[00:07]  * funkyHat jumps on tucemiux 
[00:07] <tucemiux> greetings funkyHat
[00:08] <tucemiux> I guess the place to start is the server guide
[00:08] <funkyHat> tucemiux: I would tend to start by just having a go, and then look up a guide when you don't know how to do something
[00:09] <funkyHat> But the server guide is probably good too ⢁)
[00:10] <tucemiux> funkyHat, nowadays I have a machine that is strictly a server, I installed ESXi on it, slapped ubuntu server on top of it, and migrated to the ssh server that I configured
[00:10] <tucemiux> funkyHat, is it a good idea to have a server to just serve web pages?
[00:12] <funkyHat> tucemiux: I can't really answer that question without some more information :/
[00:13] <tucemiux> well basically, im wondering if it might be a security risk to host a web server on the server that also has all my files
[00:14] <tucemiux> files meaning user data, as in hilton videos
[00:14] <funkyHat> Well of course everything has some risk, but I think as long as you're sensible you shouldn't have much trouble
[00:15] <funkyHat> Firstly web servers run as the user www-data, so an attacker can only see what that user can see if they do manage to gain some kind of control
[00:16] <funkyHat> Things like php (or any other server side scripting) increase the risk over just serving static files
[00:16] <funkyHat> But if you've already got ESXi running, why not create a separate VM for your web server if you're concerned?
[00:17] <tucemiux> funkyHat, i wont be doing anything serious with that web server, maybe just serve static stale data, nothing fancy
[00:18] <tucemiux> yes, I was thinking about creating another server just for web serving, how much hard drive space you think I should commit to a web server?  Like 40 gigs?
[00:19] <funkyHat> tucemiux: again I have no idea, depends what you're going to be serving. I have a production server which hosts several client's sites which only has 16GB storage
[00:19] <funkyHat> *clients'
[00:20] <tucemiux> funkyHat, ill probably serve just my blog, my cheat sheets, stuff like that, any file that's big I would most likely use other sites, like youtube to share my videos
[00:23] <tucemiux> i wonder if it's difficult to configure an ldap server to use for authentication?
[00:27] <funkyHat> tucemiux: I've not done it, but I have shell accounts on a couple of servers that use LDAP for auth, so it's certainly possible ⢁)
[00:28] <tucemiux> funkyHat, basically I want to be able to configure a username and a password and any machine that attaches to my network will use the server to authenticate
[00:30] <funkyHat> tucemiux: I think you would have to configure each computer to use your directory server, but that should be possible, yes
[00:32] <tucemiux> funkyHat, exactly, you set up the username and password then you configure the machine, this way I dont have to worry about adding user accounts every time I reimage my machines
[00:34] <funkyHat> tucemiux: This might be useful https://help.ubuntu.com/community/LDAPClientAuthentication
[00:34] <funkyHat> Although possibly outdated...
[00:35] <tucemiux> funkyHat, configuring the client shouldnt be too bad, configuring the ldap server is a b@#h!, im basically trying to nail down a nail with a sledgehammer
[00:39] <tucemiux> funkyHat, i guess im going to have to wait until someone who has done before logs in, that documentation is overkill for what I want to do
[00:46] <tucemiux> funkyHat, that's over kill!! LoL  I guess I'll take a look into that later on tonight, I'm going to head back home and install that ubuntu web server
[00:48] <funkyHat> tucemiux: http://www.howtoforge.com/linux_ldap_authentication looks like it might be useful, though it isn't Ubuntu specific, and I haven't really looked at it in detail
[00:49] <tucemiux> openssh-client is installed on the client machines, right?  Not on the server themselves?
[00:49] <Pici> It wouldn't hurt to install it on your server.
[00:50] <Pici> I have it installed on all of mine, as I regularly ssh from them to other things.
[00:51] <funkyHat> It's installed everywhere by default
[00:52] <tucemiux> Pici, I was just wondering if the client packages installed more than just the command line ssh, I was going to do that on my laptop, actually
[00:53] <tucemiux> ok well its kinda cold in this starbucks, im going to go home and finish what I started, ill be back later on tonight, hopefully
[01:16] <|rt|> I'm playing with the alpha version of 10.10 server anyone know how to setup btrfs multidisk pool in the installer?
[01:17] <|rt|> I could set that up with another live cd then try to install on it
[01:48] <ryanakca> ScottK: The Kolab server will have to wait until tomorrow, yesterday's live CD doesn't boot and for some reason, no LiveCD was built today. Hopefully there will be a working live CD tomorrow.
[02:45] <robertpayne> Directories should only need chmod 644 for a webserver to serve static content from them no?
[02:47] <jmarsden> robertpayne: Directories should probably be 755, files in them 644.
[02:47] <robertpayne> jmarsden: ok.. script files 755 as well?
[02:47] <robertpayne> jmarsden: I thought directories could be lower.. why does the server need execute access for them?
[02:47] <jmarsden> Yes, that's conventional.
[02:48] <jmarsden> "execute" on a directory means something else -- search permission.    man 2 chmod
[02:49] <robertpayne> jmarsden: ahh.. thanks :) Is there any easy way to recursively set just directories to 755 without modifying files?
[02:49] <robertpayne> nvm.. find . -type d -exec chmod 755 {} \;
[02:50] <jmarsden> Yes, something like chmod ugo+X /path.to.directory
[02:50] <jmarsden> Well, you can do that too.
[02:50] <robertpayne> I guess I mean subdirectories etc..
[02:50] <jmarsden> chmod -R ugo+X /path.to.directory
[02:50] <robertpayne> wont that hit files as well?
[02:51] <jmarsden> man chmod and look at the difference between +x and +X :)
[02:51] <robertpayne> doh
[02:56] <robertpayne> sweet all fixed up :)
[03:00] <robertpayne> over past three weeks went from knowing pretty much nothing about full sys admin to running my own box completely without a control panel.. been fun
[03:48] <ScottK> ryanakca: OK.  Thanks for the update.
[05:05] <chrislabeard> What is the best way to allow my server to send emails ?
[05:05] <twb> Install and configure an MTA
[05:05] <chrislabeard> mta
[05:06] <chrislabeard> like postfix
[05:06] <chrislabeard> I don't have a mail server do i need to set this up as well
[05:06] <twb> That depends what you mean by "mail server", but: probably not.
[05:07] <chrislabeard> well it says system mail name
[05:07] <chrislabeard> just make something up?
[05:07] <twb> Er, no.
[05:08] <chrislabeard> that is one of the things that confuses me
[05:08] <twb> You should read the Ubuntu Server Guide, if you haven't already.
[05:08] <twb> A host's mailname is typically the same as its FQDN.
[05:08] <chrislabeard> right which is wwmcd.org
[05:08] <twb> For a satellite MTA, it doesn't really matter.
[05:08] <chrislabeard> Yeah its going to serve all my drupal and wordpress installs
[05:30] <chrislabeard> How can I check to see if postfix it working
[05:31] <twb> Send an email?
[05:31] <chrislabeard> is there anyway to test it form the potfix prompt
[05:31] <chrislabeard> postfix**
[05:32] <twb> What is a postfix prompt?
[05:33] <chrislabeard> when you telnet into it
[05:33] <chrislabeard> or something
[05:33] <twb> Well, yes, if you connect to the submission or smtp port, you could talk SMTP to it.
[05:43] <robertpayne> chrislabeard: are you testing simple SMTP or TLS etc?
[05:44] <chrislabeard> uhhh
[05:44] <robertpayne> chrislabeard: http://flurdy.com/docs/postfix/edition10.html#test-postfix-receive
[05:44] <robertpayne> right below it is can postfix send
[05:45] <robertpayne> if it works inside the box then hook it up to a mail client like thunderbird and tail -f /var/log/mail.log
[05:45] <robertpayne> chrislabeard: of course you have to be running a pop3/imap mail server as well postfix is only a MTA
[05:46] <chrislabeard> Oh well than maybe I should use simple SMTP
[05:46] <chrislabeard> since all I want is the server to send me messages
[05:48] <robertpayne> chrislabeard: yeah then just need postfix and telnet localhost 25
[05:48] <robertpayne> chrislabeard: follow the can postfix send on http://flurdy.com/docs/postfix/edition10.html#test-postfix-receive
[05:48] <robertpayne> to test
[05:49] <chrislabeard> anytime I try to to RCPT TO: it says Relay Access Denied
[05:50] <twb> That's because it's not an open relay and you're connecting on a non-lo interface.
[05:50] <robertpayne> means you configured it wrong
[05:50] <twb> Or you configured it wrong, yeah
[05:50] <chrislabeard> hmmm k
[05:51] <robertpayne> chrislsbeard: you need to open the relay and then block port 25/587 etc.. do NOT run an open relay
[05:51] <twb> If you're just operating a satellite site that sends mail to a smarthost and doesn't receive mail directly, you should be able to configure postfix correctly from the debconf prompts.
[05:51] <chrislabeard> okay
[05:52] <twb> I certainly would not "open the relay" even if there was a firewall there.
[05:53] <chrislabeard> k
[05:53] <chrislabeard> I reconfigured it
[05:56] <chrislabeard> hmm I wonder if its because I'm using EHCP
[06:08] <robertpayne> twb: I should say he should set it to be an open relay from localhost
[06:08] <twb> It should be that way by default
[06:08] <robertpayne> twb: yea not sure what happened on his
[06:09] <chrislabeard> I still can't get it to work, Its all good its just a test server
[06:09] <chrislabeard> Don't really need it to send email
[06:11] <twb> All servers should be able to send mail, if only for cron and atd
[06:12] <twb> And smartctl and mdadm resync notifications and logcheck...
[06:12] <ball> twb: local mail at least, not necessarily mail to/from the outside world
[06:12] <twb> ball: right.
[06:17] <ball> What does "plymouth" do?
[06:20] <twb> ball: annoys me
[06:22] <twb> Particularly since you can't remove it and can't tell it not to run
[06:22] <ball> What is its intended purpose?
[06:22] <twb> (Booting without "splash", it'll still run and hijack the vt7, then display a faux console on it.)
[06:23] <twb> ball: it's purpose is to provide a more seamless boot sequence to desktop users (than usplash)
[06:23]  * ball finds splash screens a bit odd.
[06:23] <twb> Although canonical people have assured me "it's not just splash" and "they don't want to have the plymouth argument again", I still don't get it.
[06:26] <ball> Wierd.
[06:27] <ball> Something else that's wierd: I just rebooted an Ubuntu Server box that I was able to connect to using ssh.  Now I can't.
[06:27] <ball> I added a hard disk, but that's on a second channel and wasn't booted, so I'm not sure what's wrong.
[06:28] <twb> ball: did it boot normally?
[06:29] <ball> twb: seemed to.
[06:29] <ball> ...just can't ssh to it.
[06:30] <ball> hello chilicuil
[06:30] <ball> I'll try rebooting, just in case.
[06:31] <twb> So it's in runlevel 2?
[06:31] <twb> It wouldn't surprise me if it got confused and stuck in S
[06:34] <ball> Seems to have booted this time.
[08:43] <robertpayne> is there any reason usermod -G grouptoremovefrom username wouldn't remove the user from the specified group?
[08:43] <robertpayne> oh nevermind a bit of a noob reading man pages now
[08:46] <twb> Suggest delgroup, not the low-level usermode.
[08:46] <twb> Er, deluser --group
[09:08] <robertpayne> twb: thanks.. I locked myself out of ssh by accidentially adding my sudo user to sftp group that is chrooted
[09:15] <xampart_> what does "<defunct>" mean when i do "ps aux"
[09:16] <xampart_> nevermind
[09:38] <IRConan> hi there... can anyone give me definitive information about the differences between -virtual and -ec2 kernel versions? specifically do they both support use in a Xen domU? and if so are either recommended over the other?
[09:49] <twb> IRConan: diff the /boot/config-* files of both
[09:52] <soren> I'm afraid there's more to it than that.
[09:52] <soren> linux-ec2 is built from a different source package.
[09:52] <soren> (at least in Lucid and before)
[09:53] <soren> IRConan: -virtual is the same as -server, only with a bunch of modules surgically removed from the binary package.
[09:53] <soren> IRConan: It supports pvops, so it can run as a Xen domU as well as in KVM or on real hardware.
[09:54] <soren> IRConan: -ec2 only works as a domU and you should only need it if you're running on ancient versions of Xen.
[09:56] <robertpayne> Is there anywhere you can find a list of how a package was compiled? I'm looking to compile/install php5.3.3 but was curious if there's a way to see how php5.3.2 was compiled so I can make it as compatible as possible
[09:57] <Jordan_U> robertpayne: apt-get source to grab the source package.
[09:57] <joschi> robertpayne: sure. get the source package
[09:57] <robertpayne> joschi: jordan_U: awesome thanks :) should help a lot
[09:57] <Jordan_U> robertpayne: You're welcome.
[09:58] <robertpayne> php 5.3.3 is scheduled for 10.10 but I kinda want it for the ease of nginx integration
[10:03] <robertpayne> joschi, jordan_U: technically if I downloaded the source package for php5 from the macerick I could compile it from there?
[10:03] <Jordan_U> robertpayne: Yes.
[10:04] <Jordan_U> !prevu | robertpayne
[10:04] <robertpayne> Jordan_U: sweet. have to get all the dependencies myself but will be much easier than setting up the configuration file myself etc..
[10:04] <soren> robertpayne: Just do apt-get build-dep php5
[10:04] <soren> robertpayne: That installs all the build dependencies.
[10:05] <robertpayne> soren: sounds good but if I'm running 10.04 and it made 10.10 it could fail..
[10:05] <soren> If you add a deb-src line for maverick to your sources.list and do "apt-get build-dep", it'll grab the build dependencies of php5 in maverick.
[10:06] <robertpayne> soren: but will it compile them or grab the binaries? You can't be sure the binaries are compatible?
[10:07] <robertpayne> Jordan_U: prevu looks good, but it wants to install a ton of other things along with it not sure I really up for that
[10:22] <robertpayne> soren: should I add the maverick sources above or below the lucid ones? Not sure how apt-get resolves which ones to grab
[10:41] <soren> robertpayne: It'll grab the binaries from lucid.
[10:42] <soren> robertpayne: ...but get the list of them from maverick.
[10:42] <robertpayne> soren: ahh yes ok awesome
[10:42] <soren> robertpayne: the order of the lines in sources.list does not affect the choice of version.
[10:42] <soren> robertpayne: It only affects the order of preference for mirrors and such.
[10:43] <robertpayne> soren: ok I see.. also I'm a bit naive here.. the source package ./configure still requires you to setup all the options correct? It doesn't automatically set all the options/patches the binary dist uses?
[10:43] <soren> robertpayne: Depends on how you use it.
[10:43] <soren> robertpayne: PAckage builds in ubuntu are fully automated.
[10:43] <robertpayne> soren: I'd like to grab the 5.3.3 source from maverick and automatically build it using ubuntu's settings I don't need any custom ones
[10:43] <soren> robertpayne: If you call it like "dpkg-buildpackage -rfakeroot -b" it'll do whatever the proper ubuntu builds do.
[10:44] <robertpayne> soren: ok awesome!!
[10:44] <robertpayne> soren: they just need to backport php 5.3.3 to 10.04
[10:44] <soren> Why?
[10:44] <robertpayne> soren: it includes FPM which makes working with nginx + php a lot easier and stable
[10:45] <robertpayne> php-fpm*
[10:45] <soren> Are you implying that php for the first time in history has made a release that doesn't have any regressions at all compared to the previous version?
[10:45] <robertpayne> soren: oh probably not.. to be honest I'm sure there's some issues I'll find out...
[10:46] <soren> Maybe you won't. Maybe it's just some small thing that only affects very few users.
[10:46] <robertpayne> soren: True it's a small dot release too that enabled it 5.3.2 to 5.3.3
[10:47] <soren> That's the thing, really. Ubuntu feels it's better to ensure that stuff that worked yesterday also works tomorrow than making stuff that didn't work yesterday work tomorrow.
[10:47] <soren> People are depending on stability.
[10:47] <robertpayne> soren: That is true.. stability is really key
[10:48] <soren> Exactly. Much more important than adding any new feature.
[10:48] <soren> during development (like in Maverick right now) everything is fair game. People know that they're using a moving target. Once stuff is released, they should be able to rely on it for production use.
[10:58] <robertpayne> soren: ok one last q if you don't mind to use dpkg-buildpackage do I have to have the source package downloaded or will it fetch it?
[11:10] <soren> robertpayne: You need the source package.
[11:11] <robertpayne> soren: yup figured that out :) it's building right now..
[11:11] <robertpayne> soren: I think the php5 source literally builds an entire lamp stack.. or at least the build-dep required it
[11:18] <soren> robertpayne: I don't know what you mean. It builds neither Linux, Apache or MySQL. It only builds PHP, which seems perfectly reasonable for a php package to me.
[11:19] <robertpayne> soren: doing apt-get build-dep php5 requires a ton of stuff from mysql/apache etc..
[11:19] <soren> robertpayne: Sure.
[11:19] <robertpayne> soren: I'm not sure why :S
[11:20] <soren> robertpayne: It needs MySQL librarires and development headers to build php5-mysql. It needs apache headers and stuff to build libapache5-mod-php5.
[11:21] <soren> libapache2-mod-php5, I mean.
[11:21] <soren> you know... The usual reasons for needing build-dependencies.
[11:22] <robertpayne> soren: ah yes.. I understood those but it also was requiring apache2.2-bin
[11:23] <soren> robertpayne: Not sure why it would need that.
[11:25] <robertpayne> soren: yea I'm not sure kinda sucks I don't want apache at all.. not even mod-php
[11:25] <robertpayne> http://packages.ubuntu.com/lucid/php5 says it only requires php5-cgi OR libapache2
[11:25] <soren> robertpayne: Well, you did ask for a build identical to the one in Ubuntu.
[11:25] <robertpayne> soren: yes
[11:25] <soren> Those are runtime dependencies.
[11:26] <soren> Not build-dependencies. They're not strictly orthogonal, but almost.
[11:26] <robertpayne> soren: I see hmm I may have to dig deeper into how it runs the ./configure and build process for php5
[11:27] <soren> ...or take another look at prevu.
[11:28] <robertpayne> I wasn't quite sure what prevu does it looks like it's for backports? thought backports were for previous version binaries
[11:28] <soren> No, future.
[11:29] <soren> You're taking something from .e.g maverick and putting it into Lucid, which is older than Maverick. Hence, backport.
[11:29] <IRConan> soren: what is an ancient version of Xen?
[11:30] <soren> It does all the same stuff you did above, only automatic, and in a chroot, so you don't end up with all the build-dependencies installed on your system after you're done.
[11:30] <IRConan> my host is debian lenny so compared to ubuntu is would be considered ancient
[11:30] <robertpayne> soren: oh very nice
[11:30] <soren> IRConan: Which version of Xen do you havE?
[11:30] <IRConan> 3.2.1
[11:30] <soren> IRConan: That's fine.
[11:30] <IRConan> so I should use -virtual ? and that will do paravirt?
[11:31] <soren> IRConan: Yup.
[11:31] <IRConan> cool
[11:31] <soren> It'll detect at boot time that it's running under PV Xen and Do The Right Thing[tm].
[11:31] <IRConan> cool
[11:31] <IRConan> you can tell based on what it calls the disks ofc
[11:31] <IRConan> if they're sd* it's para if they're xvd* it's HVM
[11:32] <soren> Don't count on that.
[11:33]  * soren lunches
[11:35] <robertpayne> soren: btw thanks a ton for helping me out with this - can't say how much I appreciate
[11:55] <robertpayne> soren: http://pastie.org/1095318 definitely can't just compile php from the ubuntu source package using dpkg it loads in all the apache and mysql stuff for all the modules
[12:11] <soren> robertpayne: Sorry, what?
[12:11] <robertpayne> soren: it looks like the php package builds a bunch of modules for apache by default it's a shame it doesn't just build php and let the modules be a seperate package install
[12:12] <soren> robertpayne: They are separate package installs.
[12:12] <soren> robertpayne: They're just built from the same source package.
[12:12] <robertpayne> soren: I see but is there a way to exclude them from building from the source package?
[12:13] <xampart> getting http://pastebin.com/WQ8xLSwH when trying to lvcreate and system hangs. using 2.6.32-24-server. need fix.
[12:13] <soren> I don't mean to be rude, but srsly... This isn't gentoo. We have one build from one source package, we build a stack of different binary packages, and people can use whichever ones of those they please.
[12:14] <robertpayne> soren: no worries none taken at all I totally understand
[12:14] <soren> Some packages may make it simple to only build a subset of the binary packages, but generally, they don't.
[12:14] <soren> php5 might. I don't know.
[12:14] <soren> But really, it doesn't matter.
[12:14] <soren> You're just building it once.
[12:14] <soren> If there are some of the binaries you don't need, don't install them.
[12:15] <rnewson> does anyone know if there's a netboot.tar.gz for the server install like there is for desktop (and Debian)?
[12:15] <soren> xampart: That looks like a kernel bug. Perhaps you can try in #ubuntu-kernel?
[12:16] <soren> rnewson: Just use the desktop one and don't install the desktop bits?
[12:16] <xampart> soren: ok, i try there then.
[12:17] <rnewson> soren: the server install has different boot options. I guess I can unpack the iso and read the menu.cfg's. thanks.
[12:18] <soren> rnewson: What iso?
[12:18] <rnewson> soren: the server iso. it's the boot menu options I need to get hold off. I thought there was more of a difference between server and desktop that just package selection, but I'm happy to be wrong.
[12:19] <soren> Nah, it's just the kernel and the task set.
[12:19] <rnewson> yep. and now I mount the iso I see the fileset I was after too. :)
[12:20] <rnewson> fancied converting some of my lab machines to UEC, but didn't fancy moving a CD around the server room (especially as it's 3000 miles away).
[12:20] <soren> slacker
[12:20] <soren> :p
[12:26] <eagles0513875> hey guys
[12:27] <eagles0513875> how can i create a raid mirror array
[12:27] <eagles0513875> i have a workstation ive converted into a server and just installed 2 identical hdds of 2tb
[12:28] <pmatulis> eagles0513875: use the installer?
[12:28] <eagles0513875> ?
[12:28] <eagles0513875> what do you mean use the installer pmatulis
[12:30] <pmatulis> eagles0513875: create the array during installation - you havn't provided much detail, so i'm giving you the standard method
[12:30] <eagles0513875> at what point do i do that though
[12:30] <eagles0513875> im not that far into the installation process
[12:30] <eagles0513875> if i dont do it there can i do it via command line once the server is online?
[12:30] <pmatulis> eagles0513875: during partitioning
[12:30] <eagles0513875> ok
[12:31] <pmatulis> eagles0513875: yes, provided you don't try to touch active partitions
[12:31] <eagles0513875> so you recommend the best and safest way is to do it during partitioning
[12:32] <pmatulis> eagles0513875: this is the standard yes.  and it's the only way if you intend to put system files on the array
[12:33] <sanderj> Hi.. Anyone know how to install uec on one machine instead of two as normal?
[12:33] <eagles0513875> pmatulis: thanks
[12:34] <eagles0513875> pmatulis: which option do i choose guided use entire disk then 2nd option is with lvm and 3rd is same as 2nd but encrypted lvm
[12:35] <pmatulis> eagles0513875: there should be a 'manual' option in there somewhere
[12:35] <eagles0513875> pmatulis: there isnt there is when i click manual it says configure iscsi volumes
[12:37] <eagles0513875> pmatulis: figured it out
[12:37] <pmatulis> eagles0513875: what was it?
[12:38] <eagles0513875> have to create an empty partition table on the device
[12:39] <eagles0513875> these are brand new drives
[12:49] <eagles0513875> pmatulis: then once the machine is online how can i check and verify the raid array ?
[12:50] <xampart> cat /proc/mdstat
[12:50] <pmatulis> eagles0513875: short answer: mdadm
[12:51] <eagles0513875> thanks pmatulis
[12:51] <eagles0513875> pmatulis: i have 2x2tb hdds is raid 1 a good choice?
[12:51] <eagles0513875> or should i go with something along the lines of raid0
[12:52] <pmatulis> eagles0513875: it depends what you're after
[12:52] <eagles0513875> i know the problem with raid1 if one drive goes everything is gone
[12:52] <eagles0513875> would i be able to rebuild the array?
[12:53] <_ruben> that's with raid0 (losing all when 1 fails)
[12:53] <_ruben> raid1 can be rebuilt
[12:53] <_ruben> 0=stripe, 1=mirror
[12:54] <eagles0513875> _ruben: ok so if i loose one with mirrored array
[12:54] <eagles0513875> im fine
[12:54] <eagles0513875> :)
[12:54] <eagles0513875> what strikes me as odd on ubuntu server is why it asks for how many spare drives you have
[12:55] <dom_dom> hello! im having problem with installing ubuntu server on ibm x3550. i know its certified to work with it, but when it comes to driver selection none woks. any suggestions?
[12:55] <pmatulis> eagles0513875: you can have a spare so rebuild process is smoother
[12:57] <eagles0513875> what if i dont have a spare one pmatulis
[12:57] <pmatulis> eagles0513875: then you'll need to physically put one in when you need it
[12:58] <eagles0513875> ok
[12:59] <dom_dom> \j #ubuntu-pl
[12:59] <eagles0513875> dom_dom: its /
[13:00] <dom_dom> eagles0513875: i know:P, how would i get here :) maybe you can help me with my problem ?
[13:00] <eagles0513875> dom_dom: exactly what drivers are you having an issue with. its funny i have an ibm x server and it has given me no problems with drivers
[13:01] <dom_dom> eagles0513875: megaraid_sas
[13:01] <eagles0513875> dom_dom: i dunno but mine is a simple x server
[13:01] <eagles0513875> non swap and it has onboard raid
[13:02] <dom_dom> eagles0513875: never had to use raid so i'd appreciate any help :) but i think that i'll have to ask later
[13:03] <eagles0513875> dom_dom: im learning how to setup raid as we speak lol
[13:03] <eagles0513875> software raid on a machine here at work lol
[13:04] <dom_dom> eagles0513875: that's funny :) i think i need a break. maybe then something will come to my mind
[13:04] <eagles0513875> dom_dom: lol google
[13:04] <eagles0513875> pmatulis: strange i created my array but for some reason i dont have a root partition anymore :(
[13:04] <dom_dom> eagles0513875: propably :)
[13:13] <zul> morning
[13:13] <eagles0513875> _ruben: you still here im confused about something
[13:18] <eagles0513875> !raid | eagles0513875
[13:18] <zul> jdstrand: what are those two mysql bugs you wanted me to fix?
[13:19] <jdstrand> zul: that is an overstatement. I wanted you to consider (and if rejecting, commenting why) points 2 and 3 in comment #1 of bug 578922
[13:20] <jdstrand> zul: my thoughts were otoh without regard to what might break.
[13:20] <jdstrand> zul: you'd want to read the whole bug for context of course
[13:20] <eagles0513875> is there any good apparmor documentation?
[13:20] <zul> of course
[13:22] <jdstrand> eagles0513875: sure. see https://wiki.ubuntu.com/AppArmor, especially the bottom for more info
[13:22] <jdstrand> zul: thanks! :)
[13:22] <eagles0513875> jdstrand: no , at the end there hehe
[13:24] <zul> jdstrand: about the /tmp stuff see 375371
[13:26] <eagles0513875> jdstrand: is anyone doing app aromor documentation on the wiki for lucid
[13:26] <jdstrand> zul: oh cool
[13:26] <hggdh> kirkland: there (and good morning)?
[13:26] <zul> hey hggdh
[13:26] <jdstrand> eagles0513875: https://help.ubuntu.com/10.04/serverguide/C/apparmor.html
[13:26] <hggdh> morning zul
[13:26] <jdstrand> eagles0513875: the wiki page I pointed to just hadn't been updated. I'll update it
[13:27] <eagles0513875> ok thought i was goign nuts for a sec
[13:29] <jdstrand> eagles0513875: k, I updated it
[13:29] <eagles0513875> thanks
[13:32] <jdstrand> zul: k-- I'll point that bug at 375371
[13:44] <eagles0513875> jdstrand: raid question for you when i select the partitions i want to use in the raid array on both drive it wont let me continue it goes back and has me choose the partitions
[13:44] <eagles0513875> wait hold on
[13:45] <eagles0513875> this makes no sense im trying to choose which partitions i want in the raid array
[13:45] <eagles0513875> they are my ext4 partition and swap partition and i am choosing them on both drive and its not eletting me continue :(
[13:47] <eagles0513875> nm i got it
[13:48] <Jinxed-> If I wanted to record a webcams attached to my network, be able to view them live from the network, and be able to store a buffer of a few hours for each camera, what would be the best way to do this with ubuntu?
[14:05] <zul> hggdh: can you look at #617982 a quick sec?
[14:12] <zoopster> ubottu is slacking today! it didn't pick up bug 617982
[14:24] <hggdh> bug 617982
[14:24] <hggdh> zul: looking
[14:25] <zul> hggdh: thanks
[14:42] <hggdh> zul: done
[14:42] <zul> thanks
[15:13] <hggdh> smoser: good mornig, are you genning an UEC image for 10.04.1?
[15:13] <smoser> no.
[15:13] <smoser> do you care to know why ?
[15:13] <hggdh> smoser: can I use the 20100813 one?
[15:13] <smoser> for testing ?
[15:13] <hggdh> smoser: yes, I do care to ;-)
[15:13] <hggdh> yes
[15:14] <smoser> if you're just testing "does it boot", i'd suggest trying with 20100813, and then if you see anything you dont like falling back to released.
[15:14] <smoser> the reasons are
[15:15] <smoser> bug 605079 and bug 574910
[15:15] <smoser> landscape considers the first a "must fix", and that will be ready as soon as -proposed is re-opened and the 1 week wait passes.
[15:15] <smoser> the kernel issue we want fixed too.... jjohansen is working on it.
[15:16] <smoser> i dont see a lot of reason in releasing something since i want to relesae *very* soon after that.
[15:17] <smoser> also would really like to have bug 613309 in a refresh (which, is also just waiting to get into -proposed)
[15:17] <kaushal> hi
[15:18] <kaushal> can some one please guide me regarding http://paste.ubuntu.com/478876/ ?
[15:18] <kaushal> I am on Ubuntu Server 8.04
[15:18] <hggdh> smoser: ack, makes sense
[15:19] <kaushal> i followed the suggestion as mentioned here ---> https://answers.launchpad.net/ubuntu/+source/dpkg/+question/50778
[15:19] <kaushal> but it didnot work for me
[15:53] <kaushal> checking in again for my query ?
[15:58] <sherr> kaushal: looks like this bug : https://bugs.launchpad.net/ubuntu/+source/collectd/+bug/297288
[15:59] <kaushal> sherr, have already tried out the suggestion referred by that bug
[15:59] <kaushal> did not worked for me atleast
[16:02] <kaushal> https://lists.ubuntu.com/archives/ubuntu-server/2010-August/004541.html
[16:04] <kaushal> sherr, any further suggestion ?
[16:05] <soren> jdstrand: Now that sudoers has a sudoers.d, would you consider that a proper way to let an application do stuff as root or is there any reason to use a rootwrap thingamajig?
[16:06] <soren> jdstrand: Of course assuming that it doesn't just do "whatever ALL=(ALL) NOPASSWD: ALL", but actually enumerates the specific commands it needs to run as root?
[16:06] <DaveWM> hi,  i'm running apache2 with userdir mod enabled,  if users have their own domain setup to resolve to my servers ip,  how do i get apache to point it to their public_html or ~username site ?
[16:11] <sherr> kaushal: it's an init script that fails? Maybe try debugging that - run the startup command yourself, see if there are any helpful options, debug or logging options?
[16:18] <kaushal> sherr, if you can give some hint regarding that init script ?
[16:20] <jdstrand> soren: I think that a properly enumerated sudoers file is just fine, assuming it is an otherwise non-privileged user and the commands that are given sudo access don't have a way to break out into shell (classic example, vim)
[16:21] <jdstrand> soren: it isn't as portable, and is potentially more difficult to maintain, but that is up to the developer.
[16:27] <soren> jdstrand: Cool. Thanks.
[16:31] <sherr> kaushal: sorry, I don't have collectd myself. But try running whatever the init script runs interactively i.e. in an xterm yourself. Maybe it logs a useful error message? Or maybe the command has useful debug or logging options (see man page). Good luck.
[16:37] <zul> SpamapS: lemme know when you are around
[16:39] <kees> soren: yeah, I'd agree with jdstrand. the thing I wrote for euca was basically just a stop-gap to help them get to using specific sudo rules and more careful wrapper scripts.
[16:39] <kees> soren: all priv-escalation tools should be carefully examined, but I think it makes sense to attempt to use sudo so as to not reinvent the wheel, etc.
[16:40] <soren> kees: Yeah, sudo's probably pretty solid by now :)
[16:40]  * kees just realized his terrible pun regarding "wheel"
[16:40] <ssureshot> when I run top I get this on the processor line.. Cpu(s):  1.7%us,  0.3%sy,  0.0%ni, 97.4%id,... what is the 97.4%id pertain too?
[16:40]  * jdstrand just got the pun
[16:40]  * jdstrand likes it :)
[16:40] <soren> kees: Heh :)
[16:40] <kees> heh
[16:41] <kees> ssureshot: "Idle"
[16:42] <ssureshot> ah thank you
[16:46] <smoser> hggdh, have you tested at all ? just curious.
[16:46] <sbeattie> soren: the other thing to consider is that the tools allowed in /etc/sudoers.d/ ought to be somewhat self-contained and that you can consider adding an apparmor profile for whatever the sudo target is to protect against any overlooked routes to a shell.
[16:49] <soren> sbeattie: apparmour profile is a good idea. I'll also try to be very specific about the arguments the application can pass to the individual commands.
[16:53] <mdeslaur> soren: the only thing that's not so nice about using sudo, is the application _stays_ as root, without being able to drop the privileges it doesn't need. Are you using that for a _helper_, or for the whole app?
[16:57] <tucemiux> what is LVM?  I'm going to install a server on ESXi -- virtually -- I just want to know what the LVM option will do?  I gave 40 gigs of space to ubuntu server
[16:58] <hggdh> smoser: running 20100813, no problems so far
[16:58] <smoser> k. good.
[17:07] <tucemiux> dawnn, how long does it take for the ubuntu server OS itself to install????
[17:09] <pmatulis> tucemiux: not long
[17:10] <jpds> tucemiux: Time is an illusion, lunch time, doubly so.
[17:11] <tucemiux> jpds, well excuse me if Im mistaken but it took me like 10 minutes to install just the OS itself without adding any packages
[17:11] <jpds> Sounds about right.
[17:15] <tucemiux> incredible!
[17:17] <tucemiux> ok so if I have a stable system running and I want to update it just : sudo apt-get update  ???  I dont need anything else like apt-get dist-upgrade
[17:18] <delimiter> I'm wondering why -proposed isn't listed for search/browse @ http://packages.ubuntu.com/
[18:02] <jiboumans> smoser: sweetness on the cloud-images-sans-cloud spec :)
[18:03] <smoser> thank... i'm hoping if openstack guest doesn't "just work" that it shouldn't be much work.
[18:04] <racquad> hi guys, does anyone knows where can I find a good tutorial about headless installation of a ubuntu server?
[18:11] <hallyn> racquad: I don't know of a good tutorial, but have done it two ways - (1) hacked the install cd to work over serial (does not work out of the box, but 2-3 minor changes and it works)  2) (my preference) from another already installed debian, did a debootstrap
[18:13] <racquad> hallyn, 1) unfortunatly I don't have a serial. 2) maybe I can try that one. Any tutorial? Some have advised me to make a  pre-install on another computer than plug the HD in my headless device. Others,  using a customized installation with a script
[18:13] <hallyn> what do you have on the machine to start with?
[18:14] <hallyn> racquad: if you don't have serial, there must be an ssh you can ssh into, and that's how you want to do it?
[18:14] <hggdh> folks, we need tests on the 10.04.1 server ISO -- please see http://iso.qa.ubuntu.com/qatracker
[18:14] <AndyGraybeal> hi everyone, i need to disable 'screen locking' for all my users.  currently it locks when they are 'idle'.. is there a setting i can change without having to be around when they log in and ask them to change the setting?
[18:16] <hallyn> hggdh: thx, noted
[18:19] <hggdh> hallyn: tahnk YOU :-)
[18:21] <hallyn> AndyGraybeal: perhaps 'gconftool-2 -s /apps/gnome-screensaver/lock_enabled false' or somesuch
[18:21] <hallyn> (run as their user)
[18:21] <hallyn> not sure of the exact command, nor what you actually want
[18:21] <AndyGraybeal> ah thank you hallyn, (run as their user is the key.. i missed that)
[18:22] <hallyn> cool, np
[18:23] <tschundeee> I compiled my own nginx server.. how do I make it run as a www user for example?
[18:23] <tschundeee> I think it runs as root when it is simply added to /etc/init.d/
[18:26] <racquad> hallyn, yes. I need an SSH connection. But I have read that somehow I could do a remote boot (via PXE, maybe) and then a chroot from another computer (with keyboard and screen) and then continue the installation from this second computer
[18:26] <racquad> once installed the base system, I could reboot the headless device and continue the fine tunning from a SSH connection
[18:29] <Ununtu-install> Hello
[18:30] <Ununtu-install> hi
[18:31] <tschundeee> found it /opt/nginx/nginx.conf and there you can set the user for the worker processes
[18:31] <tschundeee> :)
[18:35] <Ununtu-install> how can I get help from this channel?
[18:37] <AndyGraybeal> is there a way to completely disable screen locking?
[18:38] <AndyGraybeal> also is there a way to auto-log out after a certain time of idle
[18:40] <lenios> Ununtu-install, just ask your question
[18:45] <tucemiux> so if I want to install a web server for my own personal use, all I have to do is: sudo apt-get install apache2 -- and the server will be installed?  I could then access the server using the ip of the server?
[18:46] <zul> tucemiux: yes
[18:50] <tucemiux> zul, nice!  Let me try it right now on the ubuntu server I just cooked up on my ESXi box :-)
[18:51] <tucemiux> zul, im going to open a port in my router, can you please help me test it?? o_O
[18:52] <zul> tucemiux: i cant right now...im too busy with things right now
[18:52] <tucemiux> zul, ok thanks anyway, it's working locally so all is good
[18:55] <lenios> tucemiux, you can test it using public ip
[18:56] <tucemiux> lenios, no, i mean it already works locally, I wanted to create a port on my router then have someone try to view my web page
[18:56] <lenios> i can try if you want
[18:56] <lenios> but that would only test router, not the web server
[18:57] <tucemiux> lenios, I already tested the server locally, I can reach it locally, what's the path to "Index.html"? o_O
[18:57] <lenios> http://IP/index.html
[18:58] <asanir22> I'm trying to install ubuntu 10.04 server on an old 32 bit system(CPU celeron 1.7 256 MB Ram), during the installation(Base system installation) i get error on "linux-generic-pae"  kernel component and fails to install, any ideas?
[19:00] <patdk-wk> asanir22, try with the linux-generic, instead of pae?
[19:00] <panfist> i'm trying to use apt-mirror to create a repo and serve packages on my network, but my server doesn't have enough space to hold all the packages...how can i direct apt-mirror to download the packages to external storage?
[19:00] <asanir22> no, I don't know how?
[19:01] <tucemiux> lenios, i mean, how do you edit index.html ?
[19:01] <asanir22> patdk-wk, should I try with linux-generic?
[19:01] <patdk-wk> I would
[19:02] <patdk-wk> how to install it from the installer, I don't know
[19:03] <lenios> tucemiux, /var/www/ ?
[19:03] <patdk-wk> I do know if you do a net-install it uses linux-generic
[19:03] <patdk-wk> and if you want something different, you have to do it yourself
[19:03] <jiboumans> team, mathias will be online a bit later today, he just sent me an sms
[19:07] <asanir22> I don't have any Idea how I could do that, I've just got the x86 iso installer and failed several times installing
[19:07] <tucemiux> lenios, can you tell me the magic word:  http://elsitio.mine.nu:1110
[19:08] <patdk-wk> asanir, heh, not sure how to say it simple :)
[19:08] <patdk-wk> asanir22, give the alternate install iso a shot
[19:08] <patdk-wk> I personally have a large pxe server setup, with like all ubuntu installations selectable from a menu, it's easy to do
[19:08] <patdk-wk> but not so esay to setup pxe and tftp and dhcp to make it all work, at first
[19:09] <asanir22> whare i can find alternate install iso?
[19:09] <patdk-wk> on the website
[19:10] <lenios> katapuff!
[19:11] <zul> SpamapS: did you do the php SRU for the 64bit problem?
[19:15] <tucemiux> lenios, so it works then, excellent, thank you very much sir !
[19:24] <asanir22> join #ubuntu
[19:28] <SpamapS> zul: it is Declined for Lucid so I can't propose it..
[19:29] <seicherlbob> hi there! I'm having troubles with my new lucid server. I copied all data from my old debian lenny server on 2 sata drives (a software-raid). now when i attach one of them to my new system, the bootscreen only shows "fsck from util-linux-ng 2.17.2 \n /dev/sda1: clean, X/Y fiels, A/B blocks" and does nothing (for minutes now). I can switch to the other ttys, but they are blank. When i do Ctrl+Alt+Del it shutsdown gracefully and reboots to the same
[19:29] <seicherlbob>  screen, so it is not hanging. plz, i need help! BTW: the hdds where installed as IDE on the old server, now i use AHCI (if that matters).
[19:30]  * asanir22 discovering channel
[19:30] <zul> SpamapS: wtf?
[19:32] <seicherlbob> i can even already ping it, but ssh is not running, nor DHCP or other stuff.
[19:40] <panfist> would i be out of line asking this: why does ubuntu choose not to include nfs packages by default?
[19:45] <seicherlbob> ok, it seems that, if you have a softwareraid configured with device-names and you attach a new drive, (even if you use higher sata-ports), the device-order gets changed, so the raid cannot start up. But this message is missing on the boot-prompt.  I just started hitting any keys. when i pressed "S", the raid initialization was skipped. Why is the message from mdadm not showing on the boot-screen?
[19:46] <zul> SpamapS: bug number again?
[19:46] <SpamapS> bug 564920
[19:47] <Wyleyrabbit_> can anyone point me to a guide or howto on how to upgrade a driver on ubuntu server?
[19:48] <Wyleyrabbit_> specifically, I would like to upgrade the megaraid_sas driver that comes with Ubuntu Server 10.04 LTS, it's pretty old.
[19:59] <seicherlbob> can i define the devices of a soft-raid (mdadm) by UUID aswell?
[20:05] <soren> mdeslaur: Oh, just a helper. Well, a bunch of helpers.
[20:05] <mdeslaur> soren: ah, ok. sounds fine then
[20:07] <pmatulis> zul, SpamapS: declined?
[20:07] <SpamapS> seicherlbob: mdadm writes UUID's to the devices, which it uses to assemble the raid
[20:08] <zul> SpamapS: there you go fixed
[20:10] <SpamapS> zu	zul     
[20:10] <SpamapS> doh
[20:10] <SpamapS> window locked up
[20:12] <SpamapS> zul: I brancehd lucid-updates to do the SRU.. is that correct?
[20:12] <zul> SpamapS: yes unless there is a security updates
[20:13] <SpamapS> https://code.launchpad.net/ubuntu/lucid/+source/php5
[20:13] <SpamapS> I see not security
[20:13] <seicherlbob> SpamapS: yes, but the device-names got shuffled and at boottime, i could not see any output. now i re-configure it with UUIDs for the participating devices, so whenever i change the devicenames, it doesnt care.
[20:14] <SpamapS> seicherlbob: I guess what I'm saying is.. simply by adding "/dev/sda1" to a RAID device, it assigns a UUID, and writes a raid superblock saying "RAID with UUID=xxxx has members with UUID=aaaaa and UUID=bbbbb"
[20:14] <SpamapS> seicherlbob: so, you don't need to be explicit about the UUID's, its all in the superblock.
[20:15] <seicherlbob> SpamapS: isnt mdadm doing that when i configure it to do it? i dont know what happend but obviously it was complaining about wrong devices and so couldnt start the raid.
[20:16] <SpamapS> seicherlbob: you can verify this with mdadm --detail /dev/mdX
[20:16] <seicherlbob> SpamapS: anyway, i will just define the arrays members by uuid and thats it
[20:16] <SpamapS> seicherlbob: you can also do   mdadm --query /dev/sdX#
[20:17] <seicherlbob> SpamapS: now i see the dev/sd[bd]1 devices as members. but they got changed! so i will define them with uuid so that can no longer happen
[20:19] <pmatulis> Wyleyrabbit: it's not something you're supposed to do. what's wrong with the driver in lucid?
[20:19] <SpamapS> seicherlbob: I'm fairly certain that they already have UUID's, and its just showing you the device names.. but maybe you created it a weird way that I've never seen.
[20:20] <seicherlbob> SpamapS: I followed the howto at http://www.linuxhomenetworking.com/wiki/index.php/Quick_HOWTO_:_Ch26_:_Linux_Software_RAID
[20:20] <Wyleyrabbit> pmatulis, the driver that comes with lucid is version 04.01, whereas the current driver is 04.30 with a list of bug-fixes to go along with it. Are you saying I should just stick with the existing one?
[20:22] <pmatulis> Wyleyrabbit: there will always be improvments (bug fixes).  is it's current state preventing you from using it?
[20:22] <SpamapS> seicherlbob: this HOWTO makes an incorrect statement.. "Your system doesn't automatically remember all the component partitions of your RAID set." .. thats just not true.
[20:23] <seicherlbob> SpamapS: so what shall i do?
[20:23] <SpamapS> seicherlbob: it does in fact remember, and will re-assemble as long as all of the partitions are marked as type "FD" or "Linux raid autodetect"
[20:23] <SpamapS> seicherlbob: Is your raid broken in any way?
[20:23] <SpamapS> seicherlbob: maybe paste the warnings you're seeing
[20:23] <seicherlbob> SpamapS: no. I had one running and attached another one (2 discs) from my old server
[20:24] <Wyleyrabbit> pmatulis, hmmm, not that I can see. My understanding is the newer driver might offer better performance or something. I'm not generally a "bleeding edge" kinda guy, but I do like things reasonably current.
[20:24] <pmatulis> SpamapS: mdadm scans the system for raid sets during boot.  maybe that's what the howto is getting at
[20:25] <pmatulis> Wyleyrabbit: that's understandable.  but, really, i would just use what's available as well as possibly test stuff in 10.10
[20:26] <seicherlbob> SpamapS: the only syslog i got from the broken boot is: mdadm[1581]: DeviceDisappeared event detected on md device /dev/md0
[20:26] <SpamapS> pmatulis: all of the information needed to re-assemble an mdadm created RAID set is contained on the raid partitions themselves.
[20:26] <qman__> yeah, you can even reassemble live, no configuration
[20:26] <Wyleyrabbit> pmatulis, 10.10??? I don't see that available yet. Is it?
[20:26] <SpamapS> pmatulis: so saying that your system doesn't automatically remember all the component partitions of your RAID set is quite wrong.
[20:27] <seicherlbob> ok, guys. so what shall i do? do i need a mdadm.conf with the raid definitions or not? and if, what shall i define there?
[20:27] <qman__> well, strictly speaking, the system doesn't, the RAID itself does
[20:27] <SpamapS> Now, it may not remember which RAID is /dev/md0, and which is /dev/md1 .. but *that* is where you use the UUID of the filesystem to make sure the same volumes are mounted as /, /home, etc., each time.
[20:27] <qman__> and the system auto scans it on boot
[20:27] <pmatulis> qman__: yeah, it's semantics
[20:28] <SpamapS> qman__: lol, system is a pretty broad term... good point.
[20:28] <seicherlbob> SpamapS: the UUIDs of the raid as such. yeah. but the members of the raid need no configuration?
[20:28] <SpamapS> seicherlbob: mdadm writes the configuration to the members.
[20:29] <pmatulis> seicherlbob: you can use /etc/mdadm.conf to make it elegant (self documenting too)
[20:29] <Wyleyrabbit> pmatulis, I guess what would be more important to me than having the current "bleeding edge" driver is having the ability to talk to the RAID card to find out if a drive has failed, etc. I'm shocked that Ubuntu comes with the driver, but no utility to know the status of the array health.
[20:29] <seicherlbob> so i just remove the raid-definition from the mdadm.conf and let mdadm do its job alone?
[20:29] <SpamapS> seicherlbob: I'm not sure what you even want. It sounds like maybe one of your raid partitions got wiped out? or just disconnected briefly?
[20:29] <qman__> mdadm.conf is generated by scanning the RAID
[20:29]  * SpamapS is going cross eyed
[20:29] <qman__> I have one because I remember running into some trouble after adding disks to my array
[20:29] <seicherlbob> SpamapS: no, none was removed. another raid was added, but then the devicenames were incorrect.
[20:30] <pmatulis> qman__: unless it already exists?
[20:30] <seicherlbob> SpamapS: so mdadm "thought" that the devices was removed, because sdb1 was no longer part of the raid it used to be.
[20:30] <qman__> pmatulis, not automatically generated, you create it by doing something like mdadm --scan > /etc/mdadm.conf
[20:30] <qman__> I forget the exact switches
[20:31] <pmatulis> qman__: sure, but it's not required
[20:31] <qman__> seicherlbob, mdadm doesn't care about "sda" or "sdb", it only cares about the UUIDs saved on the disks themselves
[20:31] <seicherlbob> qman__: I'll just try it now. we will see.
[20:32] <pmatulis> Wyleyrabbit: with s/w raid there is no raid card
[20:33] <pmatulis> Wyleyrabbit: and you can query the array with mdadm
[20:33] <Wyleyrabbit> pmatulis, right. I have a hardware raid card though, and absolutely must be able to know the array health
[20:33] <qman__> Wyleyrabbit, mdadm has its own system to determine failure, and you can use smartctl to check the disks individually
[20:33] <pmatulis> Wyleyrabbit: ok, i thought you were using s/w
[20:33] <qman__> ah
[20:34] <qman__> well, that utility would be up to the vendor
[20:34] <qman__> since it would be specific to your hardware
[20:34] <Wyleyrabbit> yeah, no. I have one of these: Intel Raid RS2BL080 with batter backup.
[20:34] <Wyleyrabbit> http://www.intel.com/products/server/raid-controllers/RS2BL080/RS2BL080-overview.htm
[20:34] <pmatulis> Wyleyrabbit: 10.10 alpha 3 is released, yes, you can run it right now
[20:35] <Wyleyrabbit> ummm, that would be a "battery backup"
[20:36] <Wyleyrabbit> so anyways, back to what I was saying. It's great that Ubuntu includes the driver to talk to these LSI cards out of the box, but nothing to query the array health. The driver doesn't even write to any of the log files, from what I can tell.
[20:37] <qman__> drivers don't typically write to log files, they just send messages that would show up in dmesg
[20:37] <seicherlbob> well, mount complaints that there are no devices with these UUIDs (i used the ones, mdadm told me). but i have 2 others. what is this?
[20:37] <qman__> which you could then log and parse
[20:38] <Wyleyrabbit> Although my card is made by Intel, it's really an LSI card. Much in the same way, I suppose, that one can buy an ASUS video card, but it's really an ATI or Nvidia card.
[20:44] <seicherlbob> why does mdadms UUID not fit with dev/disk/by-uuid ?? the uuids there are different to the ones mdadm gives them. I only find the same IDs in dev/disk/by-id with a "md" prefix
[20:45] <Wyleyrabbit> qman__, the only message in dmesg about megaraid_sas is the IRQ it's on and something about the latency timer setting. Currently the array is building (therefore degraded) so there should be tons of messages about this, but there's nada.
[20:48] <asanir> I'm trying to find an iso suitable for my old intel celeron architecture, some ppl recommends me to get alternate package, should I find it here: http://cdimage.ubuntu.com/ubuntu-server/ ??
[20:49] <qman__> asanir, any "i386" version of ubuntu will work
[20:50] <asanir> but it fails on my system
[20:50] <asanir> seems to have problem with linux-generic-pae
[20:51] <qman__> well, you could try installing the linux-386, but a celeron is i686 and should work with the generic kernel
[20:52] <qman__> I have a K6 which is an i586 and have to use the linux-386 kernel
[20:52] <asanir> could u pls tell me how to use linux-386 instead of default installer?
[20:53] <Jordan_U> asanir: What exactly happens when you try to use the Ubuntu installer?
[20:53] <qman__> what I did was let it install the linux-generic kernel with the system, then reboot with the live CD, choose the "shell on your system" option, and install the linux-386 package
[20:54] <qman__> with apt-get
[20:56] <seicherlbob> SpamapS: ok, now i restarted without any new lines in the config and mdadm misses 1 device of each array. wtf?!
[21:00] <seicherlbob> SpamapS: i will just configure it by IDs, i dont care about how it should work. damn it.
[21:07] <seicherlbob> perfect. the members of the array dont have UUIDs set in dev. thats great! what am i doing wrong?
[21:08]  * seicherlbob was sarcastic
[21:13] <alvin> seicherlbob: bug 27037 I have to recreate an array on reboot (without data loss)
[21:43] <Wyleyrabbit> Rats.
[21:44] <Wyleyrabbit> I'm about to give up on Ubuntu Server and go to Centos. ARGH! I really really don't want to do that, but I've now blown 8+ hours trying to get ANYTHING to allow me to talk to the hardware raid array. Nothing works.
[21:48] <sherr> Wyleyrabbit: The megaraid cards have a utility to query the RAID, status etc. e.g. MegaCli
[21:48] <ivoks> ?
[21:49] <ivoks> Wyleyrabbit: you'll have the same problem on *any* linux
[21:49] <sherr> See : http://hwraid.le-vert.net/wiki/LSIMegaRAIDSAS
[21:49] <ivoks> 3ware has excelent tools for linux
[21:49] <sherr> It's a bit of a pain though ...
[21:49] <Wyleyrabbit> sherr, right! But I cannot for the life of me get that utility installed so it will talk to the Raid card.
[21:50] <ivoks> there's no installation
[21:50] <ivoks> you just run it
[21:50] <Wyleyrabbit> ivoks and sherr, The MegaCli utility only seems to be available as an RPM. I did manage to get it to run well enough to show me the help file, but nothing else
[21:50] <Wyleyrabbit> it wouldn't show any installed adapters.
[21:51] <sherr> I'm running it on Debian Lenny.
[21:51] <sherr> http://hwraid.le-vert.net/debian
[21:53] <sherr> It's not brabd new h/w though - Dell PE1800 + LSI MRSAS 8888ELP + HP MSA60 array
[21:53] <Wyleyrabbit> sherr, omg! this is awesome. I just might be able to get this going after all. I didn't have ANY .deb files to work from, only RPMs that failed the "alien" process.
[21:53] <sherr> *brand
[21:53] <Wyleyrabbit> is lenny equivalent (or at least very similar) to Ubuntu Server 10.04 LTS ?
[21:54] <sherr> A bit older.
[21:54] <sherr> But just as good. If your hardware is supported (our new Dell R410 at work was not - so it runs 10.04 now)
[21:55] <sherr> (the h/w raid1 disk for the OS was not seen in Lenny)
[21:55] <sherr> Note also - the "megacli" programs are horrible ....
[21:56] <sherr> This is what I have installed :
[21:56] <sherr> megacli 5.00.12-1
[21:56] <sherr> megaclisas-status 0.5
[21:57] <sherr> Good luck - I am off very shortly.
[21:57] <Wyleyrabbit> thanks sherr !!!
[21:57] <Wyleyrabbit> sherr, what driver version are you running?
[21:59] <mathiaz> hggdh: hi!
[21:59] <mathiaz> hggdh: how is the 10.04.1 UEC testing going/
[21:59] <sherr> Not sure. Let me have a quick look ... whatever's in the package above.
[21:59] <mathiaz> hggdh: ?
[22:00] <sherr> Wyleyrabbit: Sorry - this is what is in the dmesg :
[22:00] <sherr> megasas: 00.00.03.20-rc1 Mon. March 10 11:02:31 PDT 2008
[22:00] <hggdh> mathiaz: one gone, one to go
[22:01] <mathiaz> hggdh: great!
[22:01] <hggdh> mathiaz: got a misbehaving server in the middle
[22:01] <mathiaz> hggdh: everything is working well?
[22:01] <Wyleyrabbit> sherr, ok, thanks.
[22:01] <sherr> + Fusion MPT base driver 3.04.06
[22:01] <sherr> Copyright (c) 1999-2007 LSI Corporation
[22:01] <sherr> Fusion MPT SPI Host driver 3.04.06
[22:02] <sherr> (sorry - this is in the distro, not the megacli package ofcourse)
[22:02] <hggdh> mathiaz: so far yes, except for this misbehaving server (which I think I am bypassing the issue)
[22:03] <hggdh> mathiaz: I already marked done the required UEC test
[22:03] <Wyleyrabbit> sherr, that helps.
[22:03] <Wyleyrabbit> I think U.S. 10.04 came with 04.01 (which is also old)
[22:03] <Wyleyrabbit> what version of debian is ubuntu server 10.04 lts most similar to?
[22:03] <hggdh> mathiaz: we need coverage on the other tests
[22:04] <mathiaz> hggdh: I'm working on these
[22:11] <Wyleyrabbit> sherr, SUCCESS!!! Thanks
[22:26] <bogeyd6_> Anyone feel like helping on this? I had a truecrypt drive in linux (encrypted the whole drive) and now Im on windows, how do i access the encrypted data?
[22:27] <Wyleyrabbit> bogeyd6_, should be as simple as telling windows truecrypt to mount the drive
[22:28] <bogeyd6_> Wyleyrabbit, i did but it sets it up as a RAW type and dinwos 7 wont recognize it as a formatted drive
[22:29] <Wyleyrabbit> bogeyd6_, can you assign the drive a letter (in Control Panel - Admin Tools - Create/Format partitions)?
[22:30] <bogeyd6_> Wyleyrabbit, yup, just double clicking on it tells me it needs formatted
[22:31] <Wyleyrabbit> bogeyd6_, hmmm. I'll have to stop there, I don't use Truecrypt much.
[22:31] <Wyleyrabbit> sorry I couldn't be of more help
[22:32] <bogeyd6_> sux
[22:58] <funkyHat> bogeyd6_: which filesystem did you use on the drive?
[22:58] <bogeyd6_> funkyHat, its been too many years to remember, is there a way to find out?
[22:59] <funkyHat> bogeyd6_: yes, boot up Linux and the command mount will tell you
[22:59] <funkyHat> bogeyd6_: Windows can't read most Linux filesystems. There is some 3rd party support for ext2(/3) I think
[23:00] <bogeyd6_> im getting vmware rite now, ill boot up linux and copy my files out
[23:04] <g0rd0n> evening
[23:09] <bogeyd6_> TEAM PlanB
[23:17] <ryanakca> ScottK: Hmm. Yet another day without an i386 ubuntu-server daily. Maybe I'll have better luck tomorrow.
[23:17] <ScottK> OK.
[23:18] <g0rd0n> i'm a bit confused by the numerous virtualization possiblities... i
[23:18] <g0rd0n> i'm only experienced with vserver, but it seems like xen or kvm have more to offer
[23:19] <g0rd0n> now some websites say kvm is better, most say xen is
[23:20] <ScottK> g0rd0n: My recommendation is to look at the Ubuntu Server Guide (URL in /topic) and see what it suggests.  Generally sticking with the supported/documented solution is a safe approach in Ubuntu.
[23:20] <failover> g0rd0n, http://itmanagement.earthweb.com/osrc/article.php/3898441/Linux-KVM-Eyes-World-Domination.htm
[23:21] <g0rd0n> world domination lol
[23:22] <failover> =]
[23:24] <g0rd0n> yeah i will probably try kvm with command line tools then
[23:45] <g0rd0n> is there something like debian volatile for ubuntu server?