[00:00] <Masshuu> http://pastebin.com/G0fCNr03
[00:00] <Masshuu> uuh now what
[00:10] <Psi-Jack> I understand that OpenVZ support has been long since removed from Ubuntu since after Ubuntu 8.04, but word around says it was being replaced with LXC. Does Ubuntu 10.04 have LXC support to run LXC containers?
[00:21] <JanC> Psi-Jack: I haven't tried it, and I don't think there is official support, but I suppose it has the necessary bits for LXC in the kernel and in userspace
[00:22] <JanC> maybe somebody else knows more about it
[01:12] <sdfoj8>  hello gys, sorry for disturbing, i am trying to accomplish a project, a file storage system like rapidshare . what clustering open source software do you recommend me ? so i can setup my private cloud ? otherwise, i have to store files distributed to servers but those servers wont be able to handle huge traffic. so i need to make it clustered.
[01:14] <jmarsden> sdfoj8: Maybe read https://help.ubuntu.com/community/UEC
[01:15]  * sdfoj8 hello gys, sorry for disturbing, i am trying to accomplish a project, a file storage system like rapidshare . what clustering open source software do you recommend me ? so i can setup my private cloud ? otherwise, i have to store files distributed to servers but those servers wont be able to handle huge traffic. so i need to make it clustered.
[01:16] <sdfoj8> jmarsden
[01:16] <sdfoj8> instead of ubuntu itself
[01:16] <sdfoj8> i want to install a cluster version
[01:16] <jmarsden> ?  UEC is developed by and part of Ubuntu...
[01:16] <sdfoj8> to be honest
[01:16] <jmarsden> Did you actually read the info?
[01:16] <sdfoj8> when i see
[01:17] <sdfoj8> "ec2" somewhere
[01:17] <sdfoj8> i run away from it
[01:17] <sdfoj8> and what i got from ubuntu-cloud site was ec2
[01:18] <jmarsden> So you want a cloud, but not EC2.  Ubuntu has chosen EC2 compatibility, so that means you are on your own, basically.
[01:18] <sdfoj8> why do i need amazon services if i want to setup my own cloud
[01:18] <sdfoj8> didnt get the point
[01:18] <jmarsden> You don't.  UEC runs on your own computers and is 100% free.
[01:18] <sdfoj8> cloud world is not based on amazon on amazon s3/ec2
[01:19] <sdfoj8> so why i see amazon in that site ?
[01:19] <jmarsden> It is your PRIVATE cloud and does not need or use Amazon anything.  It just uses cpmpatible protocols
[01:19] <sdfoj8> https://help.ubuntu.com/community/UEC/PackageInstallSeparate
[01:19] <jmarsden> Because you are not reading carefully enough.
[01:19] <jmarsden> Read, instead of skimming for keywords.
[01:19] <sdfoj8> hmm i saw it now.
[01:19] <sdfoj8> jmarsden
[01:19] <sdfoj8> i am realyl confused
[01:20] <sdfoj8> next semester i am going to take distributed systems course
[01:20] <sdfoj8> and i wish i learn more about cloud thing
[01:20] <sdfoj8> what comes strage to me is, there has been never a really working open source cloud service
[01:20] <sdfoj8> i saw open source projects like glusterfs and others
[01:20] <sdfoj8> but those are not cloud as i know.
[01:20] <jmarsden> There is now, it is called UEC.  WHich is why I pointed you at it.
[01:20] <sdfoj8> they dont have automatic intelligen replication
[01:21] <sdfoj8> do you know what i am talkn about ?
[01:21] <sdfoj8> suppose, i got a video on 124123servernode.mysite.com/stream/birthday.avi
[01:21] <sdfoj8> and i am sending this video to my friends
[01:22] <sdfoj8> and suddenly 10000 of my friends want to watch hat video
[01:22] <sdfoj8> and they open the url, but since the server has 1gbit, which means 12.5MB per sec. and each user will get only 12.5KB per sec.
[01:22] <sdfoj8> 12.5Kb per sc.
[01:22] <sdfoj8> KB i mean.
[01:23] <sdfoj8> so the server should understand and track the files, if there is a hude demand to a file, then it should replicate that file through other nodes before streaming it
[01:23] <sdfoj8> and it shouldnt replicate all the files as well.
[01:24] <sdfoj8> so jmarsden what do you say
[01:28] <sdfoj8> hello ?
[01:42] <sdfoj8> anyone can talk about UEC ?
[02:03] <MTecknology> !enter > sdfoj8
[02:04] <sdfoj8> MTecknology
[02:04] <sdfoj8> sorr.
[02:06] <diago> does a drive have to have a partition? I mean, can I just mkfs.ext3 /dev/sda
[02:06] <diago> It seems to work fine
[02:06] <MTecknology> diago: it works - but it's easier to break things if you forget that you put an fs on the whole drive
[02:07] <sdfoj8> anyone has experience with glusterfs, mogilefs ?
[02:07] <MTecknology> !ask | sdfoj8
[02:09] <MTecknology> diago: in unix there's a drive, partition, and slice - that gets fun..
[02:10] <sdfoj8> anyone has experience with glusterfs, mogilefs ? can u compare them?
[02:11] <diago> MTecknology: my thoughts: It's a logical volume for a virt guest
[02:12] <diago> I could then lvextend and resize2fs when I need to increase
[02:12] <diago> if I start adding partitions I'm not sure how this will work
[02:12] <diago> if that makes any since
[02:14] <MTecknology> sounds like you're going to have a heck of an interesting time oif you do that.. why not make the drive use lvm, then put a volume on that, then when you need to grow what the guest has, you can follow any of the millions of guides out there?
[02:15] <MTecknology> let the guest use that partition which can easily grow as opposed to the whole drive which you can't grow
[02:15] <diago> I don't think I explained it right...
[02:17] <MTecknology> diago: I may have just understood it wrong - i think if it works for you then there's no issues with doing it that way :P - yay linux
[02:17] <diago> I just tested and it's pretty easy.  shut the virt down, on the KVM machine do lvextend -L+5g /dev/vol/drive
[02:18] <diago> start the virt and resize2fs /dev/vda
[02:18] <diago> BAM!
[02:18] <diago> awesome... I love linux
[02:20] <MTecknology> diago: :P
[02:21] <MTecknology> diago: you might actually be able to grow before even shutting down and just not be able to use the additional space until you reboot
[02:21] <diago> You're probably right, LVM is awesome... over iSCSI at that!
[02:22] <MTecknology> :)
[03:02] <deslector> hi, I'm trying to use smb_auth with squid on lucid server but it is not working
[03:02] <deslector> it used to work with the same config on a debian etch system
[03:02] <deslector> the original page of the smb_auth author doesn't exist anymor
[03:06] <deslector> any ideas?
[03:57] <mcarse> I am trying to get my drobo with multiple luns running again in Ubuntu 10.04, apparently I need to load scsi_mod, but when I try "modprobe scsi_mod" i get "FATAL: Module scsi_mod not found" where do I find it?
[04:03] <mcarse> any thoughts?
[04:24] <tyler_d|werkin> whats a 64 bit alternative to linux-generic?
[06:10] <amstan> hey guys
[06:10] <amstan> i'm having some speed issues with my ssh
[06:11] <amstan> i recently upgraded my network to gigabit, i can download stuff from the http server at 60MB/s, but any sftp transfer to the same server is still at 3MB/s
[06:12] <amstan> how could i speed up the ssh connections?
[06:12] <pwnguin> amstan: can you download via https?
[06:12] <amstan> well.. i would also like to upload
[06:12] <amstan> it's a home server
[06:12] <pwnguin> ssh has encryption which slows thing down
[06:12] <amstan> ok, how can i disable it?
[06:13] <amstan> like.. an ondemand thing, i don't care about encryption within my lan
[06:13] <pwnguin> maybe there's an identity cipher
[06:13] <pwnguin> but this is just a theory without evidence
[06:14] <pwnguin> if you monitor CPU usage durting ssh transfer
[06:14] <amstan> ok, on it
[06:14] <pwnguin> my theory might be validated if your CPU spikes
[06:16] <amstan> yep, my server is getting 100% cpu
[06:16] <amstan> on one of the cores
[06:18] <pwnguin> ideas
[06:18] <pwnguin> 1 turn off compression
[06:19] <pwnguin> 2 weaken the cipher
[06:23] <pwnguin> http://serverfault.com/questions/125610/is-there-a-way-to-make-scp-run-faster-on-a-mac-os-x
[06:23] <pwnguin> amstan: one last idea might be to try out webdav
[06:24] <amstan> pwnguin: http://serverfault.com/questions/116875/how-can-i-disable-encryption-on-openssh
[06:24] <amstan> pwnguin: see that first post? what file is that to be edited?
[06:24] <joebob> How can I get grub2 on two raid 1 drives. It's only on one currently...
[06:58] <Callum__> joebob: so that is a single RAID 1 array consisting of two drives?
[07:19] <ChrisRut> When might Kernel 2.6.34 be available for Lucid?
[09:01] <jo-erlend> I setup a server yesterday. It has four disks, which I setup in a raid5. I used two partitions; one for swap and one for root. The installation seemed to go well, but when I rebooted, it dropped to an initramfs, with an error that the disk with a certain uuid wasn't available.
[09:01] <jo-erlend> I'm fairly certain that I did everything correct. Is this a known bug in lucid server? It's 64bit btw.
[09:02] <jo-erlend> how do I fix it?
[10:18] <jo-erlend> It's really annoying that ubuntu server doesn't display the grub menu by default. I have a USB keyboard, and it seems it's too slow to load that I can use shift to display the menu. And since I can't boot without changing the parameters, I'm having difficulties.
[10:50] <jo-erlend> I think I'll install karmic server instead, and attempt an upgrade afterwards.
[11:05] <ruben23> hi guys im installing a ubuntu server with 1 terabyte HDD, is it ok to set on default one partition all--> 999.9Gb this wont affect performance..?
[11:08] <ruben23> any idea guys
[11:41] <larsemil> ruben23: why do you want to do that?
[11:42] <ruben23> larsemil:i just a 1 ter byte hdd, i ahve install just a default, just want to know if it will affect server perfomance.
[11:42] <larsemil> ruben23: i would say that 99% of the time you would want to seperate root partition and datapartition.
[11:43] <ruben23> hmm ok
[11:43] <ruben23> for backup purpose right..?
[11:44] <ruben23> any suggested partion scheme..?
[11:44] <ruben23> i mean the division of hdd, i dont have in particular, any deafult settings can do
[11:45] <larsemil> ruben23: i always use something like /boot 200mb /10-15gb and then depending on use /usr or /home for alot of data. i like using /home. also if it is for files sometimes i just use a /stuff that can be shared in the network.
[12:36] <jo-erlend_> dammit... Karmic has the same problems. I'm installing from a USB stick. Might that be the cause?
[13:27] <RoyK> http://pastebin.com/mUmPu6Sw <-- I get these errors from my vm
[13:37] <jo-erlend_> this was the tenth time in two days I've installed Ubuntu 10.04 Server 64bit without any success. I've setup software raid and formatted using a swap area and an ext4 partition. When I reboot after the installation, it complains about a disk with a uuid doesn't exist.
[13:38] <jo-erlend_> I then check /dev/disk, and find that there are no /dev/disk/by-uuid directory at all.
[13:42] <ruben23>  hi guys, how fo i configure my network on my ubuntu server having a PPPoE DSL connection but its not username/pass , it automatic IP detection form the ISP...any suggestion..?
[13:45] <jo-erlend_> I think there is an application called pppoeconfig or something
[13:46] <jo-erlend_> pppoeconf
[13:49] <ruben23>  jo-erlend_: sorry im wrong, the connecyion of my DSL is auto assigneing of publci IP so it like DHCP.
[13:50] <ruben23> my setup is like this : internet---->modem---->ubuntu server
[13:50] <jo-erlend_> that shouldn't have anything to do with pppoe. You'll configure your interface to use dhcp in /etc/network/interfaces.
[13:51] <ruben23>  jo-erlend_: yes i have set it to DHCP but the porblem, my linux server get this IP address, 192.168.1.10 (whihc is a private IP) why not a publci IP..?
[13:52] <jo-erlend_> then you have router in your network. Your router gets the public ip and has its own dhcp server which your server obtains an IP from.
[13:52] <ruben23>  jo-erlend_: why is it preivate IP not public..?
[13:53] <ruben23>  jo-erlend_: so your saying that my modem is in a router mode not on bridge mode..?
[13:53] <jo-erlend_> looks like it, yes.
[13:54] <ruben23> jo-erlend_: how do i make my linux server, be accessible with SSH..? over the internet..?
[13:55] <jo-erlend_> by setting up port forwarding in your router.
[13:55] <ruben23>  jo-erlend_: hmmmm but its a modem not a router, which my ISP owned it..
[13:55] <ruben23> i cant touch..
[13:56] <jo-erlend_> then you'll have to talk to them. It doesn't have anything to do with Ubuntu.
[13:57] <ruben23> jo-erlend_: thanks, for pointing mer to a clearer path now..
[13:59] <ruben23>  jo-erlend_: i can used DDNS tobe able to connect always even do the Public IP changes right..?
[14:07] <ruben23> hi
[14:28] <jo-erlend_> ruben23, it can be set to update a domain name, yes. It has nothing to do with actual connections.
[14:28] <jo-erlend_> is it possible that ubuntu cannot boot from a software raid?
[16:06] <ruben23> hi what is the preempt kernel of the ne ubuntu-server 10.04
[16:34] <Guest50155> how can I grow ext3 filesystem on a raid device? I've upgraded from 320GB to 1.5TB raid 1
[16:43] <jo-erlend_> if you have free space on the disks, then you should be able to use resize2fs as usual, I think?
[16:44] <jo-erlend_> you'll have to resize the partition first, of course. If you're using lvm, that's easy. If you're not, it's not so easy.
[16:45] <jo-erlend_> but a raid1 cannot be bigger than your smallest disk. Perhaps you should provide more details?
[16:46] <Guest50155> The partitions are already bigger. running resize2fs /dev/md1 now
[16:47] <Guest50155> it wasn't too bad to upgrade drives individually without lvm. i've never used lvm and it scares me to try to start ;)
[16:48] <Guest50155> jo-erlend_, working thanks!
[16:48] <Guest50155> another issue I have is my cdrom drive just never mounts properly and if I run mount /media/cdrom0 the light never comes on and after about 20 seconds I get mount: no medium found on /dev/sr0
[16:48] <Guest50155> other people have had the same cd issue as me: http://ubuntuforums.org/showthread.php?t=1235151
[16:49] <jo-erlend_> I have no idea about that as I stopped using cd and dvd a while ago.
[16:49] <Guest50155> yeah me too but it's not my box...
[16:53] <sdfoj8> what is the command line that i can test my speed with speedtest.net do you know ?
[17:00] <jo-erlend_> what amazes me, is that nobody in the entire world seems to know how to install ubuntu to boot from a raid partition.
[17:04] <jo-erlend_> but in order to preserve my sanity, I think I'll leave it for today and have another go tomorrow.
[17:18] <RoyK> hi all. I keep getting hangs on an ubuntu guest (ubuntu lucid kvm host, ubuntu lucid guest)
[18:00] <RoyK> this bug is a bitch https://bugs.launchpad.net/ubuntu/+source/linux/+bug/579276
[18:01] <RoyK> "medium" my arse - my boxes hang after a few minutes with normal nfs  traffic
[21:06] <ChrisRut> When might Kernel 2.6.34 be available for Lucid?
[21:08] <_ruben> ChrisRut: you could the mainline kernel packages provided by the kernel team, ubuntu releases usualy don't get new upstream kernels, just (security) backports
[21:09] <ChrisRut> _ruben: is there a PPA i can use to get 2.6.34 on lucid? or would I have to build it manually?
[21:10] <_ruben> http://kernel.ubuntu.com/~kernel-ppa/mainline/
[21:22] <ChrisRut> _ruben: Thanks
[21:45] <chewbranca> so I'm trying to get eucalyptus up and running with UEC, I'm using the CDInstall guide on the ubuntu help site for UEC, I've verified that my node is registering, but when I do euca-describe-availability-zones verbose, it shows 0000 for all free/max options
[21:45] <chewbranca> any ideas?
[21:48] <sdfoj8> chewbranca
[21:48] <sdfoj8> ca i ask why you use eucalyptus with uec ?
[21:48] <sdfoj8> why dont you use normal ubuntu with mogilfs ?
[21:50] <chewbranca> because I'm trying to build a eucalyptus based EC2 cluster
[21:50] <sdfoj8> on amazon ?
[21:51] <sdfoj8> may i ask why do you do that ?
[21:51] <chewbranca> why not?
[21:51] <sdfoj8> why dont you build in other hosting ? and you do it on ec2 ? i am little confused
[21:51] <sdfoj8> coz ec2 is already a cluster
[21:51] <sdfoj8> and you are building a cloud server on a cloud already ?
[21:51] <chewbranca> I need to use EC2, and I want a local dev environment, eucalyptus seems like a good way to go
[21:51] <chewbranca> no on a local box
[21:52] <sdfoj8> so is it like that
[21:52] <sdfoj8> you got 3-4 instances with amazon ec2
[21:52] <MTecknology> how can I specify no swap with vmbuilder?
[21:52] <sdfoj8> but when you connect to one server, all the servers are same ?
[21:52] <sdfoj8> ?
[21:53] <chewbranca> I'm not even using this with EC2 yet, I've got a master controller and a node server here at home, trying to get them up and running, but the node server is not showing up in the availability zone
[21:53] <sdfoj8> sorry bo bother but i am asking to learn
[21:53] <sdfoj8> why do do that.
[21:53] <sdfoj8> when you install UEC what will be the difference
[21:54] <chewbranca> the difference of what?
[21:54] <chewbranca> I'm just trying to build a dev environment that emulates EC2, I don't want to have to pay for everything I do on EC2
[21:54] <MTecknology> sdfoj8: since you're back - I wanted to mention that it's considering spamming if you ask the same question in many channels
[21:54] <chewbranca> I've got a lot of work to do with deploy management and provisioning, so I want a local version I can use to test
[21:55] <sdfoj8> MTecknology i dont know but i think you are tracking me :) i didnt ask same quest.
[21:55] <sdfoj8> i mean what if you install normal ubuntu server instead of that
[21:56] <MTecknology> sdfoj8: last time you were asking questions you were asking the same thing in #ubuntu, #ubuntu-server, #red5, and a couple others.
[21:56] <chewbranca> sdfoj8, if you want to install normal ubuntu server go for it, I'm trying to build a local version of EC2, not understanding what is complicated about understanding that
[21:57] <MTecknology> chewbranca: I can give you a pretty nifty command I use for creating vm's
[21:57] <sdfoj8> i did it one time MTecknology
[21:57] <chewbranca> MTecknology, nice, that would be cool
[21:57] <MTecknology> chewbranca: http://paste.ubuntu.com/449355/
[21:58] <MTecknology> sdfoj8: I was just informing you that you should not do that.
[21:58] <sdfoj8> ok.
[21:58] <chewbranca> MTecknology, oh niiiiiiiiiiiiiiice, that's slick
[21:59] <sdfoj8> using vm is doing and sharing the cpu right ? using virtual boxes on a same machine.
[21:59] <sdfoj8> vps
[21:59] <MTecknology> chewbranca: I modify it a little different almost every time I use it - but that's the base I use
[21:59] <chewbranca> MTecknology, yeah that's definitely handy
[21:59] <MTecknology> glad you enjoy :)
[22:01] <chewbranca> MTecknology, cool, just saved that
[22:02] <MTecknology> chewbranca: oh! word of caution- be careful with allocating swap. I had two vm's start hammering swap, got into swap death, vm's couldn't get over 1MB/s disk read
[22:02] <MTecknology> I just dealt with that last week
[22:02] <chewbranca> so I'm definitely having some trouble figuring out the debugging path with UEC and eucalyptus, I mean I've got the cloud controller and node up and running, but the node isn't showing up in the availability zones, and I just don't know where to go next
[22:03] <chewbranca> MTecknology, oh ouch, that's rough, did you just allocate too much swap? or was the vm using too much ram and just hitting swap?
[22:03] <MTecknology> The solution was to add a little ram to each vm, and remove its swap. Simple solution that brought disk read from <1MB/s to >120MB/s
[22:03] <MTecknology> it was hammering swap too hard
[22:04] <MTecknology> I'm gonna head to do homework though
[22:04] <chewbranca> yeah hitting swap is rarely good
[22:04] <MTecknology> chewbranca: best of luck - blog and link me when you finish that
[22:04] <chewbranca> ok cool, thanks for sharing that provisioning script
[22:04] <chewbranca> MTecknology, will do!
[22:31] <JackTO> hey all
[22:31] <JackTO> I have a remote KVM, and its an older Avocent Witchview, that doesn't support the latest refresh rates.   Even though I'm not running any X/Gui, the default refresh rate seems to be too high, how do i change it from the CLI?
[22:42] <qman__> I'm pretty sure the console is locked at 60Hz
[22:42] <JackTO> hmmm... someon told me to disable a framebuffer?
[22:44] <qman__> I'm also fairly sure that would require recompiling the kernel
[22:45] <qman__> you could try removing 'quiet splash' from the boot line, see if that solves your problem
[22:45] <JackTO> what does that do?
[22:45] <qman__> unhides the kernel messages
[22:45] <JackTO> like in recovery mode?
[22:45] <qman__> what version of ubuntu is it?
[22:45] <qman__> yeah, only without starting the recovery process
[22:46] <JackTO> hmmm... let me see if even the BIOS hsow up ok, through the KVM, b4 ubuntu starts...
[22:46] <JackTO> maybe its an issue with the graphics card
[22:46] <JackTO> let me see
[22:48] <qman__> I don't really get why they do that on ubuntu server anyway
[22:48] <qman__> I can see wanting to hide that stuff for desktops or whatever
[22:48] <qman__> but on my server, I want to know what's going on
[22:49] <JackTO> darn, i get flashing, eratic lines even b4 grusg load...  the bios shows up all messed up
[22:49] <JackTO> i agree, why do they hide the boot process, so odd... anyone using a server wants to see whats going on
[22:49] <JackTO> gurb*
[22:49] <qman__> I upgraded one of my servers to lucid due to a hardware issue
[22:50] <qman__> and now it has a boot splash screen
[22:50] <qman__> yech
[22:50] <chewbranca> wow... finally fixed the issues with getting eucalyptus working proplery thanks to this thread: http://open.eucalyptus.com/forum/problem-node-registration
[22:50] <chewbranca> had the exact same issues
[22:52] <JackTO> is there a difference in VGA signals from 2003 and now.. isn't vga, vga...   my Avocent KVM used to work perfectly... now all i get is flashing lines
[22:52] <qman__> ugh, nvidia.com now has a flash intro page on it
[22:52] <qman__> what is wrong with these people
[22:57] <qman__> JackTO, 2003 isn't that old
[22:57] <qman__> it should 'just work' even with the framebuffer
[22:58] <qman__> linux framebuffers only support VGA, SVGA type resolutions
[22:58] <qman__> no widescreen, no special refresh rates
[23:06] <soren> bc: It's the (old) ubottu code.
[23:21] <MTecknology> I create a disk with "qemu-img create -f qcow2 data1.qcow2 100" Then I added this to my libvirt xml  <disk type='file' device='disk'><source file='/virt/ext/images/pifio/data1.qcow2'/><target dev='hdb' bus='ide'/></disk>  Then defined it and rebooted the vm. I can see /dev/sdb in trhe vm but when I do fdisk /dev/sdb I get Unable to read /dev/sdb  -  Any ideas what I'm doing wrong?
[23:25] <jo-erlend_> MTecknology, libvirt has a channel on oftc. kvm has a channel on this network, and Ubuntu has #Ubuntu-virt, which is supposed to handle these kinds of questions.
[23:26] <MTecknology> jo-erlend_: a new channel - thanks
[23:30] <stork> hi, i'm still running gutsy on my server, what's the best way to upgrade it?
[23:32] <soren> hastily
[23:32] <stork> indeed :]
[23:33] <stork> will i need to do a clean reinstall? i don't mind but i do have a lot of stuff on a RAID array that would be a pain in the arse if i lost it
[23:34] <soren> You should never need to reinstall an ubuntu system.
[23:34] <MTecknology> stork: I would assume sudo aptitude update && sudo aptitude full-upgrade && reboot && do-release-upgrade  <- should work fine
[23:34] <chewbranca> how long does it normally take for a new instance to go from pending to running?
[23:34] <stork> no luck i don't think, gutsy isn't even supported anymore
[23:35] <jo-erlend_> how do you boot a software raid array if the first disk fails?
[23:35] <stork> none of my repos work
[23:35] <jo-erlend_> soren, you're an advanced one. If you know this, please let me know so I can sleep! :)
[23:37] <jo-erlend_> setting up / on raid5 doesn't work, but I've read that's because grub2 doesn't support booting from a raid5. I've also setup a raid1 for /boot, using four disks, and that doesn't work either. People are telling me I cannot use raid for the boot device, but this cannot be the case?
[23:37] <qman__> gutsy is too old, the repositories have been down a long time
[23:37] <jo-erlend_> gutsy is dead.
[23:37] <soren> jo-erlend_: raid1 for /boot is fine and supported.
[23:37] <qman__> you will probably have to reinstall from scratch
[23:37] <soren> No.
[23:37] <soren> 22:34 < soren> You should never need to reinstall an ubuntu system.
[23:37] <jo-erlend_> no, you can upgrade.
[23:38] <jo-erlend_> soren, what about raid1 with more than two disks?
[23:38] <soren> jo-erlend_: Doesn't matter.
[23:38] <soren> jo-erlend_: Not sure why you want to, but it's perfectly valid.
[23:40] <jo-erlend_> soren, I have four hotswappable disks. If either fails, I'd like to be able to just replace it, resync and feel calm. I have four disks and they have lots of space. Why would I not want to spend a few hundred megabytes on giving me extra redundancy?
[23:40] <soren> Why stop there?
[23:41] <soren> You could have an extra partition on each disk to mirror the /boot partition.
[23:41] <soren> It's just a few hundred MBs, after all.
[23:41] <jo-erlend_> what would you do?
[23:41] <soren> Precisely.
[23:42] <jo-erlend_> ?
[23:42] <soren> I'm using your own argument against you.
[23:42] <soren> See?
[23:42] <jo-erlend_> I don't understand. People are telling me to make one disk responsible for boot, but that would mean if one disk fails, the system cannot reboot.
[23:43] <jo-erlend_> I want at least _some_ redundancy.
[23:43] <soren> Sure, sure.
[23:43] <soren> That's fine. I get that.
[23:43] <qman__> I wouldn't do that, that's quite a waste
[23:43] <soren> i'm just not sure why you feel you need 4 copies.
[23:43] <jo-erlend_> sommer, the disks are equal in size. Why would I leave unpartitioned space instead of using it for extra redundancy for /boot?
[23:44] <qman__> use it as swap instead
[23:44] <jo-erlend_> no, swap is partitioned on the raid5.
[23:44] <jo-erlend_> I really don't want swap without redundancy either.
[23:46] <JackTO> Question:  Is there any benifit of running VNC os a server?  or is SSH sufficient?
[23:46] <JackTO> on*
[23:46] <jo-erlend_> JackTO, absolutely.
[23:46] <qman__> only if you run X
[23:46] <jo-erlend_> no.
[23:47] <soren> jo-erlend_: I think I've missed what your question is. Sorry :)(
[23:47] <soren> jo-erlend_: Yes, mirroring /boot is supported.
[23:47] <soren> jo-erlend_: even across eleventy disks.
[23:47] <JackTO> jo: you are saying SSH is sufficient?
[23:47] <jo-erlend_> soren, I need a redundant /boot partition. I've tried setting it up on RAID5 and I've tried setting it on on RAID1 (though only using four disks). Nothing works.
[23:48] <soren> jo-erlend_: raid5 will not work.
[23:48] <qman__> JackTO, I wouldn't run VNC on a server unless I needed X, and even then, it would only allow connections tunneled through SSH
[23:48] <soren> jo-erlend_: raid1 should.
[23:48] <jo-erlend_> I eventually found out.
[23:48] <jo-erlend_> soren, it doesn't. Not with four disks, at least.
[23:48] <soren> jo-erlend_: I wan't work with "nothing works".
[23:48] <JackTO> qman: thanks....
[23:49] <soren> jo-erlend_: My only useful response to that is "don't do it wrong".
[23:49] <jo-erlend_> soren, can you please have a look at this page and tell me what I'm doing wrong? https://help.ubuntu.com/10.04/serverguide/C/advanced-installation.html
[23:50] <soren> Besides, it's waaaay past my bed time. Feel free to catch me either during (European) business hours tomorrow or between 1900 and 2100 UTC.
[23:50] <soren> jo-erlend_: No.
[23:50] <soren> jo-erlend_: Because that page says how to do it. It doesn't say how you did it.
[23:50] <jo-erlend_> because I'm following the steps very closely, except for using four disks instead of two, which shouldn't make any difference?
[23:50] <JackTO> qman: Since my remote IP based KVM, is not working with my new equipment, do I really need it on a server sitting in a colo/datacenter...   i guess only if the system lokcs up, and i need access to the bios i need it...?
[23:50] <soren> jo-erlend_: ...and you still haven't said /how/ it doesn't work.
[23:51] <qman__> JackTO, VNC would not let you access the BIOS
[23:51] <jo-erlend_> soren, I have. It was a while ago though. I get an initramfs, complaning that /dev/disk/by-uuid/<uuid here> doesn't exist. In fact, the directory doesn't exist at all.
[23:52] <JackTO> qman: true... i'm starting to realli like SSH... much better than the windows world i came from
[23:52] <soren> What kind of controller are they hooked up to?
[23:52] <soren> jo-erlend_: Å
[23:52] <soren> ^
[23:52] <soren> sheesh. Typing is hard.
[23:52] <jo-erlend_> I get almost the same results when I do it in a virtual machine, btw. The errors are not identical though. In the VM, it sais: no such disk.
[23:52] <qman__> JackTO, indeed, SSH is the swiss army knife of the networking world :)
[23:53] <soren> jo-erlend_: Let's look at it tomorrow. I need sleep.
[23:53] <jo-erlend_> soren, I've had disks on this machine before, although not in a software raid, but mdadm shouldn't be specially affected by the controller, should it?
[23:53] <soren> jo-erlend_: no.
[23:54] <JackTO> I'm just amazing in linux, i don't have to Reboot the server every time i install or change something... i like that a lot...
[23:54] <jo-erlend_> soren, I do too. I'll sleep well if you promise to help me with this tomorrow :>
[23:54] <JackTO> amazed*
[23:54] <soren> jo-erlend_: but being sure that the devices are around in initrams depends on drivers being loaded and  that depends on which controller etc.
[23:54] <jo-erlend_> I've spent more than sixteen hours trying to install ubuntu server on that damned box.
[23:54]  * soren wanders off
[23:54] <jo-erlend_> soren, all the devices are there, except for the md-ones.
[23:55] <jo-erlend_> but you're right. It's time to get some sleep. :)
[23:55] <soren> jo-erlend_: but /dev/disk/by-uuid/ doesn't exist?
[23:55] <jo-erlend_> right.
[23:55] <jo-erlend_> only by-path and by-id
[23:55] <soren> jo-erlend_: that's messed up.
[23:55]  * soren falls over, sleeping
[23:55] <jo-erlend_> it's identical in karmic.
[23:56] <jo-erlend_> but sleep tight. I'll tap you on the shoulder in the morning. :)