[00:09] <DormantOden> Jumbers: sorry, got called away. I've tried using: route add 192.168.1.99 dev ppp0 . That along the right lines?
[00:13] <SpamapS> jurgen: what client OS?
[00:15] <jurgen> xubuntu
[00:15] <jurgen> runs headless
[00:15] <jurgen> oh client is windows 7
[00:17] <jumbers> DormantOden: You'd need to do route add default I believe
[00:17] <SpamapS> jurgen: I asked because I used to have jitter on linux clients.. but winxp and later seem to do ok in that regard..
[00:18] <SpamapS> jurgen: you may want to install sysstat.. which will give you some general stats every 10 minutes ... you can also run 'vmstat 5' to get a 5 second picture of blocks in / blocks out if you think the disks are really busy.
[00:19] <SpamapS> DormantOden: route add 192.168.1.99 dev ppp0 will just make a rule to send traffic bound for 192.168.1.99 down device ppp0
[00:19] <DormantOden> ah, well, its half right =P
[00:22] <DormantOden> SpamapS: So, route add default gw 192.168.1.99 ppp0 would do it?
[00:22] <twb> SpamapS: 192.168.1.99/32, yes
[00:23] <jurgen> spamams: can i check how many MB's are sent to a local network IP with sysstat?
[00:23] <DormantOden> ah, kk
[00:23] <DormantOden> twb: that meant for me?
[00:23] <twb> Sorry, yes.
[00:23] <DormantOden> awsome
[00:23] <twb> Although "ip route add 0/0 via 192.168.1.99 dev ppp0" would be more modern
[00:24] <DormantOden> hmm, ok
[00:24] <DormantOden> twb: and that would work for both incoming and outgoing too?
[00:24] <twb> There's no such thing as ingress routing.
[00:25] <twb> Packets arrive; then you do something with them
[00:26] <SpamapS> jurgen: no, sysstat just records system stats available from the kernel for the most part. To record traffic to an IP you'd need to use iptables or something else.
[00:26] <DormantOden> twb: When you say "do somthing with them" I can just leave it too sort itself out right?
[00:27] <SpamapS> twb: iptables has a PREROUTING table that might disagree with that. ;)
[00:35] <twb> DormantOden: that's the (egress) routing.
[00:56] <DormantOden> twb: Im getting "SIOCADDRT: No Such Process" When ever I try to add a route
[00:57] <DormantOden> and: "192.168.1.99/32: Unknown host" whenever I try: route add default gw 192.168.1.99/32 ppp0
[00:58] <DormantOden> oh im so close to victory xD
[01:42] <maruq> hi guys
[01:42] <maruq> experiencing high load issues under 10.04
[01:43] <maruq> getting the " System information disabled due to load higher than 1" message
[01:43] <maruq> yet nothing really running in top
[01:43] <maruq> this is running on EC2
[01:48] <erichammond> maruq: What AMI id are you running? in what region?
[01:50] <maruq> erichammond: us-east-1 a&c, running my own AMI - based on ami-6c06f305
[01:50] <maruq> all I've really done is install Chef on there
[01:51] <maruq> chef-0.9.8 installed via apt
[02:12] <MTecknology> How could I script the installation of mysql-server ?
[02:13] <MTecknology> I'm not so sure because it wants me to input a root password
[02:13] <erichammond> maruq: I see flaccid is answering your question on #ubuntu-cloud.  That's basically where I was headed before I got sidetracked here.
[02:14] <maruq> erichammond: yeah, I figured I'd post to both - wasn't sure if it was a cloud or server issue. thanks ;p
[02:19] <MTecknology> There has to be a way to break it away from whiptail so I can script the password it uses...
[02:23] <Cheri703> Hello
[02:24] <Cheri703> so I installed (with great effort) server 10.04 on an old Dell Poweredge 2450, according to the installer it was "successful" but now it won't boot.
[02:25] <pmatulis> MTecknology: yes, that's possible
[02:54] <IdleOne> Cheri703: provide a little more detail as to what you are seeing
[02:54] <IdleOne> "doesn't boot" is very vroad
[02:54] <IdleOne> broad
[02:55] <IdleOne> besides that I can't help much more :/
[02:56] <Cheri703> well, when it was booting into a non-activated windows server installation, goes through the bios screen, then shows an scsi check, then would boot into the OS. I am probably not using the right terms. Now, after the "successful" installation of 10.04 server, it stops after the scsi check
[02:56] <Cheri703> I don't really have any more detail than that :(
[03:08] <detrix42> hi I am trying to set up bind9 on a local network.  I am fairly new to setting up these things.  I just need to know what I am missing, in my config files to get this to work?
[03:15] <twb> I can't see your config files.
[03:16] <detrix42> I was waiting to see if anyone was interested in helping....I was planning on using pastebin...
[03:17]  * jauntypaperclip ubuntu-offtopic
[03:17] <detrix42> named.conf:
[03:17] <detrix42> http://www.ubuntu.pastebin.com/m2itiw3P
[03:18] <Pilif12p> sense no one will answer me in #ubuntu or #perl, and this is a server, what's the best way to update Perl? Do i have to remove the old version?
[03:18] <twb> jauntypaperclip: I think configuring bind is probably on-topic for #ubuntu-server (assuming he's configuring it on an Ubuntu host).
[03:18] <twb> Pilif12p: you should run whatever version apt provides you with.
[03:18] <jauntypaperclip> twb: I was trying to join off-topic and had a typo.
[03:19] <twb> jauntypaperclip: OK :-)
[03:19] <detrix42> named.local.conf:  http://www.ubuntu.pastebin.com/vmfcMskY
[03:19] <jauntypaperclip> sorry about that! stupid meat paws. :)
[03:20] <qman__> detrix42, that should be named.conf.local
[03:20] <twb> detrix42: presumably you want it to act as a recursive resolver?  Rather than hosting your own A records, that is.
[03:20] <detrix42> the name is correct.  just me being dyslectic
[03:20] <qman__> ok, just making sure
[03:22] <detrix42> named.conf.options:  http://www.ubuntu.pastebin.com/VM4t9QG3
[03:24] <detrix42> twb: like I said, I am fairly new to this.  I think that is what I want
[03:24] <detrix42> what other ones do you need
[03:27] <detrix42> this is what I get from dig:  http://www.ubuntu.pastebin.com/Dids6u7e
[03:27] <twb> detrix42: unfortunately I use dnsmasq, so I can't help much
[03:28] <detrix42> oh well
[03:29] <twb> If you hang around someone else might
[03:30] <detrix42> unfortunately, I have been up since 6:00am (EST),  I am getting tired....
[03:33] <detrix42> well I gotta go.  ;)
[05:06] <eroick> Hey, whenever I boot up a fresh 10.04 server install, I don't automatically get a login prompt. Instead, i get a screen which just has the message "Starting NTP server ntpd [ OK ]".  If I hit CTRL-ALT-F1 I get back to all the kernel messages and tty1 with a login prompt. Any ideas how I can start right into tty1 without switching? Also, what exactly am I being switched to/
[05:43] <eroick> I've found that i'm automatically being switched to tty7 (CTRL-ALT-F7... i don't have tty7 configured) when I boot up. How can I change this?
[05:44] <twb> eroick: this is 10.04?
[05:44] <eroick> twb: yep
[05:44] <twb> That's a known "feature" of plymouth
[05:45] <twb> Booting without "splash" in the boot options doesn't actually help IIRC
[05:46] <eroick> twb: can i disable the bootsplash somehow? i see all of the kernel messages up to this point, then it switches to a mystery tty
[05:46] <twb> afaict, you can't
[05:46] <eroick> hmm
[05:46] <eroick> this becomes a problem because I can't actually send CTRL-ALT-** to the machien
[05:47] <eroick> *machine
[05:47] <twb> i.e. if you boot without splash, plymouth STILL RUNS, but in its own window it displays the console text instead of a spinner
[05:47] <eroick> downgrade to 9.10/
[05:47] <twb> eroick: you could write an upstart task that runs "chvt 1" when plymouth finishes
[05:49] <eroick> how would i get it to execute right after plymouth?
[05:50] <twb> I assume plmouth generates an upstart event or something
[05:50] <twb> Alternatively, you could just run a getty on tty7
[05:52] <eroick> true
[05:52] <eroick> now i have another issue :P. when i ssh into it i can connect and authenticate alright but i get no shell
[05:52] <eroick> and no errors from ssh.. i've tried -vvv
[05:53] <eroick> and i've tried setting my default shell to various things (bash, dash, zsh)
[05:56] <eroick> twb: w even shows that the users are logged in and running a shell
[07:21] <neerq> When I install apache2, and than try to start it, it says there is no MPM package isntalled. So i tried to install apache2-mpm-prefork, but it still does not start
[08:05] <neerq> Omfg.. wtf is happening with my server. It looks like apt-get is messed up or something.. When i do: 'apt-get install php5' on my second server, it installed apache2... but when i do it on my production server, it complains about all sorts of dependencies who aren't satisfied
[08:06] <neerq> it says apache2.2-common isn't configured
[08:07] <neerq> and it says apache2-mpm-prefork is unconfigured
[08:09] <twb> neerq: pastebin the transcript
[08:10] <neerq> twb: sure, one second
[08:11] <neerq> oh no
[08:11] <neerq> most of the errors are in dutch
[08:11] <twb> LC_ALL=C, then reproduce
[08:11] <twb> export LC_ALL=C, that is
[08:11] <neerq> oh
[08:11] <neerq> okay
[08:12] <neerq> i'll try it agai
[08:12] <neerq> n
[08:12] <neerq> i am now trying to do aptitude reinstall php5, which gives me the same errors as when i tried to install it at first
[08:13] <twb> Uh, so "it complains about all sorts of depencies"... you just ignored them and installed php5 anyway?
[08:14] <neerq> well, that's because i have exactly the same setup at home.. and when I installed php5 here, it worked without these errors
[08:14] <neerq> twb: http://pastebin.com/ArVxfkpJ this is the output
[08:15] <twb> "The following partially installed packages [...]"
[08:16] <twb> That means that the last attempt to install those packages failed due to something going wrong in their post-install scripts, or due to you manually interrupting it.
[08:16] <neerq> yeah.. that's true
[08:16] <neerq> because it sayed something like: 'no mpm package installed'
[08:16] <twb> Which is probably the same as "ERROR: Module reqtimeout does not exist!"
[08:16] <neerq> oh.. i saw that one too.. yes
[08:17] <twb> You can ignore the mpm warning for now, that's happening because of the previous error
[08:17] <twb> You need to work out what the reqtimeout thing is about
[08:17] <twb> Maybe you're referring to that package in your apache config?
[08:17] <neerq> but what is strange.. when i do purge, or apt-get remove --purge, it looks like there still is some apache2 files left behind
[08:18] <neerq> which package? you mean mpm package, or reqtimeout?
[08:19] <twb> apache2.2-common is what's failing
[08:19] <twb> Try looking through /var/lib/dpkg/info/apache2.2-common.{pre,post}{rm,inst} for references to reqtimeout.
[08:19] <twb> Also grep -r /etc -e reqtimeout
[08:20] <neerq> i tried to install these packages manually too.. like this: apt-get install apache2 apache2.2-common libapache2-mod-php5 etc..
[08:20] <neerq> okay, i'll do that right now
[08:20] <twb> neerq: don't try to install packages manually
[08:20] <twb> You will make matters worse
[08:20] <neerq> yeah.. i was kind of out of options..
[08:21] <twb> Apart from "understanding the problem"
[08:21] <neerq> :P that's true
[08:21] <neerq> when I do grep -r /etc -e reqtimeout, i get 6 errors which say 'no such file or directory'
[08:22] <neerq> grep: /etc/blkid.tab: No such file or directory
[08:22] <neerq> for example
[08:22] <twb> neerq: ignore those
[08:23] <neerq> okay, that's all i get
[08:23] <neerq> i am now looking in /var/lib/dpkg/info/apache2.2-common.*
[08:23] <neerq> twb: this file (/var/lib/dpkg/info# vi apache2.2-common.postinst) has a reference to reqtimeout
[08:24] <neerq> a2enmod -q reqtimeout
[08:24] <neerq> with if dpkg --compare-versions "$2" lt 2.2.15-1~0; then above it
[08:24] <twb> So it's trying to enable a module that doesn't exist
[08:24] <neerq> twb: it looks like it
[08:24] <twb> So it's trying to enable a module that doesn't exist when it's upgrading (from?) a version less than 2.2.15
[08:25] <neerq> i had apache installed previously, from source
[08:25] <twb> Well, that was fucking stupid
[08:25] <neerq> really? why
[08:26] <neerq> btw, i had the same apache from source on my local server
[08:26] <twb> Because the version that apache ships doesn't understand Debian systems
[08:26] <twb> Er, s/debian/ubuntu/
[08:26] <twb> And likewise, Ubuntu infrastructure won't "see" that apache.
[08:26] <neerq> but why would it complain about it then? because, if it doesn't see it, that if-statement would be false.. right?
[08:27] <neerq> the installation from source was version 2.2.15
[08:27] <twb> Maybe one of the tests in that script looks for a program, and the program is there but (as far as dpkg knows) the apache package isn't -- so it gets really confused.
[08:27] <neerq> okay... that makes sense
[08:27] <twb> That's just one example of a whole slew of things that could go wrong
[08:28] <neerq> so what should i do now?
[08:28] <twb> The safest thing would be to backup your data and reinstall the whole OS from scratch, and this time, stick to packaged software in the ubuntu archive.
[08:29] <twb> If you're lucky/clever/careful, you might be able to clean up the messy bits and reinstall apache and friends without needing to do a full reinstall
[08:29] <neerq> so i would have to remove every old file from the old apache
[08:30] <twb> Ideally, yes
[08:30] <neerq> because, the first option would mean a 2 hour drive to the server.. which I don't like so much :P
[08:30] <twb> neerq: well, consider that the cost of this lesson :P
[08:30] <neerq> yeah
[08:31] <neerq> any way to find out which file the package's tests are detecting?
[08:31] <twb> Of course, there are ways to do remote installs...
[08:31] <neerq> really?
[08:31] <twb> neerq: not without deep knowledge of the specific package's packaging (in this case, apache)
[08:31] <neerq> hmm
[08:31] <neerq> okay
[08:32] <neerq> this sucks..
[08:33] <twb> If it's any consolation, it's a pretty common cock-up for for new admins to make
[08:34] <neerq> i think i'll have a look at the remote installation
[08:37] <neerq> twb: are you sure it's possible? I have upgraded ubuntu remotely before.. but the only thing about installation i can find are questions
[08:39] <twb> That's because Ubuntu users are idiots.  The documentation is mostly in the appendices of the installation-guide-amd64 package.
[08:39] <Daviey> twb: erm...
[08:40] <Daviey> twb: Not sure it's cool to refer to Ubuntu users as idiots.
[08:40] <twb> Assuming you have remote TCP/IP but no remote KVM, what you need to do is bootstrap the netboot d-i kernel and ramdisk with the "auto" option (and a few others), which will cause d-i to guess its way to the point where you can ssh into it and complete the install.
[08:42] <Daviey> neerq: If you haven't done it before... preseeding, and my favoured way is dropping it in a re-rolled initrd.  It *will* take you more than 2 hours to do this, but valuable learning exercise for 'next time'.
[08:43] <neerq> Twb, okay, thanks.. i'll try to find some documentation about it..
[08:43] <neerq> Daviey: to drop what in a re-rolled initrd?
[08:53] <twb> You can also provide the preseed file via HTTP, and *FINALLY* via TFTP
[08:53] <Daviey> neerq: As twb points out, you 'pre-answer' some or all of the questions the ubuntu server asks (called d-i)
[08:53] <twb> (Meaning that if you PXE, you can put it on your TFTP server right next to d-i itself.)
[08:53] <Daviey> twb: Well yes, http works.. but you still need to preseed the network configs (unless you have dhcp)
[08:53] <twb> Nod.
[08:53] <twb> If I'm doing bulk preseeding I'm usually in a DHCP-ful environment.
[08:53] <twb> Er, s/bulk preseeding/bulk (net)installs/
[08:53] <Daviey> yeah... would be foolish not to :)
[08:53] <twb> Remind me, how do you forge an proxied HTTP request?
[08:53] <twb> nc 203.7.155.7 8080 <<<$'GET / HTTP/1.0\nHost: example.net\n\n'
[08:53] <twb> ...isn't working.
[08:53] <twb> Ah, I need the full URL in the GET.  Thanks, curl -v.
[08:55] <_ruben> twb: how is that a forged proxied request? :)
[08:56] <twb> Well, only inasmuch as I forged it myself
[08:56] <twb> out of bits
[08:58] <_ruben> i was wondering about both the forged as well as the proxied part ;)
[08:58] <_ruben> seemed like a plain request to me
[09:00] <neerq> twb: okay.. i have removed every single apache2 related file i could find
[09:00] <neerq> is there anything i should do, before using apt-get install apache2 again?
[09:00] <_ruben> could try 'apt-get install -f' first
[09:01] <neerq> and should i do 'apt-get install php5' (because that installed apache2 on my server at home)?
[09:01] <_ruben> it'll try to fix any brokenness
[09:01] <neerq> shouldn't i clean/purge/remove some things first?
[09:03] <_ruben> depending on how much things are messed up by now, that might or might not be needed
[09:05] <neerq> omfg
[09:05] <neerq> okay
[09:05] <neerq> it *looks* like it is working now
[09:05] <neerq> apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName :D
[09:36] <alvin> When is it time to create your own apt mirror, and how much disk space would that require? I have here 28 Ubuntu computers. Our ISP closed down their own server, and now I'm pulling updates from another server.
[09:36] <Jeeves_> alvin: apt-get install apt-cacher-ng ?
[09:37] <Jeeves_>  A mirror consumes a few hundred gigabytes of storage
[09:38] <cemc> install apt-cacher-ng and raise the expire time so packages don't expire that quickly
[09:38] <alvin> Ah, there's an expire time
[09:39] <cemc> well, it's a cacher really
[09:39] <alvin> I'll have to think about it. These are two very different methods.
[09:40] <yann2> Jeeves_, thats not true, I mirror main, universe and multiverse for 3 versions of Ubuntu in less than 100GB.
[09:40] <yann2> of course you don't want to mirror universe.
[09:41] <yann2> (using apt-mirror here)
[09:41] <cemc> yann2: I guess the point is that you are using up space and bw for a lot of packages you'll never need. just sitting there
[09:41] <alvin> True, but space and bandwidth is something I have
[09:42] <yann2> cemc, main isnt that big
[09:42] <alvin> and I guess pulling from the country mirror with 28 machines at once isn't very polite. Hence the question.
[09:42] <Jeeves_> alvin: I wouldn't worry about 28 machines
[09:43] <Jeeves_> .nl has about 35000 machines pulling packages
[09:43] <Jeeves_> they're ok too
[09:43] <alvin> :-) nice to know
[09:43] <Jeeves_> Pulling the complete archive is much more expensive
[09:43] <yann2> depends, its also about speed, much nicer to get packages locally
[09:44] <yann2> plus if you dont have too much bandwidth like me you can sync during the night :)
[09:44] <cemc> alvin: that's why they called mirrors ;)
[09:45] <alvin> True. We have no download limit, but it's not a very speedy connection.
[09:45] <Jeeves_> alvin: I'd install a cacher
[09:46] <alvin> That'll probably be the solution.
[09:46] <alvin> Thanks for all the answers
[09:46] <cemc> alvin: try apt-cacher first, and you can always set up a local mirror later if you really need it. but if you have like 28 production servers which you occasionally apt-get update, there's no point for a whole local mirror (imho)
[10:02] <neerq> I want to install 'apache2-prefork-dev', but that also installs mysql-common, libmysqlclient-dev and libmysqlclient16
[10:03] <neerq> but i don't want mysql to be installed
[10:03] <neerq> is there a way to do this without mysql?
[10:04] <neerq> because i need apxs2.. i don't get why it needs mysql
[10:35] <tgywa> HI
[10:35] <tgywa> why is that this thing is not working ...# tar -zxvf php-5.2.14.tar.gz ... ?
[10:36] <_ruben> tgywa: because it doesn't have a job
[10:36] <tgywa> _ruben, what do u mean?
[10:37] <tgywa> what should I do to extract ... tar.gz ... files?
[10:37] <_ruben> tgywa: "is not working" is way too vague, when a person "is not working", it usualy means that person doesn't have a job
[10:41] <alvin> (or lazy) I see nothing wrong with that command
[10:43] <neerq> I got another problem with apache
[10:43] <neerq> when i do /etc/init.d/apache2 start, i get this error: (98)Address already in use: make_sock: could not bind to address 0.0.0.0:80
[10:43] <_ruben> neerq: either apache is already running, or some other process occupies port 80
[10:44] <neerq> So i searched the error.. and people suggest to look with netstat but there is no process using port 80
[10:44] <neerq> also apache isn't running
[10:45] <Jeeves_> neerq: 'ss -n | grep :80' ?
[10:45] <Jeeves_> What does it say?
[10:45] <neerq> nothing
[10:45] <neerq> no result
[10:45] <Jeeves_> that's odd
[10:46] <neerq> by the way, it also gives me this error: no listening sockets available, shutting down
[10:47] <neerq> any other ideas? :(
[10:52] <neerq> i hate this so much... when i am the only one on the entire internet with this problem.. :P
[10:54] <_ruben> perhaps the error is wrong, and you're not starting it as root?
[10:55] <neerq> i am starting it as root
[11:08] <neerq> _ruben, it works.. i had multiple 'Listen' commands in my configuration, because i copied the old configuration file to the new apache, which also had a Listen:80 line :)
[11:08] <\sh> whoever is working with amazon ec2 cloud services...which python lib for remote management is available for ubuntu and which python lib should I look at, when there is no such lib in ubuntu?
[11:23] <jmazaredo> when i add user using the command line where does the system gets its default layout for creating users?
[11:24] <cfairles> \sh, doesn't amazon provide python libs?
[11:25] <\sh> cfairles: nope..but there are third party libs, e.g. boto
[11:25] <cfairles> \sh, hm, yeah, they only provide .net, java and php only. but boto sounds familiar
[11:25] <\sh> http://boto.s3.amazonaws.com/
[11:25] <cfairles> \sh, i believe thats what the eng. team in a previous life used. i do recall some issues with it however
[11:26] <\sh> well, eventually I need to code it by myself...need it for my (DC)² project.
[12:05] <SpamapS> \sh: you might also try #ubuntu-cloud
[12:06] <\sh> SpamapS: thx for the info...:)
[12:09] <SpamapS> \sh: there's libcloud...
[12:09] <SpamapS> \sh: but thats sort of limited
[12:09] <SpamapS> \sh: what exactly are you wanting to do?
[12:11] <tgywa> How can I make a custom PHP 5.2. package on Ubuntu 10.04 ?
[12:11] <\sh> SpamapS: I need a python interface lib towards amazone EC2 / S3 services, to integrate into my project, to provision amazon and/or UEC VM instances easily
[12:18] <SpamapS> tgywa: simplest way would probably be to download the source package from karmic or hardy, one of which would be 5.2, and then build them on a lucid box and install. Downgrades can be tricky though, so I'd recommend purging 5.3 if you already have it installed...
[12:19] <SpamapS> tgywa: if you want to have 5.2 installed in parallel with 5.3 .. you'll need to put it somewhere other than /usr ...
[12:20] <SpamapS> \sh: uec-run-instances from the latest cloud-utils is pretty handy.. not a lib, but a cmdline tool that will do things like wait for ssh to be up, and even download the ssh keys from the console output and put them in your known_hosts file.
[12:20] <tgywa> SpamapS, currenlty ... I am running both ... 5.2 and 5.3 ... using fcgid
[12:20] <SpamapS> tgywa: but, 5.2 is installed in /usr/local or something?
[12:20] <tgywa> so now, I would like to build my customer package...
[12:21] <tgywa> SpamapS, /opt/php5-2
[12:21] <\sh> SpamapS: the problem is, that I want to have this tool ready not only for linux instances but also for Windows instances...I'll dive into this topic a bit more during the next days, eventually I find something good or I need really to sit down and code something
[12:21] <tgywa> SpamapS, exec /opt/php5.2/bin/php-cgi ...
[12:21] <tgywa> #exec /usr/lib/cgi-bin/php5
[12:22] <SpamapS> tgywa: yeah, I'd suggest building the 5.2 package the same way to avoid problems...
[12:23] <tgywa> SpamapS, so ... shall I leave it as it is...?
[12:24] <SpamapS> tgywa: what would be the benefit of having it packaged?
[12:24] <tgywa> SpamapS, if the web server fails ... we can get it from our org. repository and install it .
[12:26] <SpamapS> tgywa: you should have full backups anyway. ;)
[12:27] <tgywa> SpamapS, but seems that ... we will have a backup ... but I would like also to learn how to make a custom package
[12:28] <SpamapS> tgywa: you can try using the karmic package here, and changing all of the install paths to /opt/php52  https://launchpad.net/ubuntu/+source/php5
[12:30] <SpamapS> tgywa: one thing that may be difficult for you is all of the patches that are done in ubuntu's packages.
[12:30] <SpamapS> tgywa: Even simpler might be to just create your own source package that builds a more monolithic php52 package without all of the modules in separate packages.
[12:32] <tgywa> SpamapS, I use this way to compile PHP 5.2.14 and get it working on my Ubuntu 10.04 ...http://webpagedeveloper.me/blog/ubuntu-1004-lucid-lamp-server-running-php-52-and-53
[12:34] <SpamapS> tgywa: yeah, so you can put that stuff into a rules file, but it is somewhat non trivial as a "first time trying to build a package" project. ;)
[13:00] <toshko> Hi guys, again! I'm testing a drbd cluster with soft RAID0 under it. I disconnected one of the sata HDDs of the master node form the sata cable to see what happens. DRBD saw it but "cat /proc/mdstat" doesn't see the change - shouldn't it be in degraded mode??? DRBD is in diskless mode now.
[13:01] <toshko> I have never made raid0 till now (only RAID1) with mdadm, but shouldn't it be in degraded mode when I forcibly disconnect the cable of one HDD (out of 2)????
[13:03] <soren> That wouldn't count as degraded mode.
[13:05] <_ruben> raid0 can't "degrade", only "fail"
[13:05] <soren> 2 disks of 10 GB  RAID0'ed together doesn't mean the first 10 GB land on the first disk and the next 10 GB on the next. The first block is on disk 1, second block is ondisk 2, third block is on disk 1, etc.
[13:06] <soren> So missing a disk would make it fail, like _ruben said.
[13:06] <soren> Not just be degraded.
[13:06] <soren> (blocks are 64kB by deafult, by the way)
[13:07] <soren> toshko: Can you pastebin your /proc/mdstat
[13:16] <toshko> well ok, but it stays active for 15 min.now
[13:18] <toshko> http://pastebin.com/sGYHayEw
[13:29] <toshko> This is the syslog: http://pastebin.com/eS18U2jb
[13:30] <toshko> Seems that the DRBD is functioning correctly by migrating into diskless mode, but MD is doing nothing!
[13:31] <toshko> I'm suspecting a bug already! I could be hanging with 1 disk forever if a REAL failure occures :-))))
[13:32] <lau> hello I am stuck with Windows 7 talking to samba 3.0.28 on hardy ,
[13:32] <lau> user can login ok and write to the first level share but if I go down to subdirectories I get read-only
[13:32] <lau> any idea ?
[13:34] <patdk-wk> did you bother to check the permissions on your directories?
[13:34] <toshko> lau: are the permissions correct on down level subfolders?
[13:34] <patdk-wk> works fine here on hardy and win7
[13:39] <lau> thx patdk-wk toshko i fixed it !
[13:40] <lau> I did check the permissions on your directories but not the usergroup :(
[13:55] <ttx> mathiaz: o/
[13:55] <mathiaz> ttx: o/ !
[13:55] <ttx> mathiaz: did JamesPage contact you about the openldap bug he was working on ?
[13:56] <mathiaz> ttx: yop - on my todo list for today
[13:56] <ttx> mathiaz: ok cool. He is not around, so I was wondering.
[13:56] <mathiaz> ttx: I saw the bug - and put it on my todo list
[13:56]  * ttx tries to see what could cause bug 657149
[13:57] <mathiaz> ttx: wanna mumble-catch-up?
[13:57] <ttx> i can do that.
[14:02] <SpamapS> doh.. i386 buildd was available instantly.. amd64.. 11 hours. :(
[14:41] <jdstrand> soren: hey, when you get a chance, would you mind committing my patch to vm-builder trunk (if you haven't already)?
[14:45] <tgywa> SpamapS,
[14:45] <tgywa> SpamapS, so ... is there any source to read ... or should I gues guess?
[14:47] <SpamapS> tgywa: https://launchpad.net/ubuntu/+source/php5 .. you can download the source packages listed there and look at the build process.. it is .. crazy though.
[14:49] <tgywa> SpamapS, what to dounload from there?
[14:52] <SpamapS> tgywa: seriously.. I'm going to stop you. The php5 package is *really* complicated. If you haven't ever looked at a package before, its NOT the first one you should look at.
[14:53] <tgywa> SpamapS, I will do my first package building
[14:53] <SpamapS> tgywa: you're probably better off starting from scratch with the 'dh-make' package and building a new php52 package with that.
[14:54] <SpamapS> tgywa: you'll just need to add a few override rules to debian/rules to configure differently. Much much simpler.
[14:56] <zul> uh...php52?
[15:12] <zul> good luck with that
[15:13] <_ruben> reminds me of trying to package/backport php4 for hardy .. ended up giving the devs root access to sort it out themselves and completely isolated the vm networkwise
[15:13] <_ruben> "php4 is oel, dead" .. "yeah, but we have this legacy app" .. "so port it to php4" .. "that's too much work, just give us php4"
[15:14] <_ruben> err .. port it to php5 ;)
[15:16] <GMX750> Hi I need some help with LikewiseOpen issues: please read my problem at------------------>http://paste.ubuntu.com/513148/
[15:19] <GMX750> http://paste.ubuntu.com/513148/
[15:28] <tgywa> Snadder, hi
[15:29] <tgywa> SpamapS, I am doing my first packaging ... question ...about the control file ... should I create my self ... or is that included in the downloaded source package?
[15:33] <zul> JamesPage: how trivial is the patch?
[15:34]  * JamesPage looking now
[15:35] <JamesPage> zul: its not that complicated - http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=5;filename=_parseaddr.diff;att=1;bug=516300
[15:35] <zul> JamesPage: thumbs up then
[15:41] <JamesPage> zul: thanks for the guidance :-)
[15:41] <zul> JamesPage: no problem
[16:20] <JamesPage1> SpamapS: ping
[16:27] <zul> SpamapS: uploaded
[16:27] <zul> JamesPage1: hes gone to drop off his toddler
[16:28] <JamesPage1> zul: cheers
[16:29] <Thedon_> Hi, I want to automate sending emails from php but i have no idea how to install a mail server to let me do this, can somebody talk me through it?
[16:32] <cfairles> step 1: rtfm, step 2: ????, step 3: profit ;)
[16:32] <cfairles> couldn't resist
[16:33] <intel352> Thedon_, just install sendmail
[16:33] <Thedon_> ive heard sendmail is old and postfix is better
[16:34] <Thedon_> also, i have installed sendmail in the past but the mail was never sent, just waited in the queue
[16:37] <pmatulis> Thedon_: postfix is more popular in the ubuntu community.  so more support there
[16:37] <blackxored> hello guys, I'm seeking for a good solution for a desktop cloud installation on a low-bandwidth connection (meaning 15kb/s download or so) for testing, my eyes are rolling around eyeOS and FreeNX, I'm seeking for some advice, besides the one that I shouldn't be doing it on that connection ;)
[16:37] <Thedon_> well can you tell me how to install and configure sendmail properly
[16:38] <stickystyle> Thedon_: don't use sendmail, use postfix https://help.ubuntu.com/10.04/serverguide/C/postfix.html
[16:38] <Thedon_> stickystyle: thanks
[16:39] <stickystyle> sendmail is a universe package so it doesn't get updates and patches as frequently as postfix which is in main
[16:39] <stickystyle> np
[16:39] <blackxored> someone?
[16:43] <i2v8an> blackxored: are you talking about running a cloud locally?
[16:43] <blackxored> i2v8an, no, the total oposite, running my desktop remotely
[16:47] <i2v8an> oh, so you're looking for a provider?
[16:48] <i2v8an> oh, wait now I get what you're asking.  Isn't that kind of a client question rather than a server question?
[16:48] <blackxored> i2v8an, it's a server cuestion since I'm going to install freenx or eyeOS or whatever you recommend on one server i'm administering, so I wanted to hear from you
[16:48] <blackxored> then I'll connect to it over the afore-mentioned connection
[16:49] <i2v8an> I've used eyeOS in the past
[16:56] <i2v8an> blackxored, I found eyeOS to be more like a conferencing/ social hybrid.  I think freenx looks like a heftier bit of software to play with.
[16:57] <blackxored> i2v8an, I mean which is faster, I think eyeOS will restrict me a little,
[16:58] <i2v8an> this is for a ~15kb/s connection?  I would think eyeOS would work faster since it's over http and uses your browser.
[17:01] <blackxored> i2v8an, yeah I though that as well, but won't eyeOS restrict me, I mean it will let me actually "work in the cloud" I mean as I would on my desktop machine?
[17:01] <blackxored> I mean I have terminal access, install stuff, everything like that
[17:01] <i2v8an> yeah, eyeOS is really just a web server toy
[17:02] <blackxored> i2v8an, ok I'll give a try but I believe freenx will do it better, I just wish to have a faster one
[17:02] <i2v8an> freenx looks cooler... lol
[17:04] <JamesPage> zul:  that mailman bug is not quite as simple as it first looked....
[17:05] <aliverius_> where can i find a list with differences between ubuntu server 10.04 and 10.10?
[17:12] <dschuett> anyone ever have problems with linux detecting when your ipaddress has changed due to a dynamic ip through you isp provider??
[17:13] <dschuett> if my isp chages my ip address (which happens every once in a blue moon) my firewall never detects that...not even after restarting...it either holds the old ip or no ip at all
[17:13] <Thedon_> dschuett: do you have the dynip client installed?
[17:14] <dschuett> Thedon_: i have ddclient installed (is that the same thing?)
[17:14] <dschuett> or what does dynip do?
[17:14] <Thedon_> dschuett: what dynamic dns service are you using?
[17:15] <dschuett> dyndns
[17:15] <Thedon_> what isp?
[17:15] <dschuett> cox
[17:15] <Thedon_> well I dont know anything about cox, but you will need to install the dyndns client to track you ip address
[17:16] <Thedon_> the reason i asked is because on my router (supplied by isp) there is a option to set up dyndns directly in the router
[17:16] <Thedon_> so i dont need the dyndns client
[17:17] <dschuett> ddclient does that...but the problem is that my firewall that is supposed to update dyndns is what can't detect the new ipaddress... so therefore ddclient can't sent dyndns the new ip
[17:18] <Thedon_> ok, well is your server behind a router that is port forwarded ?
[17:19] <Thedon_> dschuett:
[17:19] <Thedon_> ?
[17:20] <dschuett> my server IS the firewall...
[17:20] <dschuett> and router
[17:21] <dschuett> Thedon:^^
[17:21] <dschuett> Thedon_:^^
[17:22] <Thedon_> ah ok, that may be a different problem to what i was thinking
[17:23] <dschuett> Yeah, life would be much easier if i just bought my own static ip...but that is a little spendy
[17:23] <Thedon_> Yep :(
[17:24] <Thedon_> They should be free, It doesn't cost ISPs any money to use Static I.Ps if they give them to everybody.
[17:24] <Doonz> ah but then how would they make money
[17:24] <Doonz> cause its very expensive to have ip's.....
[17:24] <Thedon_> I think they make enough from general bills to be honest
[17:25] <Thedon_> freeola for example, use static i.p's as standard for all broadband customers
[17:26] <dschuett> yeah, i don't see what the difference is...and you would have to think that it would be much easier to track if you knew which ip belonged to who at all times
[17:27] <Thedon_> it does cost them money to take dyn. ip s out of the random circuit but doesnt cost anything if they are all static, why dont they do that then? il tell you why, to rip us off
[17:27] <wizardslovak> hello people
[17:27] <wizardslovak> i am having little problems with apache
[17:28] <Thedon_> just ask your question mate, if somebody knows the answer they will help you
[17:28] <RoAkSoAx> ttx: ping
[17:28] <wizardslovak> i did modify "default" conf in sites-available
[17:28] <wizardslovak> with notepad++ in my windows pc
[17:28] <wizardslovak> saved it as txt
[17:28] <wizardslovak> but a2ensite "site does not exists"
[17:29] <Thedon_> wizardslovak: please ask questions all in one line, it makes them easy to read and more likely somebody will help you
[17:29] <wizardslovak> oki sorry about it
[17:31] <wizardslovak> should i upload new config file to sites-available and then a2ensite the name of file?
[17:34] <Pici> wizardslovak: The default site is enabled by default I beleive.  You should just need to restart/reload apache.
[17:36] <wizardslovak> yea but i did new config file
[17:37] <wizardslovak> which point to website and root folder where index.html is
[17:37] <Pici> wizardslovak: Try that then.
[17:38] <wizardslovak> yea but when i do a2ensite newconf is says "site new.us doesnt exists
[17:42] <wizardslovak> i cant enable it
[17:56] <Maleko> can i actually use packages from backports.org with ubuntu?
[17:56] <wizardslovak> so noone is here now
[18:02] <wizardslovak> so i made config file in /sites-enabled , then i use sudo a2ensite new.us and still i am getting error "site new.us does not exists!"
[18:03] <stickystyle> wizardslovak: you put the config file in sites-available
[18:03] <wizardslovak> yes i did
[18:03] <wizardslovak> basically only think i changed was serveradmin and root folder
[18:04] <Brumle> wizardslovak: did you reload apache2?
[18:04] <Pici> Putting it in sites-available and then running a2ensite makes a link of it in sites-enabled.  If its already in sites-enabled you don't need to do anything, just releoad apache2.
[18:05] <wizardslovak> still nothing
[18:05] <Pici> What exactly are you checking?
[18:06] <wizardslovak> i want to put website on it
[18:06] <Brumle> wizardslovak: does this command tell you anything:    tail -f /var/log/apache2/error.log
[18:13] <wizardslovak> "could not reliably determite the servers fully qualified domain name , using 127.0.1.1 for servername"
[18:19] <Brumle> wizardslovak: is the DNS set correct, and does the line "ServerName" in your sites-available/site.conf-file match?
[18:20] <Brumle> wizardslovak: does the DocumentRoot folder exist and contain index.html, and does the www-data user have read access?
[18:26] <wizardslovak> i dont have servername there at all
[18:27] <wizardslovak> i added ServerName with name of server
[18:27] <wizardslovak> document root exists and there is index.html
[18:28] <wizardslovak> su
[18:30] <wizardslovak> "a2ensite file " getting error "site file not properly enabled :/etc/apache2/sites-enabled/file is real file , not touching it"
[18:45] <AstralStorm> hey there
[18:45] <AstralStorm> where are SELinux rules located in Ubuntu Server? (10.x)
[18:46] <AstralStorm> oh wait, it's apparmor and not SELinux
[18:46]  * AstralStorm hides in shame
[18:46] <MmikeNekud> I'm installing Ubuntu Server 10.04 LTS from usb stick and after I choose language it complains I have no CDROM. Indeed, I don't have CDROM. Can I somehow circumvent that?
[18:47] <wizardslovak> lol
[18:47] <AstralStorm> should I use /var/lib/bind or /var/cache/bind for slave zones?
[18:48] <wizardslovak> i did few installs from usb stick and never had problems with it
[18:48] <MmikeNekud> wizardslovak, i did too, but desktop versions. Now I'm trying server and it fails :)
[18:49] <wizardslovak> ok so i added index.html into file , i enabled file in config as root document and still i dont see the site
[18:49] <wizardslovak> oh ok
[18:50] <wizardslovak> i forgot to disable default hehe
[18:50] <wizardslovak> but do you actually have cdrom in there or its dissabled in bios?
[18:57] <jauntypaperclip> hey peeps. silly question. trying to copy files from one server to another. is it cp -a ~/directory for media device ~/directory that should be copied?
[18:57] <jauntypaperclip> I keep forgetting to make a note of it and I always get mixed up
[18:57] <jauntypaperclip> thanks!!
[18:58] <Pici> cp /source/ /destination/
[18:58] <Pici> or even /source1 /source2 /source3 /destination/
[18:59] <Pici> The same for scp.
[19:04] <jauntypaperclip> Thanks Pici and that retains the permissions right?
[19:05] <Pici> jauntypaperclip: yes, using -a
[19:05] <jauntypaperclip> aw wicked! thanks for the help Pici
[19:12] <AstralStorm> jauntypaperclip: tar is recommended anyway
[19:13] <AstralStorm> should be faster than sending via scp and maybe even than sftp
[19:15] <MmikeNekud> wizardslovak, i dont have CDROM in the machine
[19:15] <wizardslovak> hmm
[19:26] <jurgen> question, if I update my xubuntu distro (have 9.10 now on server) do i need to save any files from the OS before in order for my data array to still build/work in the new distro (os isnt on array)
[19:26] <jurgen> using mdadm for aaray
[19:50] <smoser> ttx, i loaded my specs
[19:52] <RoyK> jurgen: that shouldn't be necessary, but as always, keep a backup somewhere in case things go bad
[19:59] <T3CHKOMMIE> hey guys im having a problem with virtualhosts. I had my host set up in a ports.conf file and now i cant seem to find the file anymore... i have 3 vh running and all are still running (use them to manage subdomaines) but i want to add another one and cant remember how i did it. help?
[20:00] <RoyK> T3CHKOMMIE: virtualhosts should be configured in /etc/apache2/sites-available and symlinked to from /etc/apache2/sites-enabled
[20:01]  * e_t_ is away: Gone away for now
[20:02]  * e_t_ is back.
[20:02] <T3CHKOMMIE> RoyK i followed a tut a few years back it had me messing with a ports.conf file... i just needed the document name and the document root....
[20:02] <T3CHKOMMIE> also, where can i go to see my virtual hosts that are currently set up? i cant even find the files for the ones working.
[20:02] <T3CHKOMMIE> :s
[20:05] <T3CHKOMMIE> RoyK, maybe you could help me with the more imediate problem. im trying to link a subdomain newsub.t3chkommie.com to a webgui on my server running on a random port, how do i get the subdomain to talk to the guis port?
[20:07] <T3CHKOMMIE> im thining something like <virtual host * :12345> <Documentname: newsub.t3chkommie.com>?
[20:07] <pepo> hi people!!!
[20:07] <pepo> people! I have a 300GB hard drive. when i do dd if=/dev/zero of=temp bs=1M count=3096 I get 61.6 MB/s (dual xeon 2.4, 2GB memory), but when I do the same command on a desktop with Intel i7 Core 500GB hard drive I get 46.2 MB/s. Could somebody give me some ideas where to dig?
[20:08] <pepo> 300GB hard drive 150 i/o transfer rate, 500GB hard drive 300 i/o transfer rate
[20:09] <pepo> both machines uses ext4
[20:10] <pepo> 300GB machine ubuntu desktop 10.04, 500GB machine ubuntu server 10.04
[20:11] <pepo> i-m trying to understand why my very old desktop works faster than a Intel i7 Core server.... O_o
[20:13] <e_t_> What's the interface type and rotational speed of the drives?
[20:14] <pepo> SATA, 7200 rpm
[20:14] <e_t_> Both?
[20:14] <pepo> SATA II, sorry
[20:15] <pepo> yes
[20:17] <pepo> my server (with 500GB) has a Highpoint 2320 raid controller with 8 hard drives 2TB each on raid5 and they also work slower than my old desktop
[20:17] <pepo> BUMP!!!
[20:17] <pepo> i do not understand this world =)
[20:19] <jcastro> SpamapS: know anyone @ facebook through anyone?
[20:19] <jcastro> We basically have to resort to beggin: http://bit.ly/92l06U
[20:22] <RoyK> pepo: those cheap raid controllers are usually quite bad
[20:23] <RoyK> pepo: I'd recommend opensolaris or openindiana for such a setup
[20:23] <RoyK> it beats most hardware raid controllers pretty badly
[20:23] <RoyK> with zfs
 damn with this cheap raid controller, i will change it with an adaptec 3805. but with local hard drives i'm confused totaly
[20:24] <RoyK> pepo: why? try openindiana on some cheap SATA controllers instead
[20:24] <RoyK> better performance, better safety for the data
[20:24] <RoyK> more control
 i'm too far of opensolaris and openindiana
[20:25] <RoyK> dynamic volume expansion if you replace the drives (one by one) with larger drives
[20:25] <RoyK> we're setting up a couple of 120TB openindiana boxes next week :D
[20:25] <pepo> na how about samba?
[20:26] <pepo> my users use win7 desktops
[20:26] <RoyK> pepo: seriously - try to dig into zfs - it's got a native in-kernel CIFS server that beats the shit out of samba
[20:27] <RoyK> pepo: give this 15 minutes of your time http://hub.opensolaris.org/bin/download/Community+Group+zfs/docs/zfslast.pdf
 realy?
[20:28] <RoyK> really what?
 about zfs
[20:28] <RoyK> zfs currently beats all other open source filesystems
[20:29] <pepo> but how about compatibility with win7 users?
[20:29] <RoyK> with checksumming on the block level and so on, your data is far safer there
[20:29] <RoyK> not a problem
[20:29] <RoyK> the CIFS server in osol/oi is very fast and supports native windows ACLs
[20:29] <RoyK> compatible with the ZFS ACLs
[20:29] <ivoks> not sure if in-kernel servers are smart design
[20:29] <RoyK> which is something you don't find on linux
[20:30] <pepo> ok. thanks. I will play on evening with it on a test machine
[20:30] <RoyK> ivoks: well, it works, and it's fast
[20:30] <ivoks> RoyK: and it's a security wise questionable
[20:30] <ivoks> RoyK: and hard to update
[20:30] <RoyK> ivoks: linux still uses nfs in kernel, for good reason
[20:30] <ivoks> that's true
[20:30] <RoyK> same applies to other protocols
[20:31] <RoyK> try to benchmark userspace nfs vs kernel nfs
[20:31] <ivoks> not everything is in speed
[20:31] <RoyK> ivoks: well, on large setups, speed is quite important. on your home setup, no need to do fancy stuff
[20:31] <ivoks> otherwise, we would all drive maserati :)
[20:32] <pepo> for me is very important to have a high speed file server in our network
[20:32] <ivoks> RoyK: everything depends on enviroment and type of the project
[20:32] <RoyK> ivoks: it's a wee difference here - zfs on openindiana is free
[20:33] <RoyK> ivoks: I agree, but the only reason I see not to use openindiana for storage atm is posix ACLs, which aren't supported, but then, very few uses them
[20:33] <pepo> we have 1Gbit network with good benchmarks but our file server with current setup is our bottleneck in the project
[20:34] <RoyK> pepo: if you move to zfs, you will need to backup/restore the data, you can't just migrate from hw raid to zfs
[20:34] <RoyK> but then, imho, it's worth it
[20:35] <pepo> in our case we will do it on second server with same parameters. first server will remain as a backup server
[20:36] <RoyK> pepo: also, with zfs you have zfs send/receive, which can take a full backup of the changes on the system, snapshot data etc
[20:36] <RoyK> snapshotting is a very nice thing to have
[20:36] <pepo> we already bought an adapteck 3805 raid controller with 8 WD Black hard drives. hope with openindiana I will get better performance than existing
[20:36] <RoyK> with OI and the CIFS server, they show up in Windows explorer as "previous versions"
[20:37] <RoyK> pepo: if using zfs, DO NOT use hardware RAID
[20:37] <pepo> why?
[20:37] <RoyK> pepo: spend some minutes on that presentation I posted
[20:37] <RoyK> [21:27]  <RoyK> pepo: give this 15 minutes of your time http://hub.opensolaris.org/bin/download/Community+Group+zfs/docs/zfslast.pdf
[20:37] <moonpup> anyone know if there is a way to get the built in chroot functionality of openssh to accept scp along with sftp?
[20:37] <pepo> aha. ok
[20:38] <RoyK> moonpup: IIRC I saw an article on that some time back - google for it :)
[20:38] <moonpup> RoyK: i have... no luck :(
[20:38] <RoyK> hm... ok
[20:38] <moonpup> i have the sftp piece working, but it won't accept scp
[20:39] <RoyK> sorry - don't know
[20:39] <moonpup> no prob, thanks
[20:40] <ivoks> moonpup: e
[20:40] <ivoks> he left
[20:40] <RoyK>  
[20:46] <SpamapS> JamesHarrison: sup?
[20:47] <SpamapS> JamesHarrison: oops
 shivering skin
[20:47] <SpamapS> that was meant for JamesPage1 but the's quite gone ;)
[20:47] <RoyK> pepo: huh?
[20:47] <SpamapS> jcastro: I have conversed with many facebookers, but not really connected to any. I know people who know people tho.
 looks interesting comparing with traditional file system :)
[20:48] <RoyK> it certainly is :)
[20:49] <RoyK> and the thing about common read failures not detected by the drives in a bunch-of-drives-setup is not marketing
 now i have to wait when new raid controller and hard drives will come and play with zfs =)
[20:49] <RoyK> we have ZFS filesystems on some NexSAN boxes, and on a few of them, we have gotten unrecoverable errors in files
 how long time did you use zfs?
[20:50] <RoyK> that hasn't happened (yet) with zfs, since zfs will fix that if the redundancy is sufficient
[20:50] <RoyK> pepo: in production on el-cheapo drives with el-chepo SATA controllers, about a year
[20:51] <RoyK> 2x50TB on that setup (one for backup with zfs send/receive)
[20:51] <jcastro> SpamapS: we really need to talk to someone about this API thing, any help "calling all cars" would be mostly appreciated, we're getting railed for broken facebook
[20:51] <SpamapS> jcastro: hurm? what whose API?
[20:51] <jcastro> the facebook one
[20:51] <SpamapS> Oh for gwibber?
[20:51] <jcastro> yeah
[20:52] <SpamapS> Isn't this just another cat and mouse game like gaim vs. aol ? :-P
[20:52] <zul> SpamapS: well according to the facebook movie they are a bunch emacs users
[20:53] <RoyK> pepo: and this system has been running with quite some load....
[20:53] <SpamapS> zul: no wonder
[20:53] <zul> SpamapS: agreed
 use zfs by default or I have to make it manually installing openindiana on server?
[20:53] <RoyK> no, it's the default
[20:54] <RoyK> well, sorry, do you mean zfs on linux?
 yes, in my case our users use file server very hard
[20:54] <pepo> yes
[20:54] <RoyK> you can use zfs on linux with zfs-fuse
[20:54] <RoyK> but I wouldn't recommend that
[20:55] <pepo> no, i mean if i will go with openindiana
[20:55] <RoyK> setup a test box with openindiana
[20:55] <RoyK> test zfs with the in-kernel CIFS server
[20:55] <RoyK> if you have a windows DC, it'll plug neatly into that
[20:55] <pepo> ok. i will play on evening with it
[20:56] <pepo> thank you for good idea, appreciated
[20:56] <RoyK> pepo: if you want to use that raid controller of yours, configure it as a JBOD
[20:56] <RoyK> let ZFS do the rest
 i already had experience with JBOD.... disappointed
[20:58] <RoyK> pepo: you misunderstand
[20:58] <RoyK> pepo: JBOD = just a bunch of drives - with that, you'll let zfs do the magic
 true, but it works slower than a RAID5
[20:59] <RoyK> not with zfs
[20:59] <RoyK> beleive me
[20:59] <RoyK> if you setup a raidz1, that'll be like raid5
[20:59] <RoyK> raidz2 like raid6
[20:59] <RoyK> et cetera
 is very difficult to accept something that differ of traditional hw raid configuration :)
[21:01] <RoyK> only you're no longer limited to a little chip on some RAID board to do the checksumming and so on, you'll use the CPU(s) in your system and the memory there which is vastly faster than the stuff on the RAID card
[21:02] <RoyK> pepo: well, try it, grab 4-5 drives and an USB stick, install OI on the stick and setup a raidz or raidz2 on the drives, test it, benchmark it, and my guesses are you will see the benefits
[21:02] <RoyK> not only in the performance respective, but also for safety
[21:04] <pepo> ok. i will try. already anxious =)
[21:04] <RudyValencia> Hi, how do I set up my netboot system to offer both Lucid and Maverick?
[21:05]  * RoyK has no friggin' idea, but hopes someone else knows
[21:05] <RoyK> pepo: did you read through that document?
[21:06] <pepo> yes
[21:06] <RoyK> I can confirm it's true, all of it, except perhaps the crypto part which isn't in (yet)
[21:07] <pepo> this why i'm saying that looks completely different of traditional setups
[21:08] <RoAkSoAx> ivoks: /win 4
[21:08] <RoAkSoAx> grrrrrrrrrrrrr
[21:08] <RoyK> pepo: it is
[21:13] <RoyK> pepo: the reason is the 'traditional' setups are all controlled by hardware manifacturers that wants us to beleive a $3k RAID controller can do more than a good operating system
[22:23] <NginUS> how do I enable a second bond device in 10.10? I have /etc/modprobe.d/aliases.conf set to maxbonds=2, & still returning 'bond1: device not found'
[23:16] <ideopathic> I've been trying to get vsftpd working with libpam-mysql or pure-ftpd-mysql working.  In both cases mysql fails in the password match (no errors or problems connecting to the db and finding the user).
[23:17] <ideopathic> one thing i have notices, is that mysql's md5 does not return the same value for the password as openssl's md5 does.
[23:20] <hallyn> zul: around?
[23:20] <zul> hallyn: you got me on my good side :)
[23:21] <hallyn> zul: should i created a blueprint for containers in uec?
[23:21] <hallyn> zul: or do you basically have that covered?
[23:22] <zul> hallyn: i think we have that covered, the basics should be pretty easy to do
[23:22] <zul> in theory
[23:22] <hallyn> zul: ok, won't create a blueprint then, thanks
[23:22] <hallyn> haha
[23:22] <hallyn> then maybe i'ts worth it after all
[23:24] <hallyn> you don't object either way?
[23:24] <zul> nope
[23:24] <zul> hallyn: but you might want to double check with daviey though
[23:24] <zul> since he is the lead
[23:26] <Daviey> hallyn: Sounds like a good idea.... :)
[23:27] <hallyn> Daviey: zul: ok, cool, thanks
[23:28] <RoyK> pepo: ping
 replay from pepo 1ms.
[23:30] <clayd> need a little direction.  trying to stop vsftpd on 10.04 but when i run 'sudo /etc/init.d/vsftpd stop' it tells me to use service,  but when i try using service it says unknown.  Can someone point me to what i am missing?
[23:30] <clayd> wow that was a poor question.  worry
[23:30] <clayd> sorry
 kill -9 <pid> of vsftpd
[23:32] <clayd> does that fix the problem or just kill vsftpd
[23:33] <RoAkSoAx> clayd: service vsftpd stop tells you it is unknown?
[23:34] <clayd> yes that is correct.  normally i would just stop if from /etc/init.d/ but i get a message saying that "rather than invoking init ....... use the services utility
[23:36] <RoAkSoAx> clayd: try sudo service sftpd stop
[23:36] <RoAkSoAx> clayd: or sudo service ftpd stop
[23:38] <clayd> didn't work.  when i do a service --status-all  sftpd and ftpd are not listed.  vsftpd is listed but has a ? next to it
[23:38] <RoAkSoAx> zul: any ideas ^^
[23:38] <zul> hmmm?
[23:38] <zul> sudo stop vsftpd
[23:39] <RoAkSoAx> clayd: ^^
[23:39] <clayd> this is an Ubuntu 10.04 install on rackspace cloud.  I know the use a slightly different install
[23:39] <clayd> sudo stop vsftpd = stop: Unknown instance:
[23:40] <zul> clayd: what does /var/log/daemon.log say?
[23:43] <clayd> nothing about vsftpd.
[23:47] <ideopathic> clayd: did you look at /etc/init.d/vsftpd to see if the script has been modified incorrectly?
[23:48] <clayd> i will look but that is getting well my level of understanding.  anything in particular i should look for?
[23:49] <clayd> oh wait.  it is what is giving me the message.  hrmm
[23:50] <ideopathic> it's usually just a bash script that dispatches various commands.
[23:51] <clayd> i am just learning that :)  so i looks like service is not set up to recognize vsftpd
[23:51] <ideopathic> it may at least give you an idea on where it's going wrong.  might be it can't fine the PID file where it expects it.
[23:51] <ideopathic> I can't look at mine, as I uninstalled vsftpd to use pure-ftpd
[23:51] <clayd> is that a recomendation?  :)
[23:52] <ideopathic> not really... vsftpd is supposed to be more secure, however, i could not get libpam-mysql working with it on 10.04.  pure-ftpd is more of a comfort zone for me
[23:53] <clayd> what does it mean for an initscript to be converted to Upstart?
[23:55] <ideopathic> unfortunately i don't have the answer to that one.
[23:56] <ideopathic> http://upstart.ubuntu.com/
[23:56] <cfairles> clayd, initscripts are old sys-v style, a simple script in /etc/init.d. Upstart uses a single script to start services (symbolic links in /etc/init.d to the upstart script) and some weird configuration file that describes the startup process
[23:57] <cfairles> so some service that had an initscript, but now just has a symbolic link in /etc/init.d has been converted from initscript to upstart
[23:59] <clayd> so i am starting to understand that part, but i am still a little confused why 'sudo service vsftpd stop' doesn't work.