[00:01] <rcsheets> have you looked at top?
[00:03] <failover> yeap, Lucid have much less process runing, and the one wich is using more is -bash  0.7% and 3x rsyslogd using 0.4%
[00:19] <erichammond> smoser: I found that if a download was slow from http://uec-images.ubuntu.com/ I could just interrupt the connection, try again, and eventually get a far faster rate.  Different download connections of the same file performed consistently at 2KB/s, 200KB/s, and even 2MB/s.
[00:21] <smoser> erichammond, its just the release.
[00:21] <smoser> bandwidth to data center is bottlenecked.
[00:22] <erichammond> smoser: Yes, but I was surprised that I could get far higher speeds if I just kept retrying.
[00:22] <erichammond> smoser: not optimal for the whole population, unfortunately.
[00:22] <smoser> :)
[00:22] <smoser> yeah.
[00:22] <smoser> random i guess.
[00:25] <smoser> erichammond, this isn't published, but provides a mirror of those tar files
[00:25] <smoser> http://ubuntu-data.s3.amazonaws.com/
[00:26] <smoser> "isn't published" as in isn't written anywhere. i just use it for loading things to ebs.
[00:28] <erichammond> smoser: Thanks.  Doesn't appear to have Karmic and Hardy which are what I'm trying to copy now.
[00:29] <erichammond> smoser: I've got two of the four files and am still working on the remainder.
[00:31] <erichammond> smoser: I get 2K/s about 9/10 times and then 400K/s-2M/s which completes quickly and then has some sort of error on the disconnect throwing away the file 4/5 times.
[00:45] <uiuiui> hi, i cant get dhcp3 to send out a dhcpnak...
[00:45] <uiuiui> it refuses almost all win7 machines
[00:46] <uiuiui> as they dont send a dhcpdiscover right away
[00:46] <uiuiui> it seems stuck in non-authoritative mode, even if i have authoritative in config file
[01:46] <mikelifeguard> So, when I try to do git-daemon --listen=git.my.host or svnserve --listen-host=svn.my.host, I'm told the requested socket cannot be allocated - so I guess I am doing something wrong that has nothing to do with svn or git?
[01:58] <JanC> mikelifeguard: most likely you are running it as non-root and trying to use a port < 1024
[01:58] <mikelifeguard> Git uses 9418 though :s
[01:59] <mikelifeguard> I forget what svnserve is
[01:59] <JanC> or maybe you are trying to use a port that's already in use
[01:59] <mikelifeguard> hm, that's possible... let's check...
[02:04] <mikelifeguard> no, doesn't look like the ports are in use
[02:56] <jbernard> kirkland: yep, ill take a look
[02:57] <jbernard> kirkland: i agree with soren, it would be much more useful to have the total cost, i just didn't have a way of determining time since creation
[03:34] <AdamDV> Any reason a group name can't be all numeric?
[03:45] <persia> AdamDV: You may have a chance of confusing the group-name vs. GID detector for some commands, but it may be no issue at all.
[03:47] <AdamDV> Well, I tried it.
[03:47] <AdamDV> And then when I tried to add a user to the group, it told me the group doesn't exist.
[03:47] <AdamDV> So, I'm thinking it wont work.
[03:52] <amereservant> I need some opinions on making a few VPSs, such as on a small/personal scale....would virtualbox be the way to go?
[03:52] <amereservant> I was looking at the cloud server, but that's a two computer minimum setup.
[04:06] <twb> KVM is the virtualization technology that Ubuntu endorses.
[04:08] <amereservant> twb, Well, I looked at that, but it looked rather complicated and required virtualization to be enabled on the CPU.
[04:09] <persia> It doesn't.  It just falls back to being less efficient if it's not.
[04:10] <persia> And while it can be complicated, I find the combination of virt-manager on a desktop and the virtualisation server task on a server to be trivially easy to use.  Just tell the server to create some VMs, and they just work.
[04:10] <amereservant> Oh, I gotcha.  I guess I'll give it another look, I was just curious about how to do it on a small scale.
[04:11] <twb> persia: hmm, I didn't realize that the kvm fork of qemu worked *at all* without hardware VT
[04:11] <persia> twb: It most certainly does.  Works on my powerpc, for instance, and I'm not running a new enough kernel to use KVM on that.
[04:12] <amereservant> persia, Nice, thank you.  I'll definitely give it another look.  I installed and ran two virtual machines with VirtualBox, but I wasn't convinced that was very efficient.  Plus I'm not sure VirtualBox can be ran on Server edition.
[04:12] <twb> I always just explicitly installed the qemu variant.
[04:12] <persia> There was a period when it didn't (jaunty, I think, and maybe intrepid as well), but that's behind us.
[04:12] <twb> (You may recall yesterday I was bitching about trying to get qemu 0.12 onto hardy.)
[04:12] <twb> persia: thanks for letting me know
[04:12] <persia> Sorry to not mention it before: I thought your issue was the backporting, not the variant.
[04:14] <mikelifeguard> I'm trying to reverse-proxy requests to a Jetty server with http://sprunge.us/IDDf - but I only get http://sprunge.us/YgKH telling me the proxy is being disabled. Have I configured it incorrectly?
[04:18] <ezhangin> hey dudes, i am getting the weirdest problem
[04:18] <ezhangin> apparently my raid5 just formed another partition?
[04:18] <ezhangin> and i can't mount the orignal md
[04:19] <ezhangin> this is after upgrading to 10.04 without the raid drives turned on (because grub is retarded like that)
[04:19] <ezhangin> er, more like putting a new system drive in the computer and installing 10.04
[04:19] <twb> persia: it was a hodge-podge of issues and hubris
[04:19] <twb> Don't worry about it
[04:20] <twb> On an unrelated note: where can I find a copy of putty.exe on a host that doesn't use vhosting?
[04:20] <twb> The box I'm trying to get putty onto has broken DNS
[04:21] <ezhangin> and annoyingly i just had this thing mounted prior to a restart
[06:19] <amereservant> Does anyone here use KVM?
[06:20] <persia> daily
[06:20] <amereservant> persia, Ahhh!  You're still around.
[06:21] <amereservant> persia, I just had a quick question.... are all of the steps outlined across the top at https://help.ubuntu.com/community/KVM/Networking neccessary?
[06:21] <persia> My response before looking is "No".
[06:22]  * persia looks
[06:22] <amereservant> persia, It has "Installation" "Networking" "Create Guests".....
[06:22] <persia> Yep.  I've never done any of that.
[06:22] <amereservant> I'm on Networking ATM and this is a major ordeal.
[06:23] <amereservant> persia, What approach do you take?
[06:23] <persia> Do you need a real network, or are you happy just to have your server have client access to the internet, and be able to ssh to it from the host?
[06:23]  * persia has fairly minor requirements, so doesn't tend to do complex stuff
[06:23] <amereservant> It doesn't really matter.
[06:24] <amereservant> I've done everything on the Installation page.
[06:24] <persia> OK.  So, here's the quick'n'dirty way I generate VMs for throwaway test servers:
[06:25] <ezhangin> anyone using mdadm in 10.04?
[06:25] <twb> !anyone
[06:25] <ezhangin> lol
[06:25] <persia> 1) install libvirt-bin on the host, 2) install virt-manager on some headful client, 3) copy some installation iso to /var/lib/libvirt/images/ on the client, 4) launch virt-manager, 5) connect to the host, 6) use the wizard to create a new vm (selecting "local storage" and the install iso), 7) run an install, 8) reboot the VM, 9) do whatever it was I wanted to do in the first place, and usually 10) delete the VM when I'm done.
[06:25] <ezhangin> i treid earlier
[06:26] <ezhangin> tried, even
[06:27] <persia> amereservant: Note that this procedure is not necessarily best practice, takes longer than some other procedures, and fails firmly in the "works for me" category.
[06:27] <amereservant> persia, " on some headful client..."  What do you mean?  I installed it in the installation steps.
[06:27] <persia> I tend to run virt-manager on my laptop because I find the GUI handy, and don't have any screens or keyboards attached to my servers.
[06:28] <amereservant> persia, Ohhh, gotcha.  I'm doing this all inside of Ubuntu 10.04 Desktop.
[06:29] <persia> amereservant: All on one machine?  Then do everything on one machine.  That works too.  I just tend to use multiple machines because that avoids my laptop running slow when I have a couple VMs launched.
[06:30] <amereservant> persia, Yeah, it's a workhorse.  I wanted to try it out since you recommended it earlier and I've got two PCs on the network ATM, so I wanted to do some tinkering.
[06:33] <Roxyhart0> hi there, i am installing the cups driver for samba. Im in the first step i juat download the driver from cups web page (cups-windows-6.0-1.i386.rpm). Im not sure how to install it, somebody know how install it or where decompress the files?
[06:43] <amereservant> persia, Hmmm, no luck.  It just goes to a black screen when I open it.
[06:43] <amereservant> *using Virtual Machine Manager.
[06:43] <persia> What goes to a black screen when you open which?
[06:43] <amereservant> Does it support Lucid?
[06:43] <persia> Works for me on lucid.
[06:44] <amereservant> Ok, here's what I did.
[06:45] <amereservant> I opened up Virtual Machine Manager, clicked "File > Add Connection", Selected Hypervisor "QEMU/KVM", Connection "Local" and then pressed Connect......
[06:46] <persia> OK.  Did it connect?
[06:46] <amereservant> I right-clicked on localhost(QEMU) and selected "New", then gave a name for it and selected Local install media...
[06:46] <persia> (yes) OK.  Had you copied the iso, and could you select it properly?
[06:47] <amereservant> Step 2 I located the ISO file for Server10.04 64-bit, and tried leaving the settings on Generic or Linux Ubuntu Etch, then clicked forward....
[06:47] <persia> OK.
[06:47] <amereservant> Gave it 1GB RAM, 2 CPUs....
[06:48] <amereservant> Created a 10GB drive for it and Allocate entire disk now....
[06:48] <amereservant> Then clicked "Finished".....
[06:49] <amereservant> It says it's "Running", so I then I double-clicked it and it opens a window and it's just blank.
[06:49] <persia> OK.  Wait some.
[06:49] <persia> If nothing happens, verify your iso.
[06:50] <persia> I usually end up with the initial boot screen (select language, etc.) after a bit.
[06:50] <amereservant> I've installed it 3 times in VirtualBox with no issues using that ISO, so it's not that.
[06:50] <persia> Hrm.  That's a close match to the procedure I use that works.  Dunno.
[06:51] <amereservant> persia, This is how I installed it: https://help.ubuntu.com/community/KVM/Installation
[06:52] <amereservant> Anything look off about that?
[06:52] <amereservant> Everything passed in those steps.
[06:52] <persia> I don't see anything obviously wrong about that.  I didn't do anything except install libvirt-bin, personally, but the rest shouldn't hurt anything.
[06:53] <amereservant> Well isn't that just fancy.
[06:55] <amereservant> I'm going to install and try the AQEMU package and see if that gets me anywhere.
[06:58] <persia> OK.  Good luck.
[06:58] <amereservant> persia, Thank you for your help anyways.  It's not that big of a deal since I'm just learning.
[07:11] <rahman> hi how can I move a mysql database to a new server? moving /var/lib/mysql/foobar to the new server and restarting mysql didnt do the trick
[07:12] <amereservant> rahman, mysqldump
[07:12] <amereservant> Copying the data files is the wrong way of doing it.
[07:13] <amereservant> rahman, If you have phpmyadmin, you can easily dump all of them using "Export".
[07:13] <amereservant> rahman, But mysqldump is very easy and the recommended way of doing it.
[07:14] <rahman> amereservant: will I need to do "mysqldump foobar" ?
[07:14] <amereservant> rahman, Try "man mysqldump" for how to use it.
[07:15] <rahman> amereservant: thanks
[07:15] <amereservant> rahman, You can also find "how-tos" with a quick google search, I don't know off the top of my head though.
[07:16] <amereservant> 9608852
[07:16] <amereservant> Woops, that's not a form!
[07:17] <amereservant> Hehe.
[07:40] <rahman> And I have one more question, I need to move the mail server to a new machine. It is postfix, dovecort and libsasl2 installed and webmail as frontend
[07:41] <rahman> again I installed postfix, dovecort and libsasl and copied all config files from old machine like /etc/postfix/*
[07:44] <rahman> the problem is webmail frontend errors about "can't connect to POP3 server"
[08:41] <rzm25> hey guys has anyone run an Urban Terror server before?
[08:44] <zorzar> rahman: is the pop3 daemon running?
[08:49] <rzm25> guys can anyone help me in running Urban Terror on my server? I get permission errors
[08:57] <rzm25> kk
[09:02] <AlexC_> morning
[09:03] <AlexC_> is it possible to get the total file size of an extracted archive, without extracting it?
[09:05] <jefimenko> i'm trying to do a remote install on a server that already has grub (i'm replacing a different distro with ubuntu)
[09:05] <jefimenko> can the instructions here be used for an ubuntu server install like this?
[09:05] <jefimenko> https://help.ubuntu.com/community/Installation/NetbootInstallFromInternet
[09:07] <jefimenko> i couldn't find any ubuntu-server specific initrd.gz or linux fille
[09:10] <jefimenko> i took the netboot files from here: http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/current/images/netboot/ubuntu-installer/amd64/
[09:24] <cjwatson> kirkland,hggdh: there's a truly ancient bug about the image names being the same in several places, but frankly I'd rather leave them as they are and have clients work around it
[09:24] <cjwatson> it's been established for 5+ years and I'm pretty sure some folks depend on the current structure :)
[09:26] <jefimenko> is there a separate netboot install for ubuntu-server?
[09:26] <jefimenko> i really don't have physical access to this machine and the normal PXE/bootp method is very difficult at the moment
[09:27] <azlon> i am trying to recover my software raid. when i do --examine on sda1 it shows 2 failed drives. when i do --examine on sdc1 (one of the failed drives) it shows all drives are functioning properly
[09:27] <jefimenko> but this the netboot install from the internet is working great so far
[09:28] <jefimenko> but i'm just not sure what ubuntu flavor this is doing
[09:29] <rahman> zorzar: dovecot doen't start. it gives :  dovecot: Fatal: auth(default): User doesn't exist: vmail and  Fatal: Auth process died too early - shutting down
[09:30] <rahman> so webmail users can send mail but can't recieve any
[09:34] <azlon> why would --examine return 2 different results on 2 different drives?
[09:54] <gerardnetro> can anbody help me
[09:54] <gerardnetro> i was wondering... can I use the enterprise cloud version on Lan based
[09:55] <gerardnetro> i want to deploy web apps on local but soon on cloud... is it advisable to choose enterprise cloud
[09:57] <gerardnetro> anybody here
[10:12] <brianherman> gerardnetro: uh have you tried s3?
[10:12] <gerardnetro> nope im new here :[
[10:12] <brianherman> gerardnetro: ok
[10:12] <gerardnetro> but is it possible it will work on Lan
[10:13] <gerardnetro> for now
[10:13] <brianherman> gerardnetro: what are you trying to do
[10:13] <gerardnetro> i want to host vtiger
[10:13] <gerardnetro> via lan for now
[10:13] <brianherman> why do you need cloud for this?
[10:13] <gerardnetro> but soon ill ready it for cloud
[10:13] <brianherman> uh
[10:13] <brianherman> have you ever used it?
[10:14] <gerardnetro> because i want some other branches to access it
[10:14] <gerardnetro> nope
[10:14] <gerardnetro> first time
[10:14] <brianherman> ok
[10:14] <brianherman> First things first
[10:14] <brianherman> you need to install it on a server
[10:14] <brianherman> and learn how to use it
[10:14] <brianherman> cloud is just a buzzword
[10:14] <brianherman> it doesn't instantly mean that your stuff is going to get faster
[10:14] <gerardnetro> but im already installing enterprise cloud
[10:15] <brianherman> ok
[10:15] <gerardnetro> so i cancel it?
[10:15] <brianherman> no
[10:15] <brianherman> it doesn't really matter
[10:15] <gerardnetro> oh ok
[10:15] <brianherman> just if you expect it to get instantly faster
[10:15] <gerardnetro> what happens if we lost internet connection.. and im using cloud
[10:15] <brianherman> ok
[10:16] <brianherman> you are building the cloud
[10:16] <brianherman> not renting from somewhere
[10:16] <brianherman> right
[10:16] <gerardnetro> yes
[10:16] <brianherman> so you have your own personal cloud
[10:16] <gerardnetro> cause soon i want some of the agent to access their vtiger account ... on cloud
[10:16] <brianherman> even if you lose internet connection
[10:16] <brianherman> you have a cloud
[10:17] <gerardnetro> ok
[10:17] <gerardnetro> is cloud = host?
[10:17] <brianherman> This way is probably harder but better because you have an offline backup.
[10:17] <brianherman> Cloud = your own personal s3
[10:17] <brianherman> Good parts
[10:18] <brianherman> You have root access
[10:18] <brianherman> no latency
[10:18] <brianherman> bad parts
[10:18] <gerardnetro> oh but i can host web application?
[10:18] <brianherman> yes
[10:18] <brianherman> you can do whatever the @#$% you want
[10:18] <gerardnetro> great
[10:18] <brianherman> but expect to have issues
[10:18] <brianherman> i have never done the enterprise cloud
[10:19] <gerardnetro> well have a training later
[10:19] <gerardnetro> on Enterprise cloud
[10:19] <brianherman> Good luck
[10:19] <gerardnetro> so issues will not be a problem... i hope
[10:19] <brianherman> well i am a believer in murphys law
[10:20] <gerardnetro> oh :[
[10:21] <brianherman> but I also suffer from depression
[10:21] <brianherman> so
[10:21] <brianherman> ...
[10:21] <brianherman> dont listen to me
[10:21] <brianherman> just do it
[10:21] <gerardnetro> ok...
[10:21] <brianherman> You will learn a lot this way.
[10:21] <gerardnetro> but most of the time.. when i encounter errors, i can solve it :)
[10:21] <gerardnetro> i use google :)
[10:21] <brianherman> ya
[10:21] <brianherman> thats good
[10:22] <brianherman> well i have to fix my computer
[10:22] <brianherman> brb
[10:25] <ray__> Hello, .... has anyone an Idea when the Lucid Server Guide will be available? on the http://is.gd/bOgc3 site there still is 9.10 as the lates guide ... *hrmpf*
[11:08] <rahman> is there any one that has Dovecot experience?
[11:17] <PC_Nerd101> Hi - will a standard do-release-upgrade take me from lucid RC to current release?
[11:24] <maxagaz> hi
[11:24] <maxagaz> how to check all open ports on an machine from another machine ?
[11:28] <ray__> maxagaz apt-get install nmap && man nmap
[11:29] <ray__> or
[11:29] <maxagaz> ray__, I have nmap...
[11:29] <ray__> google for nessus
[11:29] <maxagaz> ray__, do you know the nmap command to get it ?
[11:31] <ray__> maxagaz, If you do not know nmap, either use man nmap or i guess it is better for you to use nessus (which as a graphical gui) http://www.nessus.org
[11:31] <ray__> s/as/has/g
[11:32] <maxagaz> ray__, I use nmap for one port, but don't the syntax for allports
[11:35] <ray__> maxagaz, ok, just type "man nmap" without the quotes in the terminal, right there on the first page is one fast example of how to scan for multiple ports, and further down are more detailed instructions on how to be sure all ports you want to have checked are included
[11:35] <maxagaz> ray__, it'sokay I found a tool to do it on ubuntu desktop
[11:37] <ray__> maxagaz, maybe you are better of in the standard #ubuntu channel for questions like this, in the ubuntu-server channel one requirement is to be able to and understand the man-pages ....
[11:37] <maxagaz> ray__, you mean this... nmap -A -T4 scanme.nmap.org
[11:37] <maxagaz> ray__, ok, thanks
[11:37] <ray__> maxagaz, as an example, yes
[11:38] <maxagaz> ray__, it's always better to know of a command to do this, thanks
[11:38] <ray__> np
[12:05] <wtfplymouth> why is the login line in the middle of the screen?
[12:06] <johe|work> ?
[12:07] <ray__> ?
[12:07] <alexxtasi> hi everybody... a "newbie" question: shall I run the "dist upgrade" command to upgrade from karmic to lucid server? Is there any problem mentiond ? (my karmic server installation has no more software than ssh, webmin, cups and samba serer)
[12:09] <ray__> alexxtasi, the recommended way is: http://is.gd/bOmk8
[12:10] <ray__> check network upgrade for ubuntu servers
[12:14] <alexxtasi> ray__: thanks ray... are there any problems reported? I 'll try...
[12:18] <wtfplymouth> is it possible or not dangerous to uninstall plymouth?
[12:20] <ray__> wtfplymouth, why?
[12:20] <ray__> wtfplymouth, it's a central part of ubuntu, afaik, so why do you want to deinstall it?
[12:23] <wtfplymouth> ray__: just thougt about it. disabled the bootsplash and now the login line is in the middle of the screen
[12:23] <ray__> alexxtasi, other problems than in the release-notes? I am sure there will be a few but ssh, webmin, cups should be fine, samba thought seems to have some issues, like the bugs  https://launchpad.net/bugs/572210
[12:24] <wtfplymouth> and how can i get all the boot massages back?
[12:24] <alexxtasi> thank you all...
[12:24] <ray__> and https://launchpad.net/bugs/572229 seem to predict, so, i would rather make a testupgrade in a virtual machine or be prepared to fix stuff manually, if you can't do A and are not skilled enough for B I would wait a month or two to let the fist bugs settle, i do not see you having the NEED to upgrade right now with cups and samba
[12:24] <wtfplymouth> like starting samba [OK] ...
[12:25] <ray__> wtfplymouth, as far as I could READ through the RELEASE NOTES and other documentation all the bootmessages are on Console 7 and in /var/log/boot.log
[12:26] <wtfplymouth> ray__: so its not possible to get them on screen during startup
[12:27] <ray__> I'l guess not, thought not been playing around with it, in reality I do not care as most of my machines are virtualized anyway, so it really does not help me to see the messages right on boot, so, why do you care about those messages anyway?
[12:29] <wtfplymouth> ray__: just to see if one of them failes during boot ;)
[12:29] <mvo> could someone with ldap exerience please comment on bug #572105 and what the best course for this is? I'm fine adding something to u-m, but would prefer somehting in the package(s)
[12:29] <ray__> wtfplymouth, well, you should be able to check that when you have the loginprompt, switch to console 7 (ALT+F7) and you should be able to browse through ...
[12:30] <wtfplymouth> ray__: but this istnt really a problem the real problem is that the server doesnt skip mountproblems during boot
[12:30] <wtfplymouth> like 9.10 did
[12:31] <ray__> wtfplymouth, as I read in the RELEASE NOTES or in some of the documentation it should be easy, there are some key bindings you can press to skip mount stuff and ignore stuff etc. You should read it
[12:32] <wtfplymouth> ray__: you have to press s to skip but the problem is thats a headless server und the boot stops till you press s. so you cannot login per ssh
[12:32] <ray__> and why does it hang?
[12:34] <wtfplymouth> ray__: cuz the disk is not connected and there is an entry for this disk in the fstab. but i want it to automount the disk when its connectet so i cannot delete the fstab entry
[12:34] <ray__> mvo, i think I read in the release notes that you can install nscd (name service caching daemon) which should cache all your credentials (better to log out and in again afterwards) so the shutdown of ldap should not be a problem, but I advise you to read the release notes yourself before trying that.
[12:34] <wtfplymouth> ray__: i need something like autoskip this errors
[12:36] <ray__> wtfplymouth, I think you won't get it, at least without me knowing the internals of plymouth I can not provide you with a better answer, maybe you should investigate in a proper automount solution like autofs or autofs5 or pmount, pmount is easy to setup and allows you mounts as normal user ...
[12:36] <ray__> as well the package automounter exists ...
[12:40] <mvo> ray__: thanks
[12:42] <wtfplymouth> ray__: thanks too
[12:42] <ray__> np
[13:15] <kirkland> cjwatson: okay, in that case, we'll try to somehow work around it
[13:16] <kirkland> cjwatson: what if cdimage produced a full manifest of the ISOs it hosts?
[13:17] <kirkland> cjwatson: just a tab delimited text file of full paths to ISOs, and text descriptions of each?
[13:17] <kirkland> cjwatson: it wouldn't take much, and i'm happy to write the shell code to do this on cdimage (if something doesn't already exist)
[13:18] <kirkland> cjwatson: not much more than a find . -type f -name "*.iso"
[13:18] <kirkland> cjwatson: and some mapping of paths to text descriptions
[13:20] <cjwatson> kirkland: already does for releases.u.c, probably just a matter of extending it
[13:20] <kirkland> cjwatson: ah, yeah, so that "flat" structure that releases uses...  that's what i'd like to see for testdrive
[13:21] <kirkland> cjwatson: s/ubuntu-10.04/lucid/
[13:21] <kirkland> cjwatson: could be handled entirely with symlinks/hardlinks even
[13:21] <cjwatson> no plans to change the structure of cdimage
[13:21] <cjwatson> happy to emit a manifest if that helps
[13:21] <cjwatson> changing the structure is an utter royal pain
[13:21] <kirkland> cjwatson: alrighty, i 'll take the manifest ;-)
[13:22] <lool> smoser: heya
[13:22] <smoser> hi
[13:22] <lool> smoser: I wonder whether you considered gzip's --rsyncable for the UEC tarballs
[13:22] <cjwatson> kirkland: file me a bug on the ubuntu-cdimage project?
[13:22] <kirkland> cjwatson: in fact, i'm glad that you're not willing to change the structure of cdimage, as testdrive greatly depends on that structure
[13:22] <smoser> i think it uses it
[13:22] <smoser> do you know otherwise ?
[13:22] <kirkland> cjwatson: will do
[13:23] <kirkland> cjwatson: thanks!
[13:23] <lool> smoser: I wonder why the rsync "virtual" bandwidth is so low when rsyncing the image here
[13:23] <lool> it is much lower than for binary data like kernels
[13:23] <lool> it's almost as if it redownloads everything
[13:25] <lool> smoser: Anyway, if you say it's already in use then there's probably nothing we can do
[13:26] <lool> smoser: thanks!
[13:26] <smoser> lool, http://bazaar.launchpad.net/~ubuntu-on-ec2/vmbuilder/automated-ec2-builds/annotate/head%3A/build-ec2-image uses 'gzip-rsyncable' (also in that directory)
[13:26] <smoser> which, i can't see how would not get there.
[13:26] <smoser> lool, i think that the images are just very dense
[13:26] <smoser> at least thats been my explanation before, and i have noticed this too.
[13:27] <smoser> i just thought that with the tar -S and gzip that the size just is really small and ends up not working well with rsync.
[13:27] <smoser> but thats all guessing.
[13:29] <ray__> hey guys, what's up with the 10.04 Server Guide?
[13:31] <lool> smoser: This all seems correct to me, I guess we can't have it any better
[13:31] <lool> smoser: Hmm alternatively what could happen is that the files are not in the same order over runs
[13:31] <smoser> well, they're probably not, right ?
[13:31] <smoser> its a filesystem
[13:31] <smoser> with files in it
[13:31] <lool> smoser: What I'm saying is that tar might pick up files in ramdom orders from the chroot where you prepare the fs
[13:32] <lool> ideally, these would be sorted in the tarball
[13:32] <smoser> tar doesn't pick up files other than '.img, kernel, ramdisk'
[13:32] <smoser> .img is a 2GB file with a filesystem in it.
[13:32] <lool> smoser: I'm speaking of the other tar
[13:32] <lool> Oh ok
[13:32] <lool> then it's probably the fs being completely different over different runs
[13:33] <smoser> so thats almost certainly it.
[13:33] <smoser> right.
[13:33] <smoser> writes just went to other places in the filesystem.
[13:33] <lool> I guess the only way aounr that is using a fs level dump like dumpe2fs, but I'm not sure this would be a good idea
[13:33] <lool> or using tarballs of the files
[13:34] <lool> smoser: Ok; thanks for discussion; would be nice to fix for rsync users, but it's not that big and I dont think many people rsync this stuff
[13:34] <smoser> i never really thought about it like this before.
[13:40] <lool> smoser: BTW I was curious whether setting GZIP in the env with --rsyncable would work instead of using a custom program, and it does
[14:14] <oru_work> greetings how can i scan for open ports ?
[14:15] <incorrect> you could use a web browser
[14:16] <Pici> !info nmap | oru_work
[14:20] <oru_work> damn ubuntu servers are so slow right now
[14:30] <hggdh> cjwatson: OK, we will work around the image names. Thank you
[14:33] <incorrect> well gb.archieve is saturated, there used to be an apt tool to find the fastest local mirror
[14:37] <incorrect> good old university of kent
[14:47] <melter> are custom init scripts still a valid way to start & stop services in 10.04?
[14:55] <resno> is it recommended to upgrade my distro?
[15:42] <enquora> the libmozjs-dev package no longer seems to be available in 10.4. can anyone enlighten me?
[16:14] <enquora> tty1 hangs or crashes on my installation - I never get a console prompt. tty2...n working fine. Suggestions on what to look at?
[16:17] <lamont> I wish there was a way I could smash more information into the error from dpkg for the bogus hostname case
[16:17] <lamont> maybe exit 55 or some such
[16:17] <lamont> :-)
[16:25] <zul> lamont: apport hook?
[16:25] <lamont> zul: meh
[16:25] <lamont> but yeah, that might do it
[16:26] <lamont> then we could just have the hook tell the admin that he picked an invalid name for his computer
[16:26] <zul> postfix is on the list for apport-hooks for maverick
[16:34] <JanC> lamont: yesterday I had somebody with a "hostname" with a space in it  :-(
[16:37] <lamont> JanC: exactly
[16:37] <lamont> the other slightly meaner one I did is you can't tell postfix that mydomain={debian.org,ubuntu.com}
[16:37] <lamont> because, well, NO
[16:38] <lamont> and if you actually run one of those domains, you know how to deal with the block
[16:40] <JanC> lamont: apparently somehow ubiquity allowed hostnames with spaces in karmic...  :-(
[16:53] <lamont> there have always been ways to have invalid hostnames. it's a question of definitions
[16:57] <zul> lamont: the rfc text for hostnames should be pretty definitive ;)
[16:59] <lamont> yeah, but if the machine isn't on the network, then who cares.
[16:59] <lamont> see 70+% of the windows machines, which let you pick all sorts of weird stuff
[17:02] <chillindave> Is it possible to host multiple virtual machines running Apache on port 80 with one external IP?
[17:04] <chillindave> And if so, how do you route the traffic on the host computer?
[17:06] <JanC> chillindave: use a reverse proxy or such ?
[17:07] <chillindave> JanC, Oh, gotcha.
[17:13] <melter> if i have "start on started _x_" in an upstart script, will it also shut down before _x_?
[17:15] <uiuiui> hey, i cannot get dhcp3-server to reply to dhcpinform, no dhcpack is sent by the server and the client remains with the misconfigured ip
[17:15] <uiuiui> this is very annoying for clients with windows7 as it seems win7 does not send dhcpdiscover packets
[17:17] <martinjh99> Anyone had any problems with PHP and Apache in the latest 10.04? Did a upgrade and now the server wants to save my php file rather than parse it.  But it is not all the time though...
[17:18] <martinjh99> What do I need to configure and where in the apache2 config??
[17:18] <chillindave> martinjh99, If it's prompting you to download it, it's because php isn't running or isn't configured correctly.
[17:18] <martinjh99> Dave thing is though it was before the upgrade to 10.04...
[17:18] <chillindave> martinjh99, Are you sure php is installed?
[17:19] <martinjh99> So I have no idea what happened... Yeah it is and running as a simple script with phpinfo in works fine
[17:19] <chillindave> martinjh99, Try "sudo apt-get install php5"
[17:19] <chillindave> martinjh99, Hmmm.
[17:20] <chillindave> martinjh99, So it works when you create a file with phpinfo(); in it?
[17:20] <martinjh99> Yup  - http://martinjh.homelinux.com/phpinfo.php#
[17:20] <martinjh99> phpmyadmin will work too - Just sometimes it won't...
[17:21] <chillindave> martinjh99, Yeah, hide that file.  It's a bad idea to share that publically.
[17:21] <martinjh99> Yup  - I'll get rid of later... Not really a busy server though...
[17:21] <chillindave> martinjh99, That's odd.  I would try re-installing php5 and Apache2 because it should just work.
[17:22] <martinjh99> dave thats what I thought...
[17:22] <chillindave> chillindave, Make backups of your vhosts if you have them, then do "sudo apt-get --purge remove apache2 php5"
[17:23] <martinjh99> No vhosts - Just the one... Might have to try that then...
[17:23] <chillindave> martinjh99, Then re-install it and see if the problem persists.  It should either work or fail as a whole.
[17:23] <martinjh99> True... ;)
[17:24] <chillindave> martinjh99, I'm no expert on it, but it sounds like it's Apache2 causing the problem.  If php is installed, it shouldn't serve php files as downloads.
[17:26] <martinjh99> Going to try and purge and re-install then Dave... Just backing up some files...
[17:27] <whaley> or a misconfiguration of mod_php...
[17:27] <chillindave> martinjh99, It won't erase your webfiles.
[17:27] <martinjh99> what the ones in /var/www ?
[17:27] <chillindave> martinjh99, It'll tell you it didn't remove your webroot folder since it wasn't empty.
[17:27] <martinjh99> Ah Cool...
[17:27] <chillindave> martinjh99, Yeah, I did it the other day.
[17:27] <martinjh99> :)
[17:28] <ziesemer_> I understand the risks, but what would be the best way to get OpenLDAP (slapd) to use Berkley DB 4.8 instead of 4.7?  (The official package is apparently compiled against and requires libdb4.7, even though libdb4.8 is already installed by default.)
[17:28] <whaley> martinjh99: is mod_php in your /etc/apache2/mods-enabled directory?
[17:28] <chillindave> martinjh99, I *think" there's also another apache2 package you need to purge to get rid of the config files.
[17:29] <martinjh99> well there is a php5.conf and php5.load
[17:29] <chillindave> whaley, Yeah, he's saying it works but then sometimes it suddenly starts serving the php files as downloads.
[17:30] <chillindave> I haven't seen it work, then suddenly glitch like that.  Usually it's because PHP isn't installed/configured and it serves all php files as downloads.
[17:30] <whaley> chillindave: yeah... that's what i suspected
[17:30] <chillindave> martinjh99, The files being served for download do have ".php" extensions, right?
[17:30] <whaley> perhaps it's only enabled for certain <VirtualHost> and <Directory> blocks?
[17:30] <martinjh99> yeah the file it wants to save is an index.php
[17:31] <chillindave> whaley, Well, only if he set it up that way, which I'm guessing he didn't.
[17:31] <chillindave> Doing the install using aptitude will configure it for all directories automatically.
[17:31] <martinjh99> thats what I use to install aptitude
[17:32] <chillindave> Yeah, I figured it was.
[17:32] <martinjh99> I just apt-get purged the files and the config is still in /etc/apache shall I delete them before installing?
[17:33] <chillindave> martinjh99, That's what I was saying earlier, there's another package you need to purge to get rid of those...one sec and I'll find out the name of it.
[17:33] <martinjh99> ah thanks
[17:34] <chillindave> Try purging "apache2.2-common"
[17:34] <chillindave> I think that's the one that has all of the configuration files in it.
[17:35] <chillindave> martinjh99, Oh, and you'll also need to re-install phpmyadmin because it won't be configured anymore.  Same thing, just use purge and re-install it.
[17:36] <martinjh99> well this is what it said - Havent done it yet... http://paste.ubuntu.com/425347/
[17:36] <chillindave> Type "N".
[17:37] <martinjh99> problems??
[17:37] <chillindave> Then add phpmyadmin like "apt-get remove --purge apache2.2-common phpmyadmin"
[17:37] <chillindave> martinjh99, Otherwise it won't purge the phpmyadmin configuration and won't re-install correctly.
[17:39] <martinjh99> http://paste.ubuntu.com/425348/ - ok to go now...?
[17:39] <chillindave> martinjh99, Yeah, that's fine.
[17:41] <martinjh99> This look good to re-install? http://paste.ubuntu.com/425351/
[17:44] <chillindave> martinjh99, Well, just do apt-get install apache2 php5 phpmyadmin
[17:45] <chillindave> It'll automatically prompt you for the dependencies such as apache2.2-common
[17:50] <martinjh99> re-installed and rebooted the server (I know...) still getting a prompt to save the php file the filename is http://martinjh.homelinux.com/gallery2/upgrade/index.php?step=5 bug there do you think...?
[17:52] <martinjh99> Getting this error at the server start - http://paste.ubuntu.com/425354/
[17:53] <cclausen> martinjh99: edit /etc/hosts and add the correct fqdn for your system
[17:53] <martinjh99> ok thanks :)
[17:53] <cclausen> I really HATE that Ubuntu uses 127.0.1.1 by default.  it breaks Kerberos SSO and other things that need to use real FQDNs.
[17:54] <cclausen> martinjh99: that gallery upgrade page loads for me in IE8.  I am not prompted to download
[17:54] <martinjh99> Am using Firefox... Will try with IE...
[17:54] <cclausen> also works for me in FireFox as best I can tell
[17:55] <chillindave> martinjh99, Check your server log files and see what they say.
[17:55] <cclausen> I think its a browser / MIME type problem
[17:55] <chillindave> martinjh99, I would think apache2 error log would tell you something.
[17:55] <uiuiui> is tehre no way to make dhcp3 send dhcpnaks to all dhcpinform packets
[17:56] <uiuiui> it seems only win7 clients send those w/o getting the correct ips
[17:56] <martinjh99> Not that I can see - http://paste.ubuntu.com/425357/
[17:58] <chillindave> martinjh99, Going off of what cclausen said, open up your php.ini file using "sudo gedit /etc/php5/apache2/php.ini" and search for mime type.  I "Think" there's a default mimetype setting there somewhere.
[17:58] <martinjh99> ah getting segfaults somewhere - http://paste.ubuntu.com/425359/
[17:58] <chillindave> martinjh99, It should be set to something like text/html.
[17:59] <cclausen> Content-Type: text/html; charset=UTF-8
[17:59] <cclausen> (from wget -S Content-Type: text/html; charset=UTF-8)
[17:59] <cclausen> looks okay to me
[17:59] <cclausen> just close and reopen your browser and see if it just got confused
[17:59] <chillindave> martinjh99, That could be your issue then.  I have no clue on that one.
[17:59] <cclausen> also, try going to step=1 instead of step=5
[18:00] <martinjh99> php.ini - Mimetype already seet to text/html
[18:01] <JanC> cclausen: how does Ubuntu break "real FQDNs"?
[18:01] <chillindave> It shouldn't matter what browser he uses since PHP is server-side, so if it's serving PHP pages as downloadables, it's an issue with the server-side, not the client browser.
[18:01] <cclausen> JanC: it doesn't break them, but doing things like setting a Kerberos KDC and/or SSH GSSAPIAuthetnication don't work when the machine hostname is pointed back to a local only IP
[18:01] <chillindave> Browsers can't trick a server into serving the php files as downloadables.
[18:02] <cclausen> JanC: Debian has the same problem, so its not a pure Ubuntu issue.
[18:02] <cclausen> chillindave: and it shouldn't serve different pages to me where they seem to load fine...
[18:03] <chillindave> cclausen, What do you mean?
[18:03] <cclausen> chillindave: Oh, I bet the specific step=5 part doesn't have some specific module instead and a PHP error message is causing oddness before the MIME type is set.
[18:03] <cclausen> chillindave: but you are correct, should be logged (if PHP error logging is enabled)
[18:04] <chillindave> Well, if the php engine fails, it will most likely trigger the Apache error handler, which means it falls on Apache to log.
[18:04] <cclausen> yes, but stderr might not go anywhere
[18:05] <cclausen> and apache might log error 500 or might not
[18:05] <chillindave> It usually goes to the system log.  Apache creates an error log in /etc/log/apache2 by default.
[18:05] <cclausen> really?
[18:05] <cclausen> I thoguht it was in /var/log/apache2 ?
[18:06] <chillindave> cclausen, Oh, woops.
[18:06] <chillindave> That's what I meant.
[18:06] <cclausen> access.log should list the attempt and error.log an error message
[18:07] <chillindave> Yeap.
[18:07] <chillindave> martinjh99, You may also open php.ini and set "show startup errors" to true.
[18:08] <maxagaz> Hi, I've just bought an HP printer LaserJet M1213nf MFP, but impossible to make it work on my computer, and no help on the net, can someone help me ?
[18:08] <martinjh99> show startup errors =true; yes?
[18:08] <chillindave> martinjh99, No, find the directive for it.
[18:09] <chillindave> maxagaz, You might want to try #ubuntu
[18:09] <martinjh99> Do I need to restart anything?
[18:10] <chillindave> martinjh99, sudo apache2ctl restart
[18:11] <chillindave> martinjh99, Anytime you edit a configuration file, you need to restart Apache to apply it.
[18:11] <martinjh99> ok
[18:12] <martinjh99> I think that the problem is that there are segfaults in the error log...
[18:12] <chillindave> martinjh99, yeah, I would guess that's the problem too.  I dunno how to go about fixing that though, never been down that road.
[18:12] <martinjh99> I do have a backup of the data on there so I could do a re-install of 10.04 from scratch - something though is making it segfault...
[18:13] <martinjh99> Me neither to be honest...
[18:14] <chillindave> martinjh99, That might be the quickest solution, otherwise it'll consist of searching and trying any number of possible fixes hoping one eventually works.
[18:15] <martinjh99> I might do that - I was thinking about repartitioning it anyways...
[18:18] <martinjh99> Thanks for the chat and the help guys...
[18:18] <chillindave> martinjh99, np.
[18:23] <maxagaz> chillindave, nobody can help
[18:23] <maxagaz> chillindave, it seems there's no driver yet for linux for this printer
[18:24] <zul> smoser: ping
[18:27] <chillindave> maxagaz, Yeah, that's the problem with new printers.  Manufacturers don't prioritize Linux drivers if the ever make them at all.
[18:28] <JanC> depends on the manufacturer
[18:28] <maxagaz> chillindave, what can I do then ?
[18:29] <maxagaz> is there a hope to make it work quickly ?
[18:29] <chillindave> maxagaz, Find a printer that makes a driver for Linux.
[18:29] <JanC> maxagaz: HP is linux-friendly normally
[18:30] <RoyK> trying to ask for something in #ubuntu is like whispering in an crowd at an airport
[18:30] <JanC> maxagaz: did you try that printer with lucid yet?
[18:31] <RoyK> chillindave: which printer is this?
[18:31] <chillindave> HP printer LaserJet M1213nf MFP
[18:31] <chillindave> RoyK, ^^
[18:31] <JanC> maybe we need #ubuntu-printer-help ...
[18:32] <RoyK> chillindave: perhaps a generic laserjet driver might work - or generic PS or PCL
[18:32] <maxagaz> JanC, yes, I'm trying with lucid
[18:32] <chillindave> RoyK, It's for masagaz, not me.
[18:32] <RoyK> maxagaz: see above :)
[18:35] <RoyK> http://pastebin.com/CSEV2tWA <-- new toy :)
[19:03] <RoyK> maxagaz: any luck with that?
[19:06] <maxagaz> RoyK, no, unfortunately
[19:07] <RoyK> the generic drives mostly work
[19:07] <RoyK> is it pcl or ps?
[19:13] <maxagaz> RoyK, How to know ?
[19:19] <jamesturk> is this an acceptable place to ask about Ubuntu on EC2?
[19:23] <jamesturk> seeing some high (~1.2) load numbers on a brand new EC2 install when postgres is running (before even creating a database, let alone any sort of usage)
[19:24] <pipepupo> Good afternoon
[19:24] <pipepupo> Do you know?
[19:24] <pipepupo> Does exist WUBI for Ubuntu-Server 10.04?
[19:33] <RoyK> maxagaz: it's in your printer's manual
[19:33] <RoyK> maxagaz: if it's a postscript printer, though, it probably also accepts PCL
[19:35] <domas> thank you lucid for having me to use 'grub-install --force' instead of just 'grub-install', you just shaved off 10 minutes of my life because of this... ;-)
[19:38] <RoyK> domas: are you certain that was a big loss?
[19:41] <domas> royk: I am!
[19:41] <domas> royk: oh well, I've spent more than hour on this gpt issue before :)
[19:41] <domas> on hardy/karmic
[19:41] <RoyK> gpt? do you really need that?
[19:42]  * RoyK usually boots off something small and uses the large stuff for data alone
[19:48] <RoAkSoAx> kirkland, howdy!! Heya quick TestDrive question. The function is_iso() is only used when passing options/arguments to testdrive right? We can say it is a local function for the UI
[19:48] <domas> royk: ummm, thats a simple internal raid10 array
[19:48] <domas> royk: and our / is 10G
[19:49] <RoyK> why don't you just use LVS on the devs?
[19:49] <domas> performance
[19:49] <RoyK> afaik the only good reason to use GPT is if you need to boot off something >2TB
[19:50] <RoyK> and you won't get more performance with another partition table
[19:50] <domas> well, ubuntu formats it all as gpt, if any partitions are >2TB
[19:50] <domas> LVM isn't just partition table :)
[19:50] <RoyK> domas: is your boot device >2TB?
[19:50] <domas> no
[19:50] <RoyK> then don't use GPT
[19:51] <RoyK> how large is your data volume?
[19:51] <domas> I can't create >2TB partitions with MBR it seems
[19:51] <domas> >2TB
[19:51] <RoyK> do you need a partition table?
[19:51] <RoyK> mostly you don't with large volumes
[19:51] <RoyK> a mirror of 2x2TB is not > 2TB
[19:51] <domas> well, it is all /dev/sda
[19:51] <RoyK> what sort of system is this?
[19:52] <domas> 16 300GB disks in raid10 setup
[19:52] <domas> X4240
[19:52] <RoyK> why don't you just use two of them for the root and the rest for data?
[19:52] <kirkland> RoAkSoAx: right, i think so
[19:53] <RoyK> domas: or even, two small SSDs for the root and the big drives for data?
[19:53] <domas> royk: I'd lose some spindles, if I used disks for root :)
[19:53]  * RoyK still sticks to opensolaris/zfs for large storage
[19:54] <domas> royk: you're not running databases, maybe
[19:54] <RoAkSoAx> kirkland, Ok then. It works with VBox, so now I just have to do some more testing, clear few doubts and will propose it for merging
[19:54] <RoyK> domas: it's always, ALWAYS, a good idea to separate the OS from the data
[19:54] <domas> royk: I'd prefer few small disks for boot and large ssds for data :)
[19:54] <domas> royk: why?
[19:54] <domas> OS doesn't read anything, doesn't write anything
[19:54] <domas> why does it deserve two separate spindles
[19:54] <RoyK> then use two small SSDs
[19:55] <RoyK> doesn't take up much space
[19:55] <domas> all space taken by disks :)
[19:55] <domas> and why would I need that for OS?
[19:55] <RoyK> separating the OS from the data simplifies things if the system goes down
[19:55] <RoyK> s/if/when/
[19:55] <domas> if system goes down , we reimage it usually :)
[19:56] <RoyK> reimage 2,4TB?
[19:56] <kirkland> RoAkSoAx: yeah, that function should be hypervisor agnositc
[19:56] <domas> royk: yup. well, usually system goes down because of data volume failures
[19:56] <kirkland> RoAkSoAx: just tells you if you the argument looks like an ISO file
[19:56] <domas> so it doesn't matter where OS is :)
[19:57] <domas> and OS can be reinstalled in few minutes
[19:57] <RoyK> domas: it still simpliefies things a lot to keep the OS separate
[19:57] <RoyK> and reimaging 2,4TB might take a while
[19:57] <RoAkSoAx> kirkland, yeah. I guess that can still be used in the front-end to validate "others" but for now i'll leave it as it is
[19:57]  * RoyK hints about ZFS
[19:58] <domas> royk: *shrug*, maybe for some people it is easier
[19:58] <domas> but I don't want to waste disks on something that won't be used
[19:58] <RoyK> domas: adding an SSD or two, perhaps 8GB each, won't cost much
[19:58] <domas> no space :)
[19:58] <RoyK> for 2,5" 8mm SSDs?
[19:59] <domas> all 16 slots already taken by disks :)
[19:59] <domas> anyway, I can actually netboot those machines
[19:59] <domas> if I really need it :)
[20:00] <RoyK> domas: don't get me wrong, but I've only been working with this for like 15 years - separating the root from the data saves a lot of trouble
[20:00] <RoyK> well, then netboot them
[20:00] <domas> thats why root gets separate partition :)
[20:00] <domas> and, oh well, I've been working with this for 14 years I guess
[20:00] <RoyK> setup a small ramdrive and let them run in that
[20:00] <domas> anyway, OS does not use disks
[20:00] <RoyK> I've done that on some setups :)
[20:00] <domas> I can give small partition to keep the bootstrapping data
[20:00] <RoyK> video streaming stuff
[20:01] <RoyK> but if you keep the OS running in case of failure, you have a chance to debug it
[20:01] <RoyK> perhaps fix it
[20:02] <domas> well, for OS not to come up from the big array you have to get array screwed up
[20:02] <domas> and once array is screwed up, I don't trust data and reimage :)
[20:02] <RoyK> by putting all eggs in the same basket, you make things harder
[20:02] <RoyK> what sort of db is this, btw?
[20:03] <RoyK> 2TB database seems a lot
[20:03] <domas> mysql
[20:03] <domas> not to me!
[20:03] <RoyK> wtf :)
[20:03] <RoyK> what sort of data?
[20:03] <RoAkSoAx> rgreening, ping?
[20:03] <domas> royk: website
[20:03] <RoyK> domas: how many sites?
[20:04] <domas> one
[20:04] <RoyK> may I ask which?
[20:04] <domas> yep, you may
[20:04] <domas> wikipedia
[20:04] <RoyK> ah
[20:04] <RoyK> makes sense
[20:04] <RoyK> how much memory do you have in that box?
[20:04] <domas> oh well, current dataset is smaller
[20:04] <domas> 64G
[20:05] <RoyK> k
[20:05] <domas> tiny
[20:05] <domas> this is my hobby
[20:05] <RoyK> heh
[20:05] <domas> at work we have bigger mysql boxes
[20:05] <domas> :)
[20:05] <RoyK> domas: take a look at opensolaris and SSD caching - it rocks rather well
[20:05] <domas> we (my work) just released SSD caching for linux
[20:06] <domas> like, this week
[20:06] <RoyK> domas: nice - got a PR for that?
[20:06] <domas> http://www.facebook.com/note.php?note_id=388112370932
[20:06] <RoyK> still, Btrfs is far from finished, so I guess I'll stick with ZFS for a while
[20:07] <domas> XFS ftw
[20:07] <domas> =)
[20:07] <RoyK> XFS isn't really much compared to ZFS :)
[20:07] <domas> well, I used to work for sun :)
[20:07] <domas> I probably know a bit more than I want about ZFS :)
[20:08] <RoyK> domas: anything I should know?
[20:08] <domas> beware of full filesystems :)))
[20:08] <domas> well, 80% full, or 90% full too
[20:08] <RoyK> I know that one
[20:08] <RoyK> fragmentation happen
[20:08] <domas> anyway, pretty much every concept of ZFS is usually implemented at any reasonable DBMS
[20:09] <domas> checksumming, doublewrite buffering, cow, etc
[20:09] <RoyK> we have this 30TB system at work, currently filling up with satellite data from Eyjafjallajökull
[20:09] <RoyK> and dedup performance sucks unless you have a truckload of RAM or L2ARC
[20:10] <RoyK> it's still the best FS around for large storage systems (imho)
[20:10] <rgreening> RoAkSoAx: pong
[20:11] <domas> royk: except for DBs ;-)
[20:11] <RoAkSoAx> rgreening, have some little time to review testdrive code?
[20:11] <RoyK> domas: might be - not using zfs for that
[20:12] <rgreening> RoAkSoAx: unfortunately not atm. I'm just heading out for the day. However, feel free to put together any details/ideas and e-mail me. I may have some time later tonight to review
[20:12] <RoAkSoAx> rgreening, will do then. thanks :)
[20:12] <rgreening> np
[20:13] <rgreening> I should be online during the week-end. feel free to ping me
[20:13] <rgreening> RoAkSoAx: ^
[20:13] <rgreening> :)
[20:13] <RoAkSoAx> rgreening, better yet then :)
[20:13] <rgreening> hehe
[20:23]  * RoAkSoAx loves testdrive
[20:24] <RoyK> domas: how long is it since you looked into zfs?
[20:24] <domas> we still run some at wikimedia
[20:34] <switchback> hey guys, new to ubuntu server. I login and the motd gives me a notice of out packages and security updates, how do I install these updates?
[20:35] <domas> apt-get update; apt-get upgrade
[20:35] <switchback> oh
[20:35] <Pici> sudo apt-get dist-upgrade
[20:35] <switchback> simple
[20:38] <RoyK> switchback: apt-get update && apt-get dist-upgrade
[20:38] <RoyK> 'update' updates the index - it might have changed since last it was automatically updated
[20:40] <|Barcode> I can't seem to upgrade from 8.04 LTS to 10.04 LTS using do-release-upgrade without the --devel-release option.
[20:40] <|Barcode> Is the final release not ready yet or something?
[20:41] <RoyK> did you do an apt-get update first?
[20:41] <|Barcode> Yep, twice to be sure.
[20:41] <RoyK> ok
[20:41] <RoyK> 10.04 is out, but it might be some latency somewhere
[20:41] <RoyK> it shouldn't matter much, though
[20:41] <RoyK> I don't think -d sticks
[20:42] <maxb> |Barcode: IIUC, it is normal for a new LTS to not be enabled to be offered by default via update-manager for LTS->LTS upgrades until the first point release
[20:43] <RoyK> |Barcode: also, upgrading to 10.04 for a server you want to stay stable, isn't something I'd recommend - wait a month or two
[20:43] <RoyK> 8.04 is supported until 2013
[20:43] <|Barcode> Well, I'm doing this on a non-critical machine. I will wait for sure for the critical ones.
[20:44] <RoyK> then just -d
[20:44] <RoyK> I have updated to to 10.04 on a couple of boxes myself so far - works well
[20:44] <RoyK> for critical stuff, I'll wait a while
[21:25] <blistov> hey all.  i'm building ipt_netflow from cvs against 2.6.28-11-server under ubuntu 9.04.  Compiles correctly but insmod results in "invalid module format".  The headers package matches my running kernel.  Idea's?
[22:13] <|Barcode> RoyK and maxb: I did the upgrade with the --devel-release option and "lsb_release -a" does show "Ubuntu 10.04 LTS", so it all worked out. Thanks guys!
[22:16] <RoAkSoAx> kirkland, will you be available this weekend (tomorrow) to do some testdrive testing to merge the code?
[22:55] <jggpc> Hello all.  Problem with 10.04 x68 iso here
[22:55] <jggpc> sorry x86..
[22:57] <jggpc> burns ok, md5 checks out ok, but when I attempt to use it to install I get the error retrieving fs-core-modules-2.6.32.21-generic-di
[22:57] <jggpc> attempting the install on a dual core Dell poweredge 2500
[22:57] <jggpc> sorry not dual core, dual proc
[22:57] <jggpc> anyone seen this?
[23:00] <jggpc> is anyone here?
[23:02] <vraa> jggpc what kind of dual proc system? what cpus you got i am interested
[23:02] <jggpc> its an older Dell poweredge 2500  twin 1.4 gb P4's  1 gb ram..  a little test system I use for stuff.
[23:03] <jggpc> the 64 cd won't install on it, so I have to use the x86 versions
[23:03] <vraa> dang twin p4's
[23:03] <vraa> yah i didn't expect p4 to be x86_64 compatible
[23:03] <vraa> you try the alternate iso?
[23:03] <jggpc> is there an alternate for server?  I did know
[23:04] <vraa> http://torrent.ubuntu.com:6969/
[23:04] <vraa> oh hm, for server i dont think there is, but there really isn't a big difference, i think server just doesn't contain the gui
[23:05] <jggpc> well server doesn't have gui, but is a lot smaller.  the server only has a cd drive, so dvd's won;t work
[23:05] <jggpc> and the lack of network install possibilities sort of messes that idea up
[23:06] <vraa> you can get the desktop and alternate 64bit iso's in cd format
[23:06] <vraa> i mean 32bit
[23:06] <jggpc> is it a mini?
[23:06] <vraa> a33e98826003515e46ef5075fcbf4914b307abe2	ubuntu-10.04-alternate-i386.iso	689MiB
[23:06] <vraa> 3e16157f0879eb43e9e51f45d485feff90a77283	ubuntu-10.04-desktop-i386.iso	699MiB
[23:07] <jggpc> oh wow..  sort of scared to try the desktop.  Was in the other channel ealier..  ti seems to be having more issues than I am having..  hee..
[23:07] <jggpc> I may just wait a while until some stability comes to town on lucid..  9.10 is pretty good, I'll move back that way for a month
[23:07] <jggpc> thanks for the idea though
[23:08] <vraa> dont be a wuss
[23:08] <vraa> download the alternate and give it a shot
[23:08] <sbeattie> jggpc: how did you verify the md5sums?
[23:08]  * handheldCar hasn't had too many issues w/ the desktop and has solved all two of them already.
[23:08] <jggpc> not being a wuss.  we are tryign to roll out several thousands of Ubuntu machines in the enterprise.  I cannot have a instable system crashing out a production line
[23:09] <vraa> wow thousands!
[23:09] <jggpc> yeah..  we have 27K windows machines we are replacing
[23:09] <handheldCar> sounds like fun. i almost wish i could have /that/ job
[23:09] <vraa> are you switching because samba4 can now replace active directory?
[23:10] <jggpc> hehe..  no you probably wouldn't..  I having to battle minions of the evil empire daily
[23:10] <jggpc> no we're switching becuase we're tired of the micro$oft tax
[23:10] <handheldCar> yeah, my classmate told me I wouldn't like a deployment job.
[23:10] <domas> 27 thousand windows servers?
[23:10] <domas> I'd definitely not want to be there :)
[23:11] <jggpc> I've run opensuse on my box for 5 years now, but the archetect is using ubuntu for the enterprise roll.
[23:11]  * handheldCar guesses it's a nice place to visit....
[23:11] <jggpc> no, 27 thousand windows desktops, 1500 windows servers.  wsorry..  was clear there
[23:11] <jggpc> *wasn't
[23:11] <domas> ;-)
[23:11] <domas> I figured
[23:11] <domas> just thought, that in -server channel I'm allowed to make such joke!
[23:12] <jggpc> heh..  is alright.
[23:12] <handheldCar> anybody have their update set to lts? it didn't work for me yesterday until I set it back to normal.
[23:13] <jggpc> anyways.  good luck, this channel is far quieter than the desktop channel!
[23:13] <handheldCar> yeah there should be another desktop channel
[23:17] <handheldCar> when upgrading, does anyone take the new versions of config files and reconfigure those afterwards?
[23:22] <handheldCar> I had several that could have been replaced but only replaced the 1st.
[23:42] <wtfplymouth> so after spending the whole day on ubuntu server i have to say goodbye
[23:43] <domas> bye!
[23:44] <handheldCar> :D
[23:44] <wtfplymouth> or does anyone have an idea how to skip fstab problems during boot automatically?
[23:46] <handheldCar> paste your fstab here http://paste.ubuntu.com/ and maybe someone will diagnos it.
[23:47] <wtfplymouth> i know that there is a problem in my fstab! because the disk of this fstab entry is not connected. i just dont want it to wait for the s key
[23:48] <wtfplymouth> 9.10 just booted
[23:48] <handheldCar> do you have a line for virtualbox usb? is this 10.04?
[23:48] <domas> hmmmm, what 'buffers' are accoutned in /proc/meminfo 'Buffers' line?
[23:50] <domas> mmmm, and how much of that overlaps with 'Slab'
[23:51] <wtfplymouth> me?
[23:55] <wtfplymouth> is there anywhere something like a plymouth.conf