[00:40] <VSpike> qman__: I went with imap because i was using a desktop and a laptop, and dual booting on both (win/linux on one and ubuntu/arch on the other) ...
[00:41] <VSpike> qman__: and also because I couldn't decide which email client i liked, so i wanted to have the spam filtering and folder filtering all done server side
[00:41] <VSpike> qman__: it's hard to give up that flexibility once you get used to it
[01:50] <Sorell> does unbuntu server have a "sleep" mode
[01:50] <Sorell> it seems like my server likes to go to sleep
[01:50] <Sorell> and I have to ping it to wake it up.
[01:51] <Sorell> I can't visit the website when this happens or ssh in
[01:51] <Sorell> but if I ping it everything is fine.
[01:51] <Sorell> any ideas?
[02:08] <rberger> Are there examples of creating a profile for pam_auth_update so as to enable pam_limits.so for all users including those that su?
[03:04] <bikcmp> Hello everyone, I'm trying to setup UEC on my box, but, as always, it doesn't quite work as expected.
[03:04] <bikcmp> http://pastebin.com/VzVUJ2pU
[03:06] <bikcmp> Seems dead in here.  I'll try #ubuntu.
[03:35] <djveer> Hey guys, i'm getting errors saying "ata3.01: status: { DRDY ERR }" and "ata3.01: error: { UNC }" when trying to boot my ubuntu server. An attempt to fsck the volume gives me "Buffer I/O error on device sda1, logical block 548742" Error reading block... etc
[03:35] <djveer> This a corrupted file system? or could there be a hardware problem here, and how do I tell
[03:35] <djveer> *?
[03:41] <twb> djveer: your disk is dying
[03:42] <twb> djveer: go stick it in the freezer, wrapped in a towel, for half an hour.  Then take it out and dd the data off it.  Then you can attempt to fsck the dd image instead of the dodgy disk.
[03:43] <djveer> twb: okay cool thanks for the advice... what dd command would I use? dd if=/dev/sda1 of='filename' bs=1M   or something like that
[03:44] <twb> djveer: if you have dd_rescue, that would be easiest.
[03:44] <djveer> twb: okay, i'll look into it. Can it be done with just the standard dd utility though?
[03:44] <djveer> i've never honestly used dd_rescue, so not sure of it's function
[03:53] <twb> dd_rescue is just dd with some bodgy patches to make it slow down or speed up when reads fail
[03:54] <djveer> okay, good to know
[03:55] <twb> i.e. it's designed specifically for your current use case
[03:57] <djveer> twb: great stuff, thanks :)
[04:31] <ubuntu_user_01> my OS hdd is failing and i need to swap it out...i have a raid 6 seperate from that hdd, is it possible to do a fresh install of ubuntu on a new HDD and rebuild or boot the raid 6 post install?
[04:35] <twb> ubuntu_user_01: sorry, you have the OS on an un-raided single disk, and a separate RAID 6 array that's used for something else?
[04:38] <ubuntu_user_01> yess
[04:38] <ubuntu_user_01> sorry for the delay
[04:38] <twb> ubuntu_user_01: is it md raid?
[04:38] <ubuntu_user_01> yes mdadm raid 6
[04:38] <twb> Are you raid6-ing the disks, or partitions on the disks?
[04:39] <ubuntu_user_01> raid 6 the disks
[04:39] <twb> Bummer.
[04:39] <ubuntu_user_01> 8 full hdd in raid 6
[04:39] <ubuntu_user_01> well thats not good hahaha
[04:39] <twb> Otherwise you could easily just put /boot on a second set of raid1 partitions, and move the root partition into the array.
[04:40] <twb> Which is what I'd do for a fresh deployment.
[04:40] <twb> I *think* GRUB2 can understand a /boot stored on LVM on md RAID now, but I wouldn't want to rely on it.
[04:40] <ubuntu_user_01> can i drop another HDD in and make my failing HDD a raid 1 with the new 1, yank out the failing drive and then rebuild the raid 1 with another new hdd
[04:41] <twb> ubuntu_user_01: so you would end up with 10 disks, a raid-1 array of two disks for / and /boot, and the original, unchanged raid6 of eight disks for /srv?
[04:41] <ubuntu_user_01> wait...i think i am raid partitions
[04:41] <ubuntu_user_01> yea
[04:42] <twb> That ought to work
[04:42] <ubuntu_user_01> sweet...now i just got to go break down my raid 0 in the hackintosh till i can get some new hdd
[04:42] <ubuntu_user_01> haha
[04:42] <twb> You'll need to dance with grub during the transition, so make sure you have a live CD ready, preferably one with mdadm and lvm2 installed in it
[04:42] <ubuntu_user_01> alternate cd comes with both correct?
[04:43] <twb> Not in rescue mode, although you can probably anna-install them from the CLI while in rescue mode.
[04:43] <twb> I would normally use a hand-rolled rescue image, or the CentOS 5 live CD, or knoppix.
[04:44] <twb> IMO that's less hassle than trying to use d-i's rescue mode.
[04:45] <ubuntu_user_01> kk thanks for the help
[05:05] <axisys_> twb: hmm.. so jfs fs gets destroyed if it is external array while the internal disk has the OS which went bad?
[05:06] <axisys_> twb: hmm would be nice if jfs behaved like zfs where i can destroy my OS on internal disk and no worry
[05:06] <twb> axisys_: huh?
[05:06] <twb> I don't know anything about JFS.
[05:07] <axisys_> ubuntu_user_01's raid6 was on JFS aparently based on his comment on #ubuntu
[05:07] <twb> axisys_: surely you mean JFS was on RAID6.
[05:08] <axisys_> twb: that was what he said
[05:08] <twb> axisys_: YOU said RAID6 was on JFS, i.e. backwards
[05:08] <axisys_> twb: oops.. yep
[05:11] <twb> I don't see why the filesystem would matter.
[05:11] <twb> He had his OS filesystem separate from his data filesystem
[05:39] <carmony> What command can I run from the terminal to determine what interfaces I have available?
[05:40] <qman__> carmony, I assume you mean ethernet interfaces--ifconfig -a will show all the interfaces with drivers loaded
[05:41] <twb> qman__: ITYM "ip addr"
[05:42] <carmony> qman__: thanks
[05:42] <qman__> twb, I use ifconfig
[05:42] <qman__> old habits die hard
[05:44] <qman__> it's a temporary means anyway, anything permanent should be in /etc/network/interfaces
[06:06] <swift> hi guys, if i start 'vncserver' on my linux box, I should be able to connect to it via vnc right?
[06:06] <swift> it's not working for me
[06:32] <qman__> swift, vncserver requires an X session to be running to connect to
[06:32] <qman__> ubuntu server does not have a GUI
[06:33] <swift> qman_.. i got it goin.. thanks for your input!!
[06:42] <twb> qman__: um, I think vncserver uses a new Xvfb instance
[06:42] <twb> And installing vncserver ought to have pulled in X.
[06:42] <twb> Unless you're talking about x0rfbserver, which exports a normal X instance.
[07:00] <maxagaz> hi
[07:00] <maxagaz> mailq returns the following error for thousands of emails: mail system configuration error
[07:00] <maxagaz> how to debug it ?
[07:09] <twb> Maybe by configuring the mail system correctly?
[07:09] <twb> What MTA are you using?
[07:44] <ruben23> anyone tired vyatta CE router..?
[07:48] <twb> ruben23: does it run Ubuntu Server?
[07:50] <ruben23>  twb:yes
[08:33] <germanjew_> an anyone here please help me with setting up a cloud?
[08:34] <germanjew_> can*
[08:34] <twb> Is there an ubuntu-cloud channel we can direct these people to?
[08:34] <twb> EUC bores me.
[08:35] <germanjew_> if you help me I'll be gone and you won't be bored anymore
[08:36] <twb> germanjew_: I can't; I know nothing about EUC
[08:36] <germanjew_> hehe ok
[08:37] <twb> It's hard enough for me to keep normal hosts running, let alone a big sort of "blob" of them
[08:50] <sherr> germanjew_: https://help.ubuntu.com/community/UEC
[08:51] <sherr> There's also a PDF best practices doc on Intel's site about Ubuntu :
[08:51] <sherr> http://software.intel.com/en-us/articles/intel-cloud-builder-best-practices/
[08:51] <sherr> germanjew_: But I cannot help more either because I don't know much about it (and am not that interested currently)
[08:56] <twb> sherr: ooh
[08:56] <twb> Oh, just best practices for euc
[08:56] <twb> I thought you meant ubuntu in general
[09:01] <sherr> No, cloud - but still a useful doc with some nice diagrams. Maybe I'm missing the nice diagrams, and that's why I don't see the point ..:-)
[09:19] <germanjew_> sherr: thanks alot, but I think I figured out what the problem was
[09:24] <Akos> hi everyone
[10:19] <eagles05138785> hey guys where can i find kernel modules
[10:20] <eagles05138785> i just got a 2nd network cable for my server and its not the same as the first one and it seems like i dont have the module installed
[10:24] <_ruben> network cables dont require drivers
[10:25] <eagles05138785> what i mean though is the nics are 2 different broadcome models
[10:25] <eagles05138785> one is a pci xpress nic
[10:26] <eagles05138785> the other is a normal giga broadcome nic
[10:26] <eagles05138785> and the regualr one isnt showing up with an ip or anything when i do ifconfig
[10:26] <eagles05138785> _ruben: i got it
[10:26] <eagles05138785> had to bring it up
[10:34] <ruben23> hi, how do i install  sources for the 2.6.24-27-rt kernel
[10:43] <sherr> ruben23: apt-get
[10:44] <ruben23>  sherr: apst get..? what..? source  2.6.24-27-rt kernel ..?
[10:45] <eagles05138785> hey guys i want your feedback on something i have a router connected directly to the outside via cable modem and i have a server which has to nics and i was wondering if i set that up as a firewall should i put that directly to the cable modem and then the router or can i leave my setup the way it is
[10:45] <eagles05138785> the router has a simple nat firewall it seems like
[10:49] <sherr> ruben23: There are manuals installed on every system - useful to know how to find out information yourself i.e. man apt-get. See the "source" option.
[11:03] <Guest66551> :
[11:25] <joschi> hi, I'm trying to boot ubuntu over PXE in a vmware ESXi virtual machine. when the vm has a network interface of the types vmxnet2 or vmxnet3 it throws a kernel panic during the boot process. when the network interface is e1000 compatible, it works as expected
[11:25] <joschi> used kernel is 2.6.31-20-server
[11:26] <joschi> anyone has had similar problems?
[12:59] <pths> which package proves samba.schema for LDAP?
[13:02] <sherr> pths: you can search packages and package contents on packages.ubuntu.com
[13:09] <pmatulis> pths: that would be samba-doc
[13:10] <pmatulis> pths: (apt-file search samba.schema)
[13:41] <jalons> I have a failed drive in a 9.04 server raid5 array, when I pass bootdegraded it's still not constructing, however I can assemble it from initramfs (busybox?)
[13:41] <jalons> how can I either A. get bootdegraded (i also tried bootdegraded=true and bootdegraded=1) to work so I can copy data off then try the rebuild, or B. get network going from the initramfs environment so I can copy the data off
[13:46] <sherr> jalons: no console access?
[13:47] <sherr> I guess "bootdegraded" needs sme updates to the system you haven't got.
[13:47] <jalons> sherr: I do have console access
[13:48] <sherr> I thought that a degraded RAID would allow bootup but expects you to "hit enter" (or something)?
[13:48] <jalons> looks like in 8.10 they changed it
[13:49] <sherr> Painful problem that : https://wiki.ubuntu.com/BootDegradedRaid
[13:49] <jalons> i think it actually tries to continue on, but mdadm is throwing an error that it can't create the array
[13:49] <zul> lamont: ping when you get a chance can you have a  look at #542955
[13:49] <sherr> jalons: there's always a liveCD/recovery mode?
[13:50] <jalons> sherr: yeah, waiting on approval to get access to a CDR to burn a recovery cd
[13:50] <jalons> thanks, I'll explore that venue
[13:55] <nijaba> kirkland: hello!  do you know if there is a way in euca to prevent a node from accepting new instances?
[14:03] <diago> how can I revert a fail apt-get install
[14:05] <mealstrom> apt-get install -f
[14:05] <mealstrom> ?
[14:05] <diago> fails with the same errors
[14:09] <jalons> resolve the errors
[14:09] <sherr> diago: maybe pastebin your errors ...
[14:10] <diago> it's a syntax error in one of the packages trying to be installed. I just want to kill the attempt
[14:10] <diago> third party package of course
[14:11] <lamont> -rw-r--r-- root/root   1549844 2010-03-19 15:18 ./usr/lib/libdns.so.64.1.0 <-- zul: looks pretty non-empty to me...
[14:12] <lamont> or is that not what ldconfig.real means by that error?
[14:12] <zul> im not sure
[14:12] <lamont> anyway, afk for a couple hours - breakfast
[14:13] <jalons> diago: so the package is throwing a syntax error, so it won't install?  is it listed in dpkg as partially installed?
[14:13] <jalons> you may be able to dpkg --purge pkg and then apt-get install -f
[14:14] <diago> it is showing as partial but I'm finding out that removing it will totally uninstall open-xchange
[14:14] <diago> great, MONDAY
[14:15] <jalons> ahh dependency hell
[14:15] <jalons> Sorry man, at that point I find a new machine and start over.
[14:16] <diago> lol... not really an option right now. I'll see what #open-xchange has to say. Thanks though.
[15:06] <diago> can I run a virtualbox question by in here. Windows host - ubuntu guest?
[15:14] <soren> diago: Depends.
[15:14] <soren> diago: It's easier to say after you asked the question, I guess.
[15:15] <diago> ok. I'm having trouble using a windows host and ubuntu guest server. The network adapter in the guest can ping the gateway but if I ping an outside address the name is converted by DNS but I receive no packets
[15:19] <_ruben> sounds like a config issue in the host and/or virtualbox, as in: its not nat'ing/bridging/whatevering properly
[15:20] <soren> diago: Yeah, that doesn't sound like an Ubuntu problem, really. It sounds like something in the hypervisor or on the host side, to be honest. I suggest you try some kind of VirtualBox support forum.
[15:21] <diago> ok, I just tried a different adapter (intel server) and it seems to be working... strange
[15:21] <diago> thanks though
[15:25] <zul> ttx: nut ffe uploaded
[15:25] <zul> 2 more to go ;)
[15:26] <diago> well, it works fine when dhcp is used so it must be my config
[15:33] <diago> this is crazy, if the guest receives info from dhcp it works fine, static no dice
[15:35] <_ruben> diago: is the static config similar to the dhcp config?
[15:35] <diago> from what I can tell, the only additional param I am using is network 192.168.0.0
[15:35] <diago> but even if I remove that I still get nothing
[15:36] <diago> well, I can get around internal fine
[15:36] <zul> mathiaz: when you get a chance can have a look at https://bugs.edge.launchpad.net/ubuntu/+source/rabbitmq-server/+bug/506985
[15:55] <MTecknology> I just got my Ubuntu 9.10 CD's :D
[15:55] <MTecknology> 1 desktop, 1 server - /me is excited
[16:53] <smoser> ttx, do you happen to have any insight on https://bugs.launchpad.net/ubuntu/+source/ec2-api-tools/+bug/542395
[16:58] <ttx> smoser: not really. Maybe doko can help you
[17:00] <ttx> smoser: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=573742
[17:01] <smoser> yeah, i just saw that and was going to test the -Djava.net.preferIPv4Stack=true
[17:01] <ttx> smoser: if you confrm it's the same, you should change the package affected and link to the debian bug
[17:01] <smoser> works
[17:02] <ttx> smoser: maybe we can workaround it... but we should keep a bug open against openjdk
[17:03] <RoAk> kirkland, found a way to parse the config files
[17:03] <smoser> ttx, yes.
[17:03] <smoser> thanks.
[17:07] <ttx> kirkland, mathiaz, smoser: I'm about to call it a day -- anything you need from me ?
[17:07] <smoser> not from me.
[17:07] <mathiaz> ttx: not from me.
[17:08] <ttx> jsdelfino: I think we worked together circa 1997.
[17:08] <kirkland> ttx: nope, thanks
[17:08] <kirkland> RoAkSoAx: cool
[17:10] <RoAkSoAx> kirkland, i'll work on it and post all my code for you to take a look at it
[17:10] <RoAkSoAx> kirkland, and I uploaded a branch with the testdriverc change for obtaining the code name: lp:~andreserl/testdrive/codename-change
[17:11]  * ttx disappears
[17:20] <kirkland> RoAkSoAx: yeah, i saw that one; it's not quite what i had in mind; i have a better design in mind; don't have time to detail it right now; but i'll try to get it in tonight
[17:20] <kirkland> RoAkSoAx: shouldn't affect your modularity, though
[17:21] <RoAkSoAx> kirkland, it should since I'm planing to handle that differently, other than that I'll just merge your change sin mine
[17:22] <kirkland> RoAkSoAx: cool
[17:22] <kirkland> RoAkSoAx: basically, i'm going to move the config code that builds the list of ISOs to a different file
[17:22] <kirkland> RoAkSoAx: different config file
[17:23] <kirkland> RoAkSoAx: that only gets sourced if we're in the "Select an ISO" code path
[17:23] <kirkland> RoAkSoAx: which necessarily requires network connectivity
[17:23] <kirkland> RoAkSoAx: I want to avoid requiring network connectivity for those code paths that don't actually need it
[17:23] <kirkland> RoAkSoAx: like "testdrive -u /path/to/local/foo.iso"
[17:25] <RoAkSoAx> kirkland, woulnd't it be more user to keep the menu but show "no network connectivity, please insert path to iso"? instead of showing the whole menu?
[17:28] <RoAkSoAx> kirkland, I have just uploaded my changes to a branch for you to take a look at: lp:~andreserl/testdrive/module
[17:29] <RoAkSoAx> an initial modularization is pretty much done, but It still requires many changes to improve the code basically
[17:29] <RoAkSoAx> and I'll probably merge the launch_kvm launch_virtualbox and launch_parallels functions from the executable to the "module"
[17:30] <RoAkSoAx> kirkland, there's also a "prueba" script that shows how to use the configparser
[17:31] <RoAkSoAx> and I'm creating the ISO list in the "module" instead of creating it in the config
[17:31] <kirkland> RoAkSoAx: i'm really slammed all day today, so i won't be able to get to it until tonight
[17:31] <RoAkSoAx> so that it is used for both, the command line, and the PyGTK interface in the future to create a manu
[17:32] <RoAkSoAx> kirkland, its ok, I'm gonna do some work on it today with the way I was planing to parse the config file, so that you can see all that tonight
[17:32] <kirkland> RoAkSoAx: awesome
[17:38] <mathiaz> kirkland: http://paste.ubuntu.com/399417/
[17:52] <mathiaz> kirkland: http://paste.ubuntu.com/399432/
[18:54] <wack479> when running rsync on my ubuntu 9.10 server kernel 2.6.31-20 when it gets to a large file, it kernel panics and then reboots, anyone else had this problem?
[18:59] <RoyK^> oh
[18:59] <RoyK^> 2GB limit?
[19:01] <soren> wack479: Oh, wow. Which filesystem?
[19:04] <Callum__> wack479: are you copying to or from a FAT32 filesystem? or any other filesystem with relatively small file size limits?
[19:04] <RoyK^> even if was FAT12, a panic is bad news
[19:07] <scar_> my friend wants to build a couple of devices that will link to each other sharing processing power and other resources, I've offered to help him out on the software side. For starters we tried Eucalyptus via the 9.10 server cd inside virtualbox
[19:08] <scar_> we followed both the server guid and community documentation, but still was not able to get it working
[19:12] <sherr> scar_: can't help with the cloud stuff myself, but know you'll need to define how it didn't work in more detail. What exactly didn't work?
[19:12] <scar_> one thing that baffle me is that eucalyptus needs 40gb space for both clustor and nodes
[19:13] <scar_> sherr I had problems building my own images (with vm-builder) and starting the image afterwards
[19:13] <sherr> UEC folks : Good PDF article on Clud Best Practice (with Ubuntu UEC) here :
[19:13] <sherr> http://software.intel.com/en-us/articles/intel-cloud-builder-best-practices/
[19:14] <sherr> (scar_) that's not to help you out specifically
[19:14] <sherr> Clud? Cloud* :-)
[19:15] <scar_> right now I think it's better for me to focus on other possable solutions, since giving each device 40gb hdd space isn't really an option right now
[19:17] <nosse1> Hello. After I upgraded my server to karmic, the ssh daemon lets the users log in, even before fsck is done checking the user's disks. How can I change the order?
[19:17] <sherr> scar_: where does it state those requirements?
[19:19] <scar_> https://help.ubuntu.com/community/UEC/CDInstall
[19:21] <ttx> kirkland: did you notice
[19:21] <ttx> https://lists.ubuntu.com/archives/ubuntu-server/2010-March/003906.html
[19:22] <kirkland> ttx: sorry, no, haven't gotten to it
[19:23] <ttx> kirkland: nothing urgent, just something to keep an eye on, I guess :)
[19:23] <ttx> I'll do a few tests when I'll be back on Friday.
[19:26] <kirkland> ttx: yeah, agreed
[19:27] <ttx> kirkland: I experienced sluggish performance under kvm recentely, and it might have been when I did iso testing over qcow2 images.
[19:39] <smoser> jdstrand, i just opened bug 544435 i believe its a problem with app armor profiles for libvirt
[19:52] <a|3x> hi
[19:52] <a|3x> when i run posix_getpid in php i get pid of apache process, why is that?
[19:58] <joschi> a|3x: probably because you're running mod_php
[20:00] <a|3x> joschi: would it be possible to have apache start different process for each page request?
[20:01] <joschi> sure. run php with mod_cgi or suphp
[20:01] <joschi> but this means a certain performance hit
[20:01] <joschi> normally you want to avoid that
[20:04] <a|3x> thats how its done on centos i believe
[20:05] <a|3x> the problem is i have a long polling app and i would like to prevent clients from long-polling more than once
[20:05] <a|3x> to do that i was planning to use pid checking
[20:05] <a|3x> is there any other way to do it?
[20:09] <joschi> why not just check a cookie or (more error-prone) the ip address before starting your expensive calculations?
[20:09] <a|3x> what if they are behind a nat
[20:10] <joschi> then cookies will still work…
[20:10] <a|3x> i have non-browser client apps i and would really like to keep them simple and not require cookies
[20:10] <a|3x> plus, cookies can be hacked
[20:12] <a|3x> actually, this is more of a licensing issue rather than calculation
[20:12] <joschi> how would knowing the process id of your php process prevent that in any way?
[20:14] <a|3x> lock mysql db lock, get pid in database next to key they provided, see if that process is still alive, if yes, return with key already used error, else save pid, unlock lock, go on
[20:16] <a|3x> i could use the db lock for that but i need it for efficient long-poll
[20:16] <a|3x> i could set flags in the db but what if the process dies and the flag never gets reset
[20:17] <a|3x> how big of a performance hit is it?
[20:17] <joschi> the php binary will have to be loaded for each request
[20:18] <joschi> since it's in the cache after the first run, at least it doesn't have to be read from the disk every time
[20:18] <joschi> a|3x: just try it
[20:19] <a|3x> the question is, after each long-poll event i am going to have everybody requesting it again all at the same time
[20:20] <a|3x> question: how much of a hit it will be when 100 php processes start all at the same time
[20:20] <a|3x> compared to what i have now
[20:20] <a|3x> that is mod_php
[20:21] <joschi> 100 * size of a single php binary
[20:21] <joschi> do your maths ;)
[20:22] <a|3x> how does mod_php work?
[20:23] <a|3x> fork?
[20:23] <a|3x> no wait
[20:27] <joschi> a|3x: the interpreter is embedded in the apache processes. that's why you only see apache's pid. no forking is done on a request.
[20:29] <a|3x> would it be possible to see if a particular php script is still executing within some apache process?
[20:29] <a|3x> unique id of some sort maybe?
[20:54] <RoAkSoAx> kirkland, is it really necessary to load multiple config files in testdrive?
[20:54] <kirkland> RoAkSoAx: yes; talk to hggdh
[20:54] <RoAkSoAx> hggdh, ping
[20:54] <RoAkSoAx> kirkland, ok, because I really don't see the point of having multiple config files :)
[20:55] <hggdh> RoAkSoAx, pong
[20:56] <RoAkSoAx> hggdh, is it really necessary to have multiple config files for testdrive, why?
[20:56] <hggdh> RoAkSoAx, we need to be able to test different scenarios -- for example, no network connectivity
[20:56] <hggdh> or different memory allocations
[20:57] <hggdh> etc
[20:57] <hggdh> RoAkSoAx, you mean the successive loads?
[20:57] <RoAkSoAx> hggdh, yeah
[20:57] <hggdh> RoAkSoAx, they do not hurt -- and allows one to only change what is needed
[20:58] <hggdh> for example, I usually base mysefl on /etc/testdriverc, and only add my own changes
[20:58] <kirkland> RoAkSoAx: basically, we need 2 ... on distro level set of defaults
[20:58] <kirkland> RoAkSoAx: and 1 local one for users to override those
[20:59] <kirkland> RoAkSoAx: various people "complained" about where that local one was located
[20:59] <kirkland> RoAkSoAx: I thought ~/.testdriverc made perfect sense, but then some people wanted in ~/.config/ and other newfangled locations
[21:00] <kirkland> RoAkSoAx: so it's best to just create one process for sourcing a config file, and pass that process an array of locations, in order
[21:00] <RoAkSoAx> kirkland, hggdh, uhmmm ok, because I was thinking on doing something like if it does not exist in ~/.testdriverc use the one in ~/.config/ if not, use the default in /etc/testdriverc
[21:00] <kirkland> RoAkSoAx: that's not how traditional Unix configs work ....
[21:00] <hggdh> RoAkSoAx, I would rather first source the default in /etc/, *then* whatever the user wants
[21:00] <kirkland> RoAkSoAx: the traditional Unix config process is a) first source a global config
[21:00] <kirkland> RoAkSoAx: b) then source local configs
[21:01] <kirkland> RoAkSoAx: please don't go changing this sort of core testdrive functionality
[21:01] <kirkland> RoAkSoAx: that part is there for a reason ;-)
[21:01] <hggdh> heh. kirkland stated it more nicely
[21:01] <RoAkSoAx> kirkland, yeah I just trying to figure out the sourcing of config files since I'm using configparser for that functionality
[21:02] <kirkland> RoAkSoAx: right ... so for that, just concatenate all of them together
[21:02] <kirkland> RoAkSoAx: and then run the conf parser across it whole blob
[21:02] <kirkland> RoAkSoAx: assuming that later variables overwrite earlier ones
[21:02] <hggdh> and catenate them in the correct order ;-)
[21:02] <kirkland> RoAkSoAx: right, what hggdh said ;-)  order matters!
[21:03] <RoAkSoAx> kirkland, i was either thinking on having something like "testdrive-globals" and testdrive-locals and they *had* to be different and nto override, but I guess that we should be able to override in local config a global variable, correct?
[21:03] <hggdh> correct
[21:03] <hggdh> the global gives the user sane defaults
[21:03] <hggdh> the user config can be as messed up as the user wants, but the global *has* to be sane
[21:04] <kirkland> RoAkSoAx: the key being that ***we*** get to change the globals, but ***users*** get to override them with whatever they want
[21:04] <hggdh> ideally, nobody will ever need to change /etc (heh, at least I can hope)
[21:06] <RoAkSoAx> hggdh, kirkland ok so, now a python programming question: is it possible to for example do something like: var = KVM_ARGS, then pass that var to a function, and in the function do something like self.var = "-usb" where self.var should actually be self.KVM_ARGS
[21:07] <RoAkSoAx> so to the function I can pass functio(var,value)
[21:07] <RoAkSoAx> and var is an attribute in a class
[21:07] <RoAkSoAx> set_function(self, var, value)self.var = value
[21:07] <RoAkSoAx> something like that?
[21:07] <hggdh> so you are actually defining a class, and heper methods in it
[21:08] <hggdh> as long as the helper methods are in the class, no problem
[21:11] <hggdh> I was, in fact, thinking of setting all the config in a class, and passing it to the emulator to be used
[21:13] <RoAkSoAx> hggdh, http://paste.ubuntu.com/399537/
[21:14] <RoAkSoAx> hggdh, how can I make that self.var = value, if var is a variable, but should refer to a varibale in the class instead of to the variable var
[21:14] <RoAkSoAx> idk if you understand me :)
[21:16] <hggdh> RoAkSoAx, easier to have all the attributes in the class
[21:16] <hggdh> so that you can just do testdrive.mem=<value>
[21:18] <RoAkSoAx> hggdh, http://pastebin.ubuntu.com/399538/
[21:19] <RoAkSoAx> hggdh, i actually modifying the attributes of the class from the main(), however, I'd rather have a set_values class to be able to set the value to a specified var so that I dont have to do it from main
[21:19] <RoAkSoAx> because IIRC, it is better to modify the attributes of a class from class functions and not from main, correct?
[21:22] <hggdh> RoAkSoAx, opinions vary there ;-) I personally do not like a generic setter like this, makes following the code harder to understand
[21:23] <hggdh> if this was an extremely complex programme, then (probably) yes; but you are already using configparser...
[21:23] <hggdh> and, BTW, it is better -- following you example -- to do if x: ... elif y: .... elif z: ...
[21:24] <hggdh> may be ugly, but such is life
[21:24] <RoAkSoAx> hggdh, haha indeed, but it would be better that way in my opinion, that's why Im wondering if theres a way to do that
[21:24] <hggdh> and yes, whenever possible, modify class attributes via helper methods
[21:25] <hggdh> most probably there *is* a way. Not sure it is worth your time (right now)
[21:25] <hggdh> also -- in the helper methods you would probably add syntax and semantic, and whatever, checks
[21:26] <RoAkSoAx> hggdh, ok then, since this is just first modularization i guess I will stick to if if if :)
[21:27] <hggdh> for example, the memory size is expected to be an integer, greater than zero and less than <memory available>, expressed in Megas
[21:27] <hggdh> so 384 == 384Mbytes
[21:28] <hggdh> (of course, this is a simple example, but you could have more stringent checks for, say, URLs)
[22:03] <RoAkSoAx> hggdh, kirkland ok. The config loading is done and should work with various configs. the only thing would be that each user will have to add [testdrive] in the config file
[22:15] <RoAkSoAx> kirkland, hggdh_ lp:~andreserl/testdrive/module, just pushed my latest changes
[22:19] <a|3x> how would i make apache run php scripts with mod_cgi?
[22:41] <kirkland-suds> mathiaz: smoser: yeah ...  well, 27kB/s ... no better :-(
[22:41] <smoser> bah
[22:41] <kirkland-suds> smoser: mathiaz: smells nice in here though
[22:42] <kirkland-suds> smoser: ask nurmi if Goleta's uplink is a carrier pigeon
[22:42] <Burgundavia> kirkland-suds, ping re: byobu
[22:44] <kirkland-suds> Burgundavia: what's up?
[22:44] <kirkland-suds> smoser: actually, i just spiked to 250kB/s
[22:44] <smoser> woohoo
[22:44] <mod> hey all
[22:45] <Burgundavia> kirkland-suds, ripping out my hair for last minute Ubuntu book stuff. Is byobu seeded by default anywhere?
[22:45] <mod> is it generally true that if I am running RAID 1 with two drives, I should be able to pull one drive and mount it stand alone?
[22:45] <kirkland-suds> smoser: can you see ESSID sbsoap ?
[22:45] <mod> not matter the RAID 1 is hardware or software or which ever brand?
[22:45] <smoser> no. i've never seen it.
[22:46] <kirkland-suds> smoser: okay, i'm rocking now
[22:46] <Burgundavia> kirkland-suds, specifically, it is seeded on the Ubuntu server/desktop/netbook cds? my browsing says no
[22:46] <kirkland-suds> Burgundavia: it's a recommends of screen
[22:46] <kirkland-suds> Burgundavia: actually, it is, as of Karmic
[22:47] <kirkland-suds> Burgundavia: fire up a Lucid desktop ISO (by Testdrive, perhaps, and you should be able to open a terminal on the LiveCD and just run 'byobu')
[22:48] <Burgundavia> kirkland-suds, no kvm on this laptop. and now I feel stupid. I saw that dependency chain and for some reason it didn't connect in my mind
[22:48] <Burgundavia> thanks
[22:48] <kirkland-suds> Burgundavia: Testdrive works with VirtualBox too ;-)  but yeah, it's seeded
[22:48] <kirkland-suds> Burgundavia: glad to hear byobu's earned a mention in your book
[22:48] <kirkland-suds> Burgundavia: what book is it?
[22:49] <kirkland-suds> mathiaz: screaming through the langpacks now
[22:49] <Burgundavia> kirkland-suds, official ubuntu book, the command line chapter
[22:49] <kirkland-suds> Burgundavia: cool ;-)
[22:50] <kirkland-suds> Burgundavia: poke me if you want a byobu deep dive some time; happy to show you the advanced features if you like
[22:51] <Burgundavia> kirkland-suds, for the next edition of the server book, we should probably do something, but that isn't my baby.
[22:51] <Burgundavia> but thanks, yes
[22:51] <kirkland-suds> Burgundavia: sure thing; is that Kyle Rankin's book?
[22:51] <Burgundavia> yep
[22:52] <kirkland-suds> Burgundavia: cool; i have an email from him that i need to respond to
[22:52] <kirkland-suds> Burgundavia: thanks
[22:55] <Burgundavia> kirkland-suds, for a quick sanity check, do you have time for a 10 minute look at what I have written?
[22:56] <hggdh> RoAkSoAx, you are mixing tabs and spaces, not a good idea
[22:57] <kirkland-suds> Burgundavia: sure
[22:57] <kirkland-suds> Burgundavia: wanna email it to dustin dot kirkland at gmail dot com ?
[22:59] <Burgundavia> doing so now
[23:03] <Burgundavia> kirkland-suds, did you get it?
[23:04] <kirkland-suds> Burgundavia: got it
[23:04] <kirkland-suds> Burgundavia: where's the track changes option?
[23:07] <Burgundavia> edit-->changes-->record
[23:07] <Burgundavia> and edit-->changes-->show
[23:07] <kirkland-suds> Burgundavia: the only suggestion I have really, would be to mention F6 (detach) and byobu -x (re-attach), and F7 scrollback
[23:07] <kirkland-suds> Burgundavia: otherwise, it's a nice concise introduction
[23:08] <Burgundavia> thanks
[23:09] <kirkland-suds> mathiaz: smoser: lang packs complete
[23:09] <kirkland-suds> on to kernels :-/
[23:09] <kirkland-suds> smoser: mathiaz: bandwidth is great though; should work from here tomorrow
[23:09] <kirkland-suds>     21200896  68%  683.60kB/s    0:00:14
[23:10] <mathiaz> kirkland-suds: is there someone around?
[23:10] <kirkland-suds> mathiaz: there's a few people in here, but there's a cafe with tables and chairs just outside
[23:11] <kirkland-suds> smoser: mathiaz: do either of you have a laptop cable lock?
[23:12] <smoser> nope
[23:12] <mathiaz> kirkland-suds: nope
[23:12] <kirkland-suds> smoser: mathiaz: bummer
[23:12] <mathiaz> kirkland-suds: how long is this going to take?
[23:12] <kirkland-suds> mathiaz: it's about 20x faster, but it'll still be a while
[23:13] <kirkland-suds> mathiaz: i'm just on pool/main/l/linux/linux-image-2.6.32-17-preempt_2.6.32-17.26_amd64.deb
[23:13] <kirkland-suds> mathiaz: gotta make it through the rest of main, then universe
[23:14] <kirkland-suds> mathiaz: smoser: http://pastebin.ubuntu.com/399589/
[23:19] <a|3x> i am having difficulties setting up suphp, anybody care to help?
[23:19] <a|3x> Invalid command 'suPHP_UserGroup', perhaps misspelled or defined by a module not included in the server configuration
[23:20] <kirkland-suds> smoser: mathiaz: made it through the kernels
[23:23] <RoAkSoAx> hggdh, tou mean spaces for def and tabs for everything else?
[23:25] <hggdh> RoAkSoAx, no, actually one should use *either* spaces *or* tabs in a Python source, but not both (PEP 8 highly recommends spaces)
[23:25] <hggdh> in my vi edit of the source, everything seems to be on the same column (because my vi is configured to \t == 4 spaces)
[23:26] <hggdh> RoAkSoAx, bye either/or, I forgot to add -- but not both
[23:26] <hggdh> my, my, am I writing bad today or what?
[23:27] <kirkland-suds> hggdh: RoAkSoAx: but I *strongly* prefer tabs and tabs only!
[23:28] <hggdh> RoAkSoAx, the master's voice ;-)
[23:28] <kirkland-suds> two reasons: a) 1 tab = 1 byte, b) anyone can configure their editor to represent a tab as as much (or as little) whitespace as they like
[23:29] <maxb> a) is largely irrelevant in the face of disks, bandwidth and compression technology of today
[23:30] <hggdh> kirkland-suds, as long as it is either/or (but not both) Python feels good...
[23:30] <RoAkSoAx> kirkland, hggdh ok, better yet, I love using tabs for coding
[23:30] <maxb> b) is a noble goal which starts to fall apart once questions of where the right margin is, and how long statements should be wrapped, enter the picture
[23:31] <RoAkSoAx> hggdh, kirkland-suds NAyways, just ping if somethings is need, Call of Duty MW2 awaits me :)
[23:31] <RoAkSoAx> later
[23:31] <maxb> But really, so long as no-one mixes both in a single file, that's a good start
[23:32] <hggdh> maxb, yes indeed, and a lot of hard-core old coders (see coreutils, for example) are moving out of \t
[23:33]  * hggdh does not care, as long as it is only one of them
[23:33] <hggdh> ok. Time to get the grill
[23:37] <kirkland-suds> hggdh: agreed
[23:38] <kirkland-suds> maxb: understood; (a) is not about bytes-on-disk, it's about left-and-right-arrow to me ;-)
[23:39] <kirkland-suds> smoser: mathiaz: openjdk is figgin huge
[23:39] <a|3x> anybody care to help me figure out why my suphp setup is not working?
[23:42] <kirkland-suds> smoser: mathiaz: oh, FCOL ... openoffice :-/
[23:52] <ruben23> hi