kernel have loop built as a module, or built into the kernel?
[09:04] <ChmEarl> trying dh-make-perl ./Sys-Virt-xxx and it trys to write to my .cpan and fails
[09:06] <ChmEarl> maybe chown root:admin ./.cpan then chmod755 ?
[09:08] <ChmEarl> heck rfb libsys-virt-perl
[09:08] <ChmEarl> request for build -- its beyond my grasp
 [1]spike: what, New Zealand? - *waves from Wellington*
[09:10] <twb> Callum__: I'm assuming he's in .nz because his IP is registered there
[09:11] <Callum__> twb: yeah
[09:12] <twb> ChmEarl: why are you running dh-make-perl as root?
[09:14] <ChmEarl> twb: I ran as user, then it tried to grab a file from cpan. I might have run perl as root once and cpanned
[09:14] <ChmEarl> can I change perms on .cpan
[09:14] <ChmEarl> ?
[09:14] <twb> ChmEarl: that should be easy enough to check: find ~ -user root
[09:15] <jiboumans> ~/.cpan should be owned by you, not root
[09:15] <twb> Right
[09:16]  * ChmEarl fixing now
[09:23] <[2]spike> your correct
[09:23] <[2]spike> Hawke Bay
[09:24] <ChmEarl> it needs libvirt 7.5 or higher, or well
[09:24] <Roxyhart08> hi i got tcpdump installed in my NAT linux server and i would like to check the ips inside of the NAT, somebody know ehich command i should use (currently i just can see the external ips)
[09:28] <screen-x> Roxyhart08: use -i to specify the internal interface
[09:29] <derknecht> twb: its a module, i have added a line in /etc/modules:   loop max_loop=16
[09:29] <twb> derknecht: I think that'll only work if loop isn't already loaded by something else
[09:29] <derknecht> i will test that now (need to restart, but i can't do this now, the services from that server are needed at the moment
[09:30] <twb> derknecht: try putting an entry in /etc/modprobe.d/foo
[09:30] <twb> echo options loop max_loop=16 >>/etc/modprobe.d/foo.conf
[09:30] <twb> I haven't done it myself
[09:30] <screen-x> Roxyhart08: also have you tried tshark and iftop?
[09:31] <twb> Personally, I'd just install tcpdump, generate a trace.pcap, then read it with wireshark on a workstation.
[09:31] <derknecht> twb: i'll  try
[09:31] <twb> Avoids the bloat of wireshark-common
[09:32] <twb> (OTOH I tried to do that the other day, and couldn't work out how to make tcpdump generate a .pcap without access to man-db.)
[09:33] <screen-x> twb: sometimes I prefer the output of tshark.
[09:33] <Cappy> 194 people here and no one is talking? has someone broken freenode?
[09:34] <twb> screen-x: tshark is a pita to read
[09:35] <twb> I usually end up using the "detailed xml" format from it, then filtering out attributes with xmlstarlet
[09:35] <Roxyhart08> thanks screen-x i will try
[09:36] <screen-x> twb: yeah I use when looking for detail as well.
[09:39] <sacarde> hi
[09:39] <sacarde> what is rmnologin service for ?
[09:41] <twb> sacarde: to prevent users from logging in before e.g. they have a home directory
[09:41] <twb> rmnologin just removes the "don't let them log in" stub file, which pam_nologin checks for
[09:52] <sacarde> if I want to remote-login ?
[09:53] <sacarde> I have to stop it?
[09:58] <twb> sacarde: if your system is working, it'll be removed early during boot
[10:00] <sacarde> and server-xorg-input-wacom ?
[10:00] <twb> sacarde: what about it?
[10:00] <sacarde> what is it for?
[10:02] <twb> sacarde: apt-cache show will tell you taht.
[10:02] <sacarde> I have problem ..... I want to remote login with xdmcp with ubuntu-server
[10:03] <twb> That's not a problem, that's a goal.
[10:13] <dholbach> hiya
[10:13] <dholbach> I'm looking for speakers for https://wiki.ubuntu.com/PackagingTraining - who wants to do ara a favour?
 dholbach, one idea: web applications that use a lamp stack. A way to package them to install the lamp dependencies, and configure it directly
[10:13] <dholbach>  dholbach, i.e. drupal
[10:14] <twb> dholbach: that's not a page
[10:14] <dholbach> oops, yes
[10:14] <dholbach> https://wiki.ubuntu.com/Packaging/Training
[10:16] <twb> Packaging apache and friends is something you'd teach to newbies?
[10:16] <huats> dholbach, I really think that ara's idea is great : packaging web applications might be interesting
[10:17] <dholbach> twb: not apache, but web applications - it's very easy to include something like "this is what we expect attendants of the session to know already"
[10:18] <huats> dholbach, you might also give a shot at the mozilla team since I know they had quite a detailed process for packaging extensions... and they have a long list of extensions that needs packaging
[10:18] <twb> Heh, PHP apps have the worst packaging of anything I've ever seen! :-)
[10:18] <twb> At least stdio C utilities typically ship with a makefile, and if you're lucky even use autotools.
[10:18] <dholbach> twb: and it's not something that I'd teach :) (and I wouldn't call people who want to help out "newbies")
[10:18] <dholbach> huats: nice idea
[10:18] <twb> Hahaha
[10:18] <twb> People who WANT to help are the ones not jaded by too many years on the battlefield.
[10:19]  * dholbach adds a note to the job description of speakers: "be optimistic"
[10:19] <dholbach> thanks huats
[10:20] <huats> you're welcome dholbach
[10:23] <twb> I would think that packaging a python project would be a saner introduction to packaging niceties, since they're usually relatively sensible upstream, and there's a well-defined python policy
[10:23] <twb> Maybe there's a php policy, too...
[10:30] <sacarde> how suggest me to allow remote login xdm ?
[10:31] <twb> X -broadcast
[10:31] <sacarde> ? dont understand
[10:31] <sacarde> I have: on port 177 xdm listening
[10:31] <twb> Run that.  It'll ask the local network for its XDMCP servers, and connect to the one that responds.
[10:32] <sacarde> ok
[10:33] <sacarde> but I dont view ubuntu-server responding
[10:33] <sacarde> but i have on port 177 xdm listening
[10:33] <twb> If you know what the app server's IP is, you can use X  -query IP instead
[10:33] <twb> sacarde: xdm doesn't run on the x terminal, it runs on the app server (IIRC).
[10:34] <twb> Last time I used XDMP, the POWER G3 was in vogue.
[10:34] <twb> *XDMCP
[10:34] <sacarde> I dont understand
[10:35] <sacarde> I have ubuntu-server and another linux-client
[10:35] <sacarde> I configured ubutu-server-xdm to accept remore ;login
[10:35] <twb> sacarde: OK, then "ubuntu server" is your app server.
[10:36] <sacarde> yes
[10:36] <twb> Run X -query IP on the other host, the X terminal
[10:36] <sacarde> it have 177 port listening
[10:36] <dholbach> twb: we ran a number of those sessions already
[10:36] <sacarde> I try to logon in some ways
[10:36] <twb> dholbach: eh?
[10:36] <twb> dholbach: oh, right
[10:36] <dholbach> :)
[10:37] <screen-x> Is there any standard way of testing for powersupply failure on a machine with dual power supplies?
[10:37] <sacarde> I run: X :1 vt8 -dpi 100 -query <servIP>
[10:37] <sacarde> ut I view only black screen
[10:37] <twb> screen-x: sure: pull out the other one
[10:37] <twb> screen-x: if the machine goes down, the first one was dead
[10:37] <screen-x> twb:  :p
[10:38] <screen-x> twb: I mean in /proc or something like that
[10:38] <twb> screen-x: I expect that your hardware vendor will tell you to install something proprietary
[10:38] <screen-x> twb: :(
[10:38] <twb> screen-x: I haven't looked into dual PSUs myself
[10:39] <twb> IME the PSU is rarely the part that fails
[10:39] <screen-x> I had two fail recently :(
[10:40] <twb> screen-x: are you in a controlled environment, or are you running your server in an unventilated cupboard in the middle of the Pilbara?
[10:41] <screen-x> twb: dedicated server room with air conditioning. Both PSUs going through UPSs, no unsual spikes/dips recorded by the UPS, all very strange.
[10:41] <screen-x> But very glad of redundant PSUs at that point..
[10:41] <twb> Are they from the same batch?
[10:41] <sacarde> twb, in which forum can I ask for my problem ?
[10:41] <screen-x> twb: probably.
[10:41] <twb> sacarde: what problem?
[10:42] <sacarde> remote login in ubuntuserver
[10:42] <twb> Well, it's an X issue.
[10:42] <twb> If the Xorg people have a support forum, that'd be the place to ask.  I don't think they have one, except perhaps mailing lists.
[10:43] <sacarde> ye
[10:43] <twb> sacarde: you should be checking the stdout/stderr and logfiles, and failing that, packet sniffing with e.g. wireshark
[10:52] <sacarde> twb, installing gdm , remotelogin works OK
[10:52] <sacarde> it was xdm problem
[10:52] <sacarde> I think
[12:24] <bartmon> Hi! I'm doing a project for a college course and I need to test early KVM versions. That means kernel 2.6.20 and onwards. Can I expect a karmic server to work with such an old kernel? How would you go about doing this - Trying out old releases or can I expect stuff to break in  big way if i plug an old kernel on newer userspace?
[12:36] <persia> ttx: Hey:  I wanted to ask you about your suggested config snippet for bug #231060 : do you know of any reason that shouldn't just be uploaded?
[12:36] <ttx> persia: looking
[12:38] <ttx> persia: you also need to modify libvirt so that it calls dnsmasq with --interface=<virtual-bridge>
[12:38] <persia> lool has also been looking at this, and wondered by libvirt uses "--except-interface lo" rather than explicitly binding some virbr address.  Does anyone know the answer?
[12:38] <persia> Aha!
[12:38] <persia> So these are related.  Does libvirt know which of virbr* is <virtual-bridge> ?  Is it always virbr0 ?
[12:39] <lool> Eh seems ttx came to the same conclusion  :)
[12:39] <lool> persia: Note there might be multiple ones
[12:39] <lool> persia: These are defined in teh XML file defining the network
[12:39] <persia> Hrm.
[12:39] <persia> That gets tricky then.
[12:39] <ttx> persia, lool: yesvirbr* determination seems like the last hurdle, that's why I wanted someone with inside libvirt knowledge to fix that one
[12:40] <lool> persia: /etc/libvirt/qemu/networks/autostart/default.xml
[12:41] <lool> But the libvirt code is probably aware of the interfaces it uses
[12:41] <lool> I had a look a while ago and the code is quite heavy and has hardcoded string lengths and the like, it's not pretty
[12:48] <persia> Hm.  It *used* to have --interface virbr0, but that got commented out :(
[12:49] <persia> Just to make sure, If I restore --interface (using the correct value), I can drop --except-interfaces from the dnsmasq call, right?
[12:50] <ttx> persia: yes, --bind-interfaces makes it bind to all interfaces but the ones in --except-interfaces
[12:51] <ttx> persia: and --interface is the alternative
[12:51] <ttx> so you should also drop --bind-interfaces from the libvirt call
[12:52] <persia> Should I?  Based on man dnsmasq, I'd think I'd want to keep that.
[12:52]  * ttx rechecks
[12:52] <ttx> you're right
[12:53] <ttx> --bind-interfaces just avoids binding the wildcard address, which is desirable in all cases
[12:53] <persia> Especially when we're trying to make it possible to run two simultaneous daemons :)
[12:54] <ttx> heh
[12:54] <ttx> (in all cases = for the system dnsmasq and for the libvirt dnsmasq)
[13:14] <ttx> smoser: yo
[13:15] <jiboumans> ttx: he has a day off today
[13:15] <persia> Ah, need to preserve --except-interface lo anyway, due to how --interface works :(
[13:16] <ttx> jiboumans: arh
[13:25] <soren> persia, ttx: --interface never really worked there.
[13:26] <soren> persia, ttx:
[13:26] <soren> persia, ttx: Back when that code was first introduced, it was found to be racy.
[13:26] <ttx> soren: because the main one wasn't running with --except-interface ?
[13:26] <soren> No.
[13:26] <soren> That's been there all along.
[13:27] <persia> soren: I saw that in the code comment, and was hoping the race had been resolved somewhere else (kernel maybe?)
[13:27] <soren> I'm not sure.
[13:27] <soren> I don't think I ever worked out where it was.
[13:27] <persia> soren: I've a use case myself that requires this configuration.  Any suggestions on how to make it "Just work"?
[13:28]  * persia admits to currently being engaged in extended yak-shaving, but likes fixing bugs anyway
[13:28] <soren> persia: Not off the top of my head. I'd ask in #virt over on OFTC.
[13:28] <soren> The problem is that from the C api, to bind to a particular interface, you specify its IP.
[13:28] <soren> so by the time dnsmasq starts and tries to do this, this IP must be set.
[13:29] <ttx> hmm, that rings a bell
[13:29] <soren> For some reason, it wasn't, when Dan first wrote that code.
[13:30] <ttx> there was a fix recently, let me check if it's related
[13:33] <ttx> Re-read the set of network interfaces when re-loading
[13:33] <ttx> 	    /etc/resolv.conf if --bind-interfaces is not set. This
[13:33] <ttx> 	    handles the case that loopback interfaces do not exist
[13:33] <ttx> 	    when dnsmasq is first started.
[13:33] <garymc> Hi Peeps, im trying to install a shopping cart. It says I need cURL enabled. Anyone know how I add or enable this?
[13:33] <ttx> soren: that's in 2.52 ^
[13:34] <ttx> does not help, methinks
[13:34] <soren> garymc: Peeps is not here, but you coul try installing the curl package.
[13:34] <soren> ttx: /me concurs
[13:34] <garymc> soren lol Peeps is short for People :)
[13:34] <garymc> ok how do I install the curl package. and will it destroy my current setup?
[13:35] <soren> curl is lower caps for cURL.
[13:35] <soren> Like you install any other package on ubuntu: sudo apt-get install curl
[13:35] <soren> As to whether it will destroy your current setup, I can't say.
[13:35] <soren> Does your current setup depend on not having the curl package installed?
[13:36] <garymc> :S
[13:36] <soren> If it's set to autodestruct if /usr/bin/curl pops up, it'll not go well, I predict.
[13:36] <soren> Otherwise, you're probably fine.
[13:36] <smoser> ttx, i'm not really in today if you (or someone) could manually fix https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/525675 as I suggest in an image, it'd be great to see how much further you get.
[13:36] <garymc> OK i installed, now how do I turn it on?
[13:36] <smoser> fwiw, i'm not really sure how this code would not have failed in alpha2 if this is the problem though, so i might be off base.
[13:37] <ttx> smoser: ok, thanks for the pointer
[13:37] <screen-x> garymc: to enable it, you probably need to give your webapp the path to curl.
[13:37] <soren> garymc: It is on.
[13:38] <garymc> damn i need a decent turorial on this
[13:39] <garymc> my shop cart still says it is set to off
[13:40] <screen-x> garymc: probably because some install script initially couldn't detect curl. Now you have it installed, you need to re-run that script or find a config file and tell it you now have curl.
[13:40] <diago> we are in the middle of converting our IPs over to a new block. Currently all machines communicate through router#1 but we have setup router#2 with the new IPs and would like a slow transition. How can we tell the machines that if they receive a request from router#2 to use router#2 as the gateway?
[13:40] <garymc> screen-x : you mean with the shopping cart
[13:40] <garymc> ?
[13:41] <garymc> should I restart apache?
[13:41] <screen-x> garymc: that prob wont help.
[13:41] <garymc> no it iddnt
[13:41] <garymc> *didnt
[13:41] <diago> garymc: magento?
[13:42] <garymc> magento?
[13:42] <garymc> no open cart
[13:42] <diago> ah, I had that problem with magento
[13:42] <garymc> open cart is the shopping cart
[13:45] <persia> Hrm.  So danpb seems to suggest only using http://wiki.libvirt.org/page/Libvirtd_and_dnsmasq to make them work together, which doesn't meet our use case.
[13:45] <garymc> So anyone know anythign about cURL?
[13:45] <screen-x> diago: use alternate routing tables.
[13:45] <diago> garymc: can you recap your problem, I came in late
[13:46] <persia> soren: What do you think about adding debconf to dnsmasq to restrict to some interface (eth0 by default)?
[13:46] <diago> screen-x: I have tried route add -host <router-ip> gw <router-ip>
[13:46] <garymc> diago : I need to have cURL running for my shopping cart. I just ran "sudo apt-get install curl" now I dont know what to do now
[13:47] <diago> garymc: you probably need php5-curl
[13:47] <soren> persia: Sounds great.
[13:47] <garymc> diago : so do i uninstall curl?
[13:47] <diago> no that won't matter
[13:47] <diago> it may need that anyway
[13:48] <garymc> ok im installing php5-curl now what?
[13:48] <garymc> my shop cart still says its set to OFF
[13:48] <diago> restart apache
[13:48] <garymc> ok done
[13:49] <garymc> my shop cart now says set to on
[13:49] <garymc> :)
[13:49] <screen-x> diago: http://lartc.org/howto/lartc.rpdb.html#LARTC.RPDB.SIMPLE
[13:49] <garymc> is that it?
[13:49] <diago> that's it garymc
[13:49] <screen-x> garymc:  :)
[13:49] <garymc> Cool diago
[13:49] <garymc> Thanks Guys
[13:49] <garymc> Ubuntu is so easy
[13:49] <garymc> :P
[13:49] <diago> screen-x: thanks, I'll give that a read
[13:50] <tgardner> cjwatson, I have chroot bind and devpts mounts in /etc/fstab that appear to be stopping a Lucid boot. Have you seen this bug yet?
[13:55] <diago> screen-x: that doesn't make much since to me, I'm having a hard to putting that in as my scenario
[13:55] <diago> basically, router#2 is using iptables and DNAT to get the external IPs to the correct internal.
[13:56] <screen-x> diago: you have two routers routing packets to servers, which must reply via the same router?
[13:56] <diago> I need to leave the default gateway in place on the machine for the time being to be router#1 BUT if a request comes from router#2 I need that traffic to go back out to router#2
[13:56] <diago> screen-x: that sounds corrent
[13:56] <diago> correct**
[13:57] <screen-x> diago: yep, so you create an alternate routing table that has router2 as the default route, then you add a rule that says that traffic to router2 must be routed according to the alternate routing table.
[13:57] <screen-x> hmmm
[13:58] <diago> I feel that has what I have done and it's not working
[13:58] <diago> I'm not sure what the machines see as the source IP though.
[13:58] <screen-x> diago: I did this a while ago while changing ISPs, but the details are fuzzy now that I think about it.
[13:59] <diago> screen-x: that's exactly what we are trying to do
[13:59] <diago> We need them both up until DNS is totally finished propagating
[14:00] <screen-x> I think I might have ended up adding additional IPs to the servers, then the routers routed packets to the same servers but reffering to them by different IPs.
[14:00]  * screen-x wishes he wrote all this down at the time. 
[14:01] <diago> screen-x: do you know what iptables SNAT does exactly>
[14:01] <diago> ?
[14:01] <screen-x> source nat changes the source field of an IP packet
[14:15] <bogeyd6> !root
[14:15] <bogeyd6> !noroot
[14:19] <soren> bogeyd6: Can you do that somewhere else, please?
[14:19] <bogeyd6> soren, i took it to privmsg
[14:19] <soren> Good.
[14:20] <bogeyd6> soren, will the cloud be in the next LTS?
[14:20] <soren> Of course.
[14:21] <bogeyd6> im going to put in for a career stateside at ubuntu
[14:21] <bogeyd6> information systems manager
[14:23] <Ziber> !wfm
[14:24] <Ziber> !ping
[14:25] <inveratulo> must be written in perl heh
[14:29] <cjwatson> tgardner: no, check with Keybuk?
[14:29] <garymc> how difficukt is it to setup multiple domains on my webserver. Taking into account I already have one domain on there working at the minute?
[14:30] <garymc> How do I do it, without messing up my Website thats already up and running?
[14:30] <tgardner> cjwatson, he's on the bug subscriber list, but he  doesn't seem to be around today. I filed it against mountall for lack of a better idea.
[14:30] <inveratulo> garymc: yo ushould be able to make changes on the fly and do graceful reloads of your webserver
[14:30] <ttx> smoser: the eucalyptus metadata service is mostly alright -- it just replies error 500 to ephemeral0 request, which makes it fail when queried by boto1.9
[14:31] <garymc> ok
[14:31] <ttx> smoser: I'll let kirkland investigate further
[14:31] <garymc> So am i best to run my other websites on a seperate server ? As I use this current webserver for call centre database input
[14:31] <cjwatson> tgardner: I just got back from holiday, so conider me comprehensively out of the loop
[14:31] <inveratulo> garymc: you can utilize the "checkconfig" or "checkconf" keyword to check your config files before you reload
[14:32] <ttx> smoser, kirkland: see https://bugs.launchpad.net/ubuntu/lucid/+source/eucalyptus/+bug/525675/comments/5 for details
[14:32] <inveratulo> garymc: i suppose it ultimately depends on your usage requirements, if you expect tons of traffic it may be best to keep them separate
[14:32] <garymc> Ok how do I setup multiple domains
[14:32] <tgardner> cjwatson, buried under a mountain of email? I'm gonna be gone for a month soon and am not looking to my return wrt email.
[14:33] <inveratulo> garymc: typically you can either decide to run a subdomain or a virtual host... check it out: http://httpd.apache.org/docs/2.0/vhosts/examples.html
[14:33] <cjwatson> tgardner: bugs mostly
[14:45] <diago> screen-x: I got it
[14:46] <kpettit> Any recommendations for a good blacklist type app?  Looking for something to give me updated blacklist IP's, mainly for ssh
[14:46] <diago> using iptables snat for anything going out of the lan card and route on the internal mchines
[14:47] <bogeyd6> kpettit, dnsbl?
[14:47] <kpettit> haven't heard of that one.  I'll look for it.
[14:47] <bogeyd6> dronecheck is pretty good
[14:47] <kpettit> I'm using arno, and it has a pretty good ssh blacklist.  But I hate watching all these attempts
[14:47] <kpettit> oh cool.  I'll look at that one too
[14:48] <bogeyd6> which was i assume you are wanting
[14:48] <bogeyd6> !fail2ban | kpettit
[14:48] <bogeyd6> hmm
[14:48] <kpettit> ?
[14:48] <bogeyd6> kpettit, fail2ban is probably more realistic of what you want
[14:48] <bogeyd6> easy to setup
[14:49] <kpettit> ohhh, I like that.  It's in package manager as well
[14:49] <garymc> inveratulo I want to run virtual host
[14:49] <bogeyd6> http://www.fail2ban.org/wiki/index.php/Main_Page
[14:49] <garymc> What do i need to do
[14:49] <kpettit> looks perfect.  thanks bogeyd6
[14:49] <garymc> Do I need to create new folders for each website? where? how?
[14:49] <inveratulo> garymc: generally yes, some people use versioning systems, whatever floats your boat
[14:49] <bogeyd6> kpettit, it is perfect, i set mine at 5 fails in 10 minutes, but our admin team is the only ones using it
[14:50] <garymc> inveratulo I need the simplist one :S
[14:50] <bogeyd6> kpettit, its also available from the apt-get install fail2ban
[14:50] <inveratulo> garymc: well the apache docs are pretty explicit so they can probably walk you through some pretty good examples, if you have any specific questions i'm sure someone here can answer
[14:51] <kpettit> bogeyd6, sweet.  Just installed it vai apt-get
[14:51] <garymc> ok i want to do namebased virtual hosts. Am I thinking right?
[14:52] <inveratulo> garymc: right, that's like if you wanted to host www.foo.com and www.bar.com off the same server
[14:53] <bogeyd6> inveratulo, which is probably what is going on. i wonder how many ip's he has
[14:53] <inveratulo> garymc: well there's a gazillion ways it could be configured, so if there are specific questions, we could probably assist
[14:55] <bogeyd6> garymc, or be more specific as to your goal
[14:58] <screen-x> diago: great, how did you do it in the end?
[15:01] <kirkland> ttx: hi
[15:01] <diago> screen-x: I'll do one more test and pastie
[15:03] <stephen`> hi... I am an experiened debian user, but have been using centos for 3+ yrs; now starting to use ubuntu; is there a doc somewhere which explains core technical differences between ubuntu and debian? eg upstart... anything else?
[15:03] <kirkland> ttx: i'll get that bug fixed
[15:05] <bogeyd6> stephen`, http://help.ubuntu.com
[15:05] <bogeyd6> stephen`, more specifically https://help.ubuntu.com/9.10/serverguide/C/index.html
[15:06] <bogeyd6> s/9.10/server_version
[15:06] <Disconnect> anyone know if its possible to enable hw virt on an hp dl165 w/o having to be there? (bios setting is wrong - by the time hp finished the doa repair it was after midnight and i spaced on resetting it..) i know dell has a tool to do it but what about hp?
[15:07] <bogeyd6> Disconnect, the hp system management
[15:07] <stephen`> bogeyd6: thanks, that looks useful; I was wondering if there was more a 'here are 4 or 5 things you need to know, coming from a debian background'
[15:07] <bogeyd6> !anyone | stephen`
[15:08]  * Disconnect asked the real question even though it started with "does anyone" :)
[15:09] <stephen`> ubottu: erm?  I don't believe I asked 'does anyone' - I asked a specific question
[15:09] <stephen`> bogeyd6: ^^
[15:09] <bogeyd6> Disconnect,  http://h18000.www1.hp.com/products/servers/management/agents/index.html
[15:10] <Disconnect> awesome thanks :)
[15:10] <bogeyd6> you gotta use the debian packages
[15:10]  * Disconnect was just discovering he should have set up the bmc before leaving anyway, so a trip might be necessary. Here's hoping not :)
[15:10] <bogeyd6> HP has a hard** for redhat and sles
[15:10] <screen-x> stephen`: maybe http://www.ubuntu.com/community/ubuntustory/Debian
[15:11] <bogeyd6> Disconnect, http://h18000.www1.hp.com/products/servers/management/remotemgmt.html
[15:11] <stephen`> screen-x: yeah... i know all that! Thanks though; Maybe what I want doesn't exist; It's cool - I'll work it out.
[15:11] <diago> screen-x: http://pastie.org/836856
[15:12] <bogeyd6> !pastebin | diago
[15:12] <stephen`> I mean, for example, I know that Ubuntu uses upstart, and debian doesn't by default (ie I've never used it).  That's fine - I'll read the upstart documentation.
[15:13] <stephen`> I'm wondering if there are any other areas like that that I could focus my attention on, or if I just treat ubuntu as debian, I'll more or less be fine.
[15:13] <bogeyd6> they like paste.ubuntu.com used
[15:13] <bogeyd6> stephen`, well i use ubuntu and sles, big differences? all the config directories and especially /var/
[15:14] <diago> so touchy over a paste
[15:14] <bogeyd6> stephen`, put simply there are huge differences but they all work basically the same
[15:14] <stephen`> bogeyd6: yeah right - i know the differences between eg debian and rhel - that's fine... I'm just after a sense of how different, in real, nuts and bolts terms, debian is from ubuntu.
[15:16] <bogeyd6> ah ok
[15:16] <bogeyd6> then you want this
[15:16] <bogeyd6> stephen`, http://www.wikivs.com/wiki/Debian_vs_Ubuntu
[15:18] <stephen`> so I've done slink -> potato -> etch -> lenny; but only really experienced ubuntu on a laptop, so never really thought about what's going on underneath
[15:18] <stephen`> (whereas have run 100s of debian servers)
[15:18]  * stephen` looks at that link
[15:19] <screen-x> diago: Thanks for the info. Glad you could sort it without too much complication.
[15:20] <bogeyd6> screen-x, why did you send me the becoming a new member thing?
[15:21] <andol> zul: How sure are you about bug #403599 actually being fixed? Also, when marking bug #525041 as Invalid, did you check the potential connection to 403599?
[15:25] <ttx> kirkland: thx
[15:27] <kirkland> ttx: i don't entirely see what's causing the problem yet, but I'll work with smoser  on it
[15:27] <ttx> kirkland: I think I nailed the root cause
[15:27] <ttx> kirkland: I've an interesting eucalyptus-side log I'll attach to the bug
[15:27] <kirkland> ttx: boto?
[15:28] <ttx> kirkland: no. Metadata service
[15:28] <ttx> kirkland: it's certainly possible to work around the bug in boto
[15:31] <Italian_Plumber> hello.  I'm trying to get a program to run at boot time, and have it run my a non-priviledged user, from that user's environemnt.  "~" has to be that user (david)'s home directory.  As you can see here: http://pastebin.com/m276db94a it is not.  What am I doing wrong?
[15:34] <Italian_Plumber> man I'm not even sure if I'm asking the right question
[15:34] <inveratulo> Italian_Plumber: look up the "@reboot" directive used in conjuction with crontab for the user you want to run the script as...
[15:35] <inveratulo> Italian_Plumber: oh wait i just saw your pastebin... did you set a home directory for your user in /etc/passwd ?
[15:35] <Italian_Plumber> yes, it's /home/david
[15:35] <inveratulo> so when you do a su - david and then pwd it returns the correct home ?
[15:36] <Italian_Plumber> when root runs the rc.d script at startup, it needs to start rtorrent as the user david, with the user david's environment.
[15:36] <inveratulo> I would utilize crontab to be honest, that's why its there
[15:41] <acalvo> hi
[15:42] <acalvo> if anyone had to develop an application which has to be modular and run in a *nix enviorenment, which programming language will you choose?
[15:42] <acalvo> I've thought c++ with a plug-in web based system
[15:42] <bogeyd6> c++ or php
[15:42] <bogeyd6> python is good too
[15:43] <acalvo> maybe for the *nix backend
[15:43] <acalvo> seems good
[15:47] <ttx> kirkland: confirmed error in eucalyptus side, see my latest comments on the bug
[15:49] <kirkland> ttx: okay
[15:50] <ttx> kirkland: boto could fail a little less miserably, but the root cause is in euca.
[15:54] <Italian_Plumber> looks like the command I was looking for is "sudo -H -u david screen -d -m rtorrent"
[16:04] <zroysch> RAID 1+0 (or 10) is a mirrored data set (RAID 1) which is then striped (RAID 0), hence the "1+0" name. A RAID 1+0 array requires a minimum of four drives: two mirrored drives to hold half of the striped data, plus another two mirrored for the other half of the data. In Linux MD RAID 10 is a non-nested RAID type like RAID 1, that only requires a minimum of two drives, and may give read performance on the level of RAID 0.
[16:04] <zroysch> what so md raid10 is different than normal raid10
[16:19] <kirkland> smoser: around?
[16:20] <smoser> here
[16:20] <smoser> but only shortly
[16:20] <smoser> kirkland,
[16:20] <kirkland> smoser: right, sorry, ttx said you're on a swap day
[16:20] <ttx> smoser: saw my comments on the bug ?
[16:21] <kirkland> smoser: priv message me the bugs you want sponsored/uploaded
[16:21] <kirkland> smoser: and i'll get to them today
[16:21] <smoser> bug number?
[16:23] <smoser> kirkland,
[16:23] <smoser> ttx, i can't see how bug 525675 is not regression, i've been using the 2009-04-04 for most of the lucid cycle
[16:24] <smoser> regression in euca side
[16:24] <ttx> smoser: might have been introduced when they worked on ephemeral0
[16:24] <ttx> also it's not sensitive to the API date
[16:25] <smoser> which probably is ok... (bug really should be)
[16:25] <ttx> it's more about enumerating keys (boto 1.9) vs non enumerating keys (1.8)
[16:25] <smoser> but we have used 1.9 in alpha2
[16:25] <smoser> this code hasn't changed. i've even had several people do the exact same 'get_instance_metadata()' call inside euca guest
[16:25] <ttx> smoser: I'd say, recent regression in euca only shown by using boto 1.9-style enumeration
[16:26] <smoser> well, fine. i really dont care. recent regression in euca.
[16:26] <kirkland> ttx: do you have any pointers in the eucalyptus code for this key handling?
[16:26] <kirkland> ttx: i'm grepping around right now
[16:26] <ttx> kirkland: nope
[16:26] <kirkland> ttx: i haven't touched this bit to date
[16:26] <kirkland> ttx: okay
[16:27] <smoser> and they're infinitely going to have issues if they globally change all their api levels.
[16:27] <smoser> ie, if they fix this, they'r egoing to break someone who explicitly requested 2008-02-02 (or any other level)
[16:27] <kirkland> smoser: they = eucalyptus?
[16:27] <smoser> yes
[16:28] <smoser> the api has levels
[16:28] <smoser> and the user *should* request a specific level and be compatible with that
[16:28] <smoser> if euca changes all levels at the same time, its impossible to make everyone happy
[16:28] <ttx> smoser: ephemeral0 is 2009-04-04 ?
[16:29] <smoser> i can send you a cralw of the ec2 api that i sent to danial at the portland sprint
[16:32] <ttx> smoser: here the issue is about what http://169.254.169.254/latest/meta-data/block-device-mapping/ returns
[16:32] <ttx> if it returns 'emi\nephemeral0\nroot\nswap' it should be ready to serve any of them
[16:33] <ttx> I don't mind it not supporting ephemeral0, I mind it exposing it if it can't serve it.
[16:34] <ttx> smoser: (if that makes sense)
[16:34] <smoser> right.
[16:34] <ttx> to me it's not a boto compatibility or API issue
[16:35] <smoser> on ec2
[16:35] <smoser> $ wget http://169.254.169.254/latest/meta-data/block-device-mapping -O - -q
[16:35] <smoser> ami
[16:35] <smoser> ephemeral0
[16:35] <smoser> root
[16:35] <ttx> same for UEC
[16:35] <ttx> (+ swap) :)
[16:35] <smoser> $ wget wget http://169.254.169.254/latest/meta-data/block-device-mapping/ephemeral0 -O - -q; echo
[16:35] <smoser> sda2
[16:35] <smoser> is there a http://169.254.169.254/latest/meta-data/block-device-mapping/ephemeral0 on uec?
[16:36] <ttx> returns error 500
[16:36] <ttx> but...
[16:36] <smoser> so thats the problem.
[16:36] <ttx> smoser: and that's what my last comments say :)
[16:36] <smoser> sorry.
[16:36] <ttx> $ grep -R ephemeral0 .
[16:36] <ttx> ./clc/modules/image-manager/src/main/java/com/eucalyptus/images/util/ImageUtil.java:  public static BlockDeviceMappingItemType EPHEMERAL = new BlockDeviceMappingItemType( "ephemeral0", "sda2" );
[16:36] <ttx> ./clc/modules/cluster-manager/src/main/java/edu/ucsb/eucalyptus/cloud/cluster/VmInstance.java:    m.put( "block-device-mapping/", "emi\nephemeral0\nroot\nswap" );
[16:36] <ttx> kirkland: ^
[16:37] <ttx> looks like a good starting point
[16:38] <smoser> so anyway... then, ttx, what do we want to do here? hack around this at the moment in cloud-init ?
[16:38] <smoser> ttx, sorry for not fully understanding your last comment, you're correct
[16:38] <ttx> smoser: let's see what Dustin can find / ask eucalyptus guys
[16:38] <ttx> smoser: if it can't be fixed for alpha3 in euca, then we'll look at workaround options
[16:39] <ttx> smoser: but any option looks ugly to me
[16:39] <ttx> smoser: so you can have the rest of your swap day :)
[16:39] <smoser> ok. thanks :)
[16:39] <smoser> one last thign, then i'll go
[16:39] <smoser> could someone please sponsor https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/524999
[16:39] <ttx> kirkland: looking at the code, it should return EPHEMERAL much in the same way it returns EMI
[16:40] <zroysch> what is the best way to backup the data of a failing partition
[16:40] <zroysch> dd or osmething
[16:42] <smoser> ttx, kirkland i'm out now. if you absolutely needed me, call.
[16:42] <ttx> smoser: ack
[16:43] <sam2903> hi, i'm searching for a mailinglist which is easy to manage via browser-interface for non-technical people
[16:43] <ttx> hahaha
[16:44] <ttx> kirkland: - m.put( "block-device-mapping/ephemeral", "sda2" );
[16:44] <ttx> kirkland: + m.put( "block-device-mapping/ephemeral0", "sda2" );
[16:44] <ttx> Thhat may solve it
[16:45] <ttx> (in clc/modules/cluster-manager/src/main/java/edu/ucsb/eucalyptus/cloud/cluster/VmInstance.java)
[16:46] <ttx> kirkland: regression introduced in r906
[16:48] <zroysch> does anyone know how to stop a drive from being used
[16:48] <ttx> kirkland: commenetd on the bug
[16:48] <zroysch> it tells me its busy but nothing is using it
[16:49] <zroysch> nevermind found it
[17:01] <TeTeT> kirkland: is it soon enough to test the proposed eucalyptus package tomorrow? If not I can probably squeeze it in today
[17:01] <garymc> Could anyone be kind enough to help me configure my Virtual Hosts in Apache. Thing is I already have a website running and dont want to mess it all up
[17:02] <garymc> but I need some virtual hosts
[17:14] <Eak> hi can i ask a little
[17:15] <Eak> ubuntu can make hotspot for school ?
[17:15] <ivoks> yes
[17:15] <ivoks> there's chillispot package that can be used
[17:15] <tobibobi> Hi, i'm trying to verify a possible security issue on default images installed by german ISP "Strato". Does one of you run a server with them which was installed using the ISP provided default ubuntu image?
[17:15] <ivoks> make sure you use it on 32bit version
[17:16] <Eak> yep
[17:16] <Eak> thank you very much, sir
[17:16] <ivoks> you'll need access points
[17:16] <sam2903> garymc, edit /etc/apache2/httpd.conf
[17:16] <Eak> now am a student make a project
[17:16] <Eak> now am use linksys wrt54gl for access point
[17:18] <persia> soren: I think I found the issue with dnsmasq: http://paste.ubuntu.com/381715/ seems to be the cause of the demon just dying when the interface isn't ready yet.  Any idea how to make it just poll for the interface instead to some timeout (say 30s), and then exit?
[17:19] <garymc> sam2903 theres noting in my httpd.conf file
[17:19] <gypsymauro> hi
[17:20] <sam2903> garymc:
[17:20] <sam2903> <VirtualHost *>
[17:20] <sam2903>   ServerName myaddress.com
[17:20] <sam2903>   ServerAlias www.myaddress.com
[17:20] <sam2903>   DocumentRoot /var/www/myfolder

[17:20] <gypsymauro> I've added a disc to a raid1 device with mdadm --add /dev/md0 /dev/hda1 its synched but it seems it's used as "spare" how can I make it as default?
[17:24] <Eak> why ubuntu people use more than centos ?
[17:27] <geneticx> hi everyone.
[17:28] <geneticx> just wondering if doing "apt-get update" and "apt-get upgrade" should be done periodically?
[17:35] <geneticx> would there be any risks of applications crashing after updates..like in windows environments?
[17:37] <persia> geneticx: Depends on the application, but it's potentially possible, if there's been an ABI transition and the application dlopen()s stuff.  This should never happen in a supported release, so if you fear this, don't run the development release (other than briefly to help testing).
[17:37] <garymc> Ok anyone got tim eto help this NEWB here get my virtual hosts setup?
[17:38] <garymc> Im reading all the help stuff, but its not happening
[17:43] <geneticx> persia: Sorry not sure if I follow, but what do you mean by "don't run the development release" of what?
[17:44] <persia> Ubuntu.
[17:44] <persia> So, right now, don't run lucid.
[17:44] <persia> If you're running any of 8.04, 9.04 or 9.10, that will never happen.
[17:45] <persia> And when 10.04 comes out, it will never happen for 10.04.
[17:45] <persia> It can only happen when using the latest test release.
[17:49] <geneticx> persia: ah ok, I'm running 9.10 so it shouldn't be a problem right
[17:49] <persia> geneticx: Right.
[17:51] <geneticx> persia: I guess I'm a bit paranoid when it comes to updates to a system considering that I was once in that world where you have to "freak out" if updates make your machine unstable =D
[17:53] <Italian_Plumber> updates only make your machine unstable is you're using FreeBSD. :)
[17:56] <persia> geneticx: There are only a few cases that are special.  You don't get improvements in libc or the kernel until you reboot, you need to restart individual servers when they are updated (and it's good practice to restart when depending libraries are updated).  Some services don't autorestart on pam updates (this is good because you want to *keep* your ssh session), and need to be restarted manually after verifying everything is OK.
[17:56] <persia> There's a few other cases, but by being careful about applying the updates, and with a bit of attention, it usually goes very smoothly.
[17:57]  * persia only restarts servers for kernel upgrades, figuring everything else will just get by, but doesn't have uptime requirements that necessitate restarting services just in case
[17:57] <Italian_Plumber> I know some FreeBSD administrators who are extremely overdue (1-2 years at times) on updates because they don't want to break thier systems.
[17:59] <geneticx> Italian_Plumber: I bet there are many, but I guess they roll on the idea "if it works, don't touch it"
[17:59] <Italian_Plumber> true, but they they fall prey to the idea, "if it hasn't been updated for a while, exploit its security holes"
[18:00] <persia> nmap is handy for finding those :)
[18:00] <geneticx> persia: I see, is there such a thing as selecting only the updates that are necessary such as security updates?
[18:00] <persia> geneticx: Yep.  There's a separation between the -security archive and the -updates archive.  -security contains only security updates, and -updates contains also critical bugfixes.
[18:01] <Italian_Plumber> I think most updates that come over the server side are for security or other minor fixes -- very few of them seem to be for new features.
[18:01] <geneticx> Italian_Plumber: lolz
[18:01] <Italian_Plumber> no?  am I wrong? :)
[18:03] <persia> None of them should be for new features.  That's against policy for -updates.
[18:03] <genii> I have Intel Pro/1000 XF (fibre optic card) ethtool is reporting only 1000 Base T (copper) capability (when should be 1000 Base FX). Is there some known prob/fix for the e1000 driver?
[18:03] <persia> There's a -backports for features, but almost nobody runs that on servers.
[18:03] <Italian_Plumber> ah well that's good.
[18:03] <Italian_Plumber> I personally stick to the LTS versions for servers.
[18:04] <Italian_Plumber> (LTS comment not related to bugfixes comments)
[18:05] <Italian_Plumber> It seems like there have been very few updates to hardy lately.
[18:05] <persia> Less disturbance to users that way :)
[18:05] <Italian_Plumber> Well I'm my only user, so no worries there. :)
[18:05] <persia> Well, most of the critical bugs are gixed :
[18:05] <geneticx> persia: ah thanks for the clarification, periodical security updates are more likely
[18:06] <persia> geneticx: That's the usual model.  Just update off -security (or -updates if it's a less critical server and your users are complaining), during scheduled maintenance windows.
[18:07] <Italian_Plumber> genetix:  FYI: I blindly run all available updates maybe 1-2 times per month -- I trust that none of them will break my system, and I haven't been burned yet... nor do I expect to be.  I just run "sudo apt-get update && sudo apt-get dist-upgrade && sudo apt-get dist-upgrade" and go get a soda.
[18:08] <zroysch> is it possible to get data from a single drive that was part of a raid5
[18:08] <Italian_Plumber> ... as opposed to upgrading a FreeBSD server, where you need to free up an afternoon and sit in front of the computer reading a book or something.  That is when you're not trying to figure out how to solve this conflict and that dependency, etc.
[18:09] <Italian_Plumber> .. and you need to let the wife know you're going to be late and don't wait up. :)
[18:10] <geneticx> Italian_Plumber: hehe.. the "soda" is the best part of the whole procedure =D . Wish I can blindly update my servers like that but I still have to consider the fact that the room for something to go wrong exists =D
[18:10] <persia> zroysch: Depending on how the raid5 was configured, and how many disks were in it, the changes vary from low to almost nothing for full files, and are always 100% for partial files.
[18:11] <zroysch> persia. there were 3
[18:11] <zroysch> it was md raid5
[18:11] <geneticx> persia: sounds like a plan.
[18:11] <zroysch> i just am curious if i should bother trying to wipe it before sending it back for warranty
[18:11] <persia> zroysch: That means that each disk contains 50% of your data, so the chance of recovery is moderate to high, depending on what you mean by "recovery".
[18:12] <Italian_Plumber> geneticx: you might occasionally be prompeted to let the system know if you want to keep your configuration file or accept the package maintainer's configuration file.  The default is to keep your own.  I always say keep my own.
[18:12] <Italian_Plumber> that is absolutely the worst thing that can happen.
[18:13] <zroysch> i still have old pci board in this ubuntu server. is there a huge difference when drives can go 3gbps ?
[18:16] <geneticx> persia: one last thing, how to I run only security updates with apt-get?
[18:16] <zroysch> wow i've been dd'ing the raid5 partition for a couple hours now. only 55gbytes copied so far. out of 900
[18:17] <geneticx> Italian_Plumber: that makes sense
[18:17] <persia> geneticx: Only enable -security (and not -updates) in /etc/apt/sources.list
[18:18] <kirkland> TeTeT: tomorrow is okay
[18:21] <geneticx> persia: ah ok
[18:21] <geneticx> persia: thanks for your input
[18:22] <geneticx> Italian_Plumber: thanks for your help
[18:24] <Italian_Plumber> no prob
[18:24] <Italian_Plumber> I think persia helped a little more. :)
[18:35] <zroysch> anyone used md raid10?
[18:35] <zroysch> trying to figure out what it means when it says it can start the raid 10 with 2 drives
[18:44] <garymc> Hey Guys... I just sorted my NamedVirtualHosts, but now when I goto mydomain.com/phpmyadmin . Phpmyadmin is no longer there. Anyone know How if ix this?
[18:44] <garymc> Im in the phpmyadmin channel but not many people about
[19:14] <momelod> greetings channel
[19:15] <momelod> in ubuntu 9.04 the drbd module is suppose to be built into the kernel, but when i reboot i dont have a /proc/drbd folder.
[19:15] <momelod> how can i get the module for drbd in 9.04?
[19:37] <franjpr> can i use my network printer attached to my ubuntu server in virtualbox windows xp without samba?
[19:42] <Jeeves_> If you have a route to it, sure you can
[20:01] <Silent> allright so ive got this qeustion
[20:01] <Silent> i setup a proxy with ubuntu
[20:01] <Silent> tho
[20:02] <Silent> evertime i reset squid i get parseConfigFile: line 622 unreconized: 'Http_access allow manager localhost"
[20:05] <Silent> so?...
[20:05] <franjpr> Jeeves_: I have added the network printer
[20:05] <franjpr> print jobs are sent
[20:06] <Jeeves_> franjpr: That should be it than
[20:06] <franjpr> but printer does not complete jobs
[20:07] <Jeeves_> franjpr: The printer itself has an network interface?
[20:08] <franjpr> sending a test page from cups webpage works
[20:08] <Jeeves_> that's not my questino
[20:08] <franjpr> but sending a print job from an application in VM windows xp do not
[20:09] <franjpr> I do not know whether I have a route to it
[20:09] <Silent> Can anyone here please assit me with a proxy issue?
[20:09] <Jeeves_> Silent: change the uppercase H into an h ?
[20:09] <Jeeves_> franjpr: Can you ping it?
[20:09] <Silent> mmm ill try it
[20:10] <franjpr> ping from windows xp to the server you mean?
[20:10] <Jeeves_> franjpr: No, the printer ofcourse
[20:11] <Silent> mmm ok it compiled fine lets see what it does when i try to access threw it
[20:11] <franjpr> I do not know the address of the printer
[20:11] <franjpr> it is attached to the server via usb
[20:12] <franjpr> I am a bit lost, you see
[20:12] <Jeeves_> So it's not a network printer ...
[20:12] <franjpr> ah ok
[20:13] <Silent> ok mmm it compiled fine tho doesnt give me teh pop up authentication :S
[20:13] <franjpr> I have tryed to add it as a network printer
[20:13] <franjpr> and as a local printer creating an lpr port
[20:13] <franjpr> so if it is not a network printer
[20:15] <franjpr> it should work as a local printer through lpr port
[20:15] <franjpr> or it is necessary to configure samba in the server
[20:23] <franjpr> I have connected the printer as a local printer through lpr port
[20:24] <franjpr> it is recognized and print jobs are sent but printer does not complete jobs
[20:29] <markus27> Is the question posted here correct? http://tinyurl.com/yz2cuf7  Does Walrus really store the full files on the cloud controller?  Aren't they distributed throughout the cluster?
[20:48] <VSpike> I'm using dnsmasq on my server, and the local interface does not use dhcp (it's static).  I want dnsmasq to serve its name via dns...
[20:49] <VSpike> afaict this means it should be in /etc/hosts -- thing is, it is already as "squiddy 127.0.1.1" (as well as "localhost 127.0.0.1").  What does that entry do, and can I just change it to 10.0.0.1, the desired address?
[21:22] <neopsyche> hello all.. glad i found this channel.. does anyone here have any experience using ubuntu with squid as a cache server?
[21:22] <neopsyche> (on ISP network for example)
[21:23] <neopsyche> hello anyone?
[21:26] <ChmEarl> pong
[21:27] <marsje> neopsyche: I use squid on debian
[21:27] <neopsyche> cool
[21:27] <marsje> neopsyche: just for me
[21:27] <marsje> neopsyche: so I guess that is not what you're looking for
[21:27] <neopsyche> im wondering what exactly can be done to set up on an isp network for a specific project.. im looking to cache youtube videos through a specific site.. its kind of challenging as i am still trying to get my head around the concept of how it all works..
[21:28] <neopsyche> i am looking to cache only specific youtube videos. as well as videos from other servers.
[21:30] <marsje> not sure that is possible
[21:30] <marsje> since it uses some Flash specific streaming stuff
[21:31] <marsje> I asume squid only understands http
[21:32] <VSpike> yeah, i agree - not sure youtube would be cacheable
[21:32] <marsje> but I've heard of tools storing flash video, so nothing is impossible...
[21:33] <VSpike> I expect you could cache the html 5 version, wouldn't you think?
[21:33] <lifeless> flash content is just files
[21:33] <lifeless> biggest challenge is dealing with the CDN urls
[21:34] <marsje> CDN = content delivery network
[21:34] <marsje> you mean folks like Akamai routing you to random nearby servers?
[21:36] <marsje> neopsyche: you want to save bandwidth?
[21:36] <neopsyche> sure.!
[21:36] <neopsyche> for ISP.
[21:36] <neopsyche> marsje: yes
[21:36] <marsje> can't you setup peering with youtube?
[21:45] <MTecknology> any of you know much about nginx? I'm trying to figure out php-frm
[21:45] <MTecknology> fpm*
[22:02] <bogeyd6> !anyone | MTecknology
[22:04] <MTecknology> bogeyd6: my generic question was there
[22:07] <lamont> MTecknology: specifically what are you trying to figure out about php-frm. not that I have any clue, what with it being php in all likelihood
[22:09] <Sciri> lamont: Can I ask you a question? ;)
[22:09] <lamont> Sciri: go ahead and ask 2/
[22:09] <Sciri> lamont: Oh. Nevermind.
[22:19] <MTecknology> lamont: I want to use nginx; I can't use mod_php like in apache; I installed spawn-fcgi and php-cgi; I need to figure out how to have spawn-fcgi a daemon that will launch php fastcgi
[22:21] <MTecknology> lamont: so far it's looking like the best route will be to compile php with the fpm and fastcgi patches
[22:31] <MTecknology> lamont: any ideas on doing that?
[22:31] <dhaivat> hello, I'm facing some serious stability issues with my setup of 10 dell r610 running ubuntu 9.10 x64, is this the right place to ask questions?
[22:34] <lamont> MTecknology: you're talking about php - I'm totally without clue on php, and on nginx.  OTOH, you have now asked a specific question that someone with knowledge has a chance of answering
[22:35] <MTecknology> lamont: alrighty; it actually took me a while to figure out what it was I even had figured out :P
[22:38] <ceocoder> hello, has anyone used netxen nx3031 on ubuntu server 9.10? I'm facing some overheating  and random reboots issues
[23:39] <David_B> hello there. I am trying to get up FTP so I can upload files from my mac to my ubuntu server. I can access my home directory, but cannot upload files
[23:40] <David_B> I cannot see what could be wrong in the config, is there anything that could be causing this?
[23:41] <David_B> my end goal is to be able to get files to /var/www