[00:21] <pmatulis> rcsheets: ok
[00:21] <rcsheets> pmatulis: in fact, it turned out to be my own pure stupidity, not a bug at all
[01:00] <Roxyhart0> hi there...i got a question i change the IP in a diferent vlan for my samba PDC server but know when i try to join client to the domain or access to the domain client that were already joined, i got the message that can not find the domain (there  are no firewall) between diferent vlans. Somebody know what could be the problem?
[01:30] <mase_wk> hi guys, can anyone tell me the correct way to disable / enable the upstart enabled init scripts ?
[01:31] <mase_wk> should  update-rc.d work still or is there another way to do it ? i tried removing mysqld with update-rc.d however it still appears to start on boot
[03:18] <silentwhisper> how is the basic step by step setting up of mail server?
[03:18] <silentwhisper> postfix,dovecote,and squirel is this correct
[03:19] <silentwhisper> ?
[03:21] <mase_wk> silentwhisper: yeh that sounds about right
[03:22] <mase_wk> postfix+dovecot is a good combination
[03:28] <dominicdinada> has anybody had any problems with samba shares since the most recent update today ?
[04:08] <hungnv> hello everyone
[04:10] <hungnv> I have a question about apt-get and dpkg. Once I use apt-get install or dpkg -i packagename, some package popup a buffer ask me to input data, likes ldaputils, it asks me for ldap server address, ldap version...
[04:10] <hungnv> how can I just skip that buffer?
[04:11] <twb> hungnv: set the debconf priority to critical, or if you're scripting, set the debconf UI to "noninteractive".
[04:11] <twb> hungnv: obviously this shouldn't be done blithely, since some of those questions are important
[04:13] <hungnv> twb: yes, I'm scripting. I have about 100 clients and now I must install just one application for all. Now I'm on server side and excecute my deployment script to do that job automatically.
[04:14] <hungnv> config files was created before, after installation completed, I just copy configuration file to right place
[04:14] <twb> It'd be better to preseed the answers to ldap-auth-config
[04:15] <hungnv> twb: what do you mean?
[04:16] <twb> Rather than saying "don't ask me" and then overwriting whatever it does -- which can cause problems if there's an upgrade (such as a security update), it'd be better to provide the answers in advance, letting it write the correct file itself.
[04:20] <hungnv> it will takes a very long time to do such a simple job
[04:21] <lowridah> if the servers to be deployed to aren't already in action you can kickstart them, specify the packages to install, and configure the LDAP settings in preseed or ks
[04:22] <lowridah> otherwise you might consider looking into puppet+mcollective if you plan on doing large scale software deployments
[04:24] <lowridah> hungnv:  http://blog.hjksolutions.com/articles/2007/07/27/unattended-package-installation-with-debian-and-ubuntu
[04:25] <twb> I don't know why you'd go with kickstart over preseeding.
[04:25] <lowridah> because i remember all the kickstart directives i'd need to install ubuntu
[04:25] <lowridah> i have to look at preseed examples?
[04:25] <twb> Yeah, so it's really only for RHEL refugees
[04:26] <hungnv> lowridah: thank for the link
[04:26] <lowridah> hungnv:  np
[04:26] <lowridah> twb:  if that was true, ubuntu wouldn't be so anaconda compliant
[04:26] <silentwhisper> hi i like to setup webmail
[04:26] <lowridah> i'm not debating how to deploy, i'm just saying what i do.
[04:26] <twb> Fair enough
[04:26] <twb> As for me, I preseed Debian and kickstart RHEL
[04:26] <hungnv> lowridah: I deploy these software for clients, not for server
[04:26] <silentwhisper> postfix, them imap then squirel is this correct?
[04:27] <lowridah> i kickstart ubuntu and the redhats off a WDS server
[04:27] <lowridah> call me crazy.
[04:27] <twb> WDS?
[04:27] <lowridah> windows deployment service
[04:27] <twb> Is that Windows <something>?
[04:27] <twb> Right.
[04:28] <lowridah> yes, i deploy most OSes off a windows box.
[04:28] <twb> We don't have Windows at all
[04:28] <lowridah> ahh i work in web hosting, we have all flavors.
[04:31] <hungnv> lowridah: you link is exactly what I want
[04:31] <hungnv> :-)
[04:34] <lowridah> awesome =)
[04:40] <imp7> ok guys, the internet told me to come here and ask apparmor questions.
[04:40] <imp7> I have a java app that is spawning threads or something that look like this: profile="/server/bin/linux-x64/app-monitor//null-16//null-17"
[04:41] <imp7> The null-16 and 17 change.. is there a way to use a variable in the profile name?
[04:41] <imp7> like profile="/server/bin/linux-x64/app-monitor*"
[04:44] <imp7> I have another java app that is showing up in aa-unconfined like this: 16362 /server/runtime/bin/pc-app (../runtime/bin/pc-app) not confined
[04:45] <twb> imp7: #ubuntu-hardened is probably better for apparmor
[04:45] <twb> But we can also try to help
[04:46] <imp7> thanks :)
[04:47] <twb> (Admittedly, I don't know the answer to that one myself.)
[04:47] <lowridah> it was all moonspeak to me
[04:48] <twb> lowridah: apparmor is a LSM (linux security module) that adds MAC (mandatory access control) based on paths, where selinux uses inodes.
[04:48] <lowridah> ahh
[04:48] <twb> His problem is that his java apps change the paths they need all the time
[04:48] <lowridah> i see
[04:49] <twb> imp7: silly question: when you say "threads" are you talking about green threads, pthreads, or conventional processes?
[04:49] <imp7> not sure myself
[04:50] <imp7> they arnt in ps but show up in the kern.log while my profile is in complain mode
[04:50] <twb> Hum
[04:57] <imyousuf> I have installed UEC for the first time (total newbie) on a single node with node controller option selected and downloaded the lucid lynx 64 bit version from the store. Now when I am trying to run the image I am getting - http://paste.ubuntu.com/467353/ My machine configuration is core i7, 4G RAM, 1G VGA, Lucid Lynx 64-bit. Any idea how I could solve it?
[05:17] <hggdh> imyousuf: run 'euca-describe-availability-zones verbose', and check the values there
[05:18] <imyousuf> doing it right now hggdh
[05:24] <imyousuf> hggdh: http://paste.ubuntu.com/467812/ Not sure what is wrong
[05:25] <imyousuf> but the free/max values seems not right, but being a newbie I am not certain
[05:27] <imyousuf> hggdh: what do you think?
[05:28] <hggdh> imyousuf: the free/max columns suggest you do not have an NC active
[05:28] <imyousuf> ok
[05:28] <hggdh> (NC == Node Controller, the piece that actually runs the VMs)
[05:29] <imyousuf> I see
[05:29] <hggdh> so...
[05:29] <imyousuf> but when I installed UEC I selected NC
[05:29] <hggdh> sudo euca_conf --list-nodes
[05:29] <hggdh> wow
[05:29] <imyousuf> doing it
[05:29] <hggdh> back to it -- what exactly did you install? Just the NC?
[05:29] <imyousuf> no hggdh all of them in one single box
[05:30] <hggdh> k
[05:30] <imyousuf> and ran the command its a blank line
[05:30] <hggdh> so no NC active.
[05:30] <hggdh> try 'sudo euca_conf --discover-nodes'
[05:30] <imyousuf> yes, seems so
[05:31] <imyousuf> blank line with: INFO: We expect all nodes to have eucalyptus installed in //var/lib/eucalyptus/keys for key synchronization. at the end hggdh
[05:31] <hggdh> weird
[05:32] <imyousuf> hggdh: I did service eucalyptus-nc status and got "eucalyptus-nc stop/waiting"
[05:32] <hggdh> yeah
[05:32] <imyousuf> so does this mean I need to do - sudo /etc/init.d/eucalyptus-nc start ?
[05:32] <hggdh> ok. try 'sudo start eucalyptus-nc'
[05:33] <imyousuf> ok
[05:33] <imyousuf> it started
[05:34] <hggdh> gotta go, sorry
[05:34] <imyousuf> hggdh: It started and now running discover nodes tells me to setup login-less ssh using authorized keys, let do that and will get back
[05:34] <imyousuf> thanks a lot hggdh
[05:34] <hggdh> welcome
[05:45] <Shapeshiftr> Hey.
[05:45] <Shapeshiftr> Where can I get MemTest for ubuntu server?
[05:48] <twb> apt-get install memtest86+
[05:53] <Shapeshiftr> Thanks.
[05:53] <Roxyhart08>  hi there somebody have a doc to configurate reverse zones for different subnets?
[05:56] <Shapeshiftr> twb , erm, *relative Linux/Debain newbie talking* how do you run it, then?
[05:56] <twb> Shapeshiftr: it appears in your bootloader
[05:57] <twb> Inherently, you cannot run memtest86+ from within the normal system, although you could potentially use kexec-tools to load it without going through a POST
[05:57] <Shapeshiftr> So, how do I check if my memory is working to its fullest extent?
[05:57] <twb> There's also a userland memtest tool that CAN run from within linux, but obviously it's not as reliable
[05:57] <twb> Shapeshiftr: reboot into memtest
[05:58] <Shapeshiftr> Hm, thanks.
[06:04] <Hilikus>  is there any way to keep a server in a low-power state (stand-by) and make it wake up when there are network requests? apache request or ssh request for example
[06:09] <dominicdinada> anybody else having problems with samba shares after the update today ?
[06:15] <twb> Hilikus: basically, no.
[06:15] <twb> Hilikus: assuming by "low power" you mean suspend-to-ram with wake-on-lan.
[06:16] <twb> Hilikus: you can, of course, reduce the CPU frequency and spin down hard disks and similar tricks
[06:22] <Hilikus> twd are those tricks already part of the normal kernel or i have to enable them or install some daemon for it?
[06:24] <twb> Depends
[06:25] <twb> What's your actual goal?  To reduce the electricity bill for your machine room?
[06:35] <mase_wk> Hilikus: if you have so little requests that you can sleep afford for a significant portion of time then your probably better off getting a very low power machine
[06:35] <mase_wk> geode LX or arm based machine
[06:36] <Hilikus> twd, yes. and lower heat production, and also increase the usable life of the parts
[06:36] <twb> Certanly *halting* a machine is actually bad for its overall lifetime
[06:36] <mase_wk> for 7watts i can power a geode Lx500 and get around 33req/s on a completelye dynamic website
[06:37] <twb> Yeah, using low-power gear like the Pentium III or an arm/geode/atom is the most obvious approach.
[06:37] <Hilikus> mase_wk: i'll look into it
[06:37] <mase_wk> ie completely non optimised
[06:37] <Hilikus> thanks guys
[06:38] <twb> Alternatively, if you have a large number of machines, consolidating them into VMs on a single host can help
[06:38] <mase_wk> the new intel chips are very efficient / cool
[06:38] <mase_wk> twb: was just about to suggest that :)
[06:38] <mase_wk> they also come with some advanced VM stuff that KVM will use
[06:39] <twb> As can migrating them to VPSs living in a larger colo that someone else takes care of -- since they have the economy of scale to get that kind of thing right
[06:39] <mase_wk> i'm currently running 12 VM's on a server which is purring along nicely
[06:39] <mase_wk> yeh linode do whole VM's for like $20 pm
[06:40] <twb> mase_wk: we plan to reduce our core network to one soekris router and one Q9550 running everything as a jail (linux) or vm (windows)
[06:40] <twb> Currenty we have more core servers than staff :-(
[06:40] <mase_wk> haha . those soekris boards are good
[06:41] <mase_wk> i'm waiting for someone to make something like the vayetta/ cisco routers with dual core arm CPU's
[06:41] <twb> Why dual core?
[06:42] <mase_wk> so i can run openvpn on the other :)
[06:42] <twb> Uh?
[06:42] <twb> Surely it'll be SMP
[06:43] <mase_wk> well smp would be better :)
[06:43] <twb> So both cores would be running linux
[06:43] <mase_wk> yeh , one for switching the other for interesting thigns
[06:44] <mase_wk> things*
[06:44] <mase_wk> vpn/ firewall etc..
[06:44] <twb> I don't think switching is done in the CPU
[06:44] <mase_wk> well forwarding is done in hardware
[06:44] <twb> Even if it was, I don't think it'd require significant
[06:44] <twb> load
[06:44] <mase_wk> but anything more advanced, vlan etc.. is done on the cpu
[06:45] <mase_wk> depends on what bandwidth you have
[06:45] <mase_wk> and what your doing
[06:45] <twb> Oh, you're talking saturated gige?
[06:46] <twb> I mean, I do tagging and firewall on 100baseT with a 200MHz MIPS, and that doesn't even notice the load
[06:48] <mase_wk> i've found that with those MIPS chips the second they start getting a decent load , adding complex routes or VPN services tends to reduce the throughput
[06:48] <mase_wk> mind you that is with GigE , not 100base as you mentioned
[06:48] <twb> Fair enough
[06:48] <mase_wk> but still, the cisco's i have here are all 30+ wattss
[06:48] <twb> I'm a sysadmin, not a netadmin :-)
[06:49] <mase_wk>  some of the arm chips are measured in milliwatts
[06:49] <twb> Don't forget that each of your lightbulbs will be 25W or more
[06:49] <mase_wk> pft light ? =)
[06:49] <mase_wk> i thought you were a sysadmin
[06:50] <mase_wk> my monitor provides me with a healthy glow
[06:50] <mase_wk> but i see what you mean
[06:51] <twb> Oh, you're still WEAK enough to need a screen?
[06:51]  * twb inflates chest, strikes dramatic pose
[07:10] <Hilikus> if a script run by cron writes to stderr will it be mailed to root or just if it writes to stdout?
[07:11] <twb> Yes
[07:12] <Hilikus> cool, thanks
[07:29] <SandGorgon> anybody using 10.04 servers here ? how has the experience been vs 8.04 ? which hosting are you using ?
[07:38] <thrain][> SandGorgon: using 10.04 in production for some smaller LAMP servers I maintain (mediawiki, dokuwiki, dotproject). It's just as solid as the 8.04 servers I have. They are not in hosting, though; virtual machines in our local VMWare install.
[07:40] <SandGorgon> thrain][, thanks!
[08:03] <Hilikus> thanks guys
[08:26] <huats> moring
[08:26] <huats> morning
[08:28] <SandGorgon> hi guys.. i recently got a dedicated server and see that they have given me a "root" user. How do I disable that and go back to the safer Ubuntu way of "sudo" ?
[08:48] <mase_wk> SandGorgon: install / configure sudo and either disable the rootuser or set a long random password
[08:51] <SandGorgon> hmm... cool
[08:56] <SandGorgon> one more question - in some of my servers, when I log on it gives a nice welcome screen with System Load, Memory Usage, # of logged on users, etc. how do I get this ?
[08:59] <rcsheets> SandGorgon: i believe that is the landscape-common package
[09:26] <SandGorgon> rcsheets, bingo.. thanks
[09:37] <larsemil> imyousuf: any luck with UEC?
[10:16] <hggdh> Daviey: there?
[10:18] <Daviey> hggdh: o/
[10:19] <hggdh> Daviey: http://pastebin.ubuntu.com/467895/
[10:21] <hggdh> Daviey: and the nodes are still not shown on 'euca_conf --list-ndoes'
[10:24] <Daviey> hggdh: OK.. can you raise a bug?
[10:24] <Daviey> Now it's in the archive.. i really want to track issues via bugs
[10:26] <hggdh> Daviey: certainly
[10:26] <Daviey> hggdh: rocking.
[10:27] <Daviey> hggdh: can you try removing the nodes.list file in /var/* ?
[10:36] <hggdh> Daviey: will do
[11:05] <glen1> is colocation? when you put your hardware in someone elses datacenter?
[11:07] <skuld_kid2> I need some help please.  I tried to add a third hard drive to my volume group, and it didn't work right ,so i removed the drive, but it looks as though i DEGRADED my volume.  and now I can't login  can anybody help?
[11:27] <skuld_kid2> I need some help please.  I tried to add a third hard drive to my volume group, and it didn't work right ,so i removed the drive, but it looks as though i DEGRADED my volume.  and now I can't login  can anybody help?
[12:20] <hggdh> Daviey: https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/609112
[12:40] <Italian_Plumber> should I reboot my server after a kernel update?
[12:47] <pmatulis> Italian_Plumber: if you want to use the new kernel, yes
[12:49] <Italian_Plumber> how would I decide if I wanted to use the new kernel?
[12:58] <qman__> Italian_Plumber, that's a matter of risk analysis
[12:58] <Italian_Plumber> ah
[12:58] <qman__> look at the list of what was fixed, decide how important that is in your environment, and determine whether/when downtime is worth the benefit
[13:05] <Italian_Plumber> where can I find the list of what was fixed?
[13:09] <qman__> I'm sure it's on launchpad, but finding it is proving non-trivial
[13:10] <qman__> that, and it keeps failing while I'm trying to click links
[13:10] <qman__> here you go: https://launchpad.net/ubuntu/+source/linux/+changelog
[13:11] <qman__> err, those are for maverick
[13:11] <Italian_Plumber> i'm on hardy
[13:11] <qman__> ah, ok
[13:11] <qman__> https://launchpad.net/ubuntu/hardy/+source/linux/+changelog
[13:11] <pmatulis> Italian_Plumber: what kernel is it?
[13:12] <Italian_Plumber> well I have one server that I reboot every day, another production server that I probably should reboot off hours, and two other servers that are non-production.
[13:14] <Italian_Plumber> http://pastebin.com/Mc1ArmcM  <-- update log
[13:14] <Italian_Plumber> I bleieve it is 2.6.24-28 ... but it seems to have updated two kernels.
[13:15] <Italian_Plumber> thanks qman__
[13:15] <qman__> looks like there were a lot of fairly important bugfixes in that version
[13:16] <Italian_Plumber> mmk
[13:16] <qman__> so I'd suggest rebooting at next convenient opportunity
[13:16] <pmatulis> Italian_Plumber: try /usr/share/doc/linux-image-2.6.24-whatever/changelog.Debian.gz
[13:18] <Italian_Plumber> cool... thanks pmatulis
[13:21] <pmatulis> Italian_Plumber: are you really an Italian plumber?
[13:21] <Italian_Plumber> me and my brother Mario, yup...
[13:21] <pmatulis> Italian_Plumber: nice
[13:23] <pmatulis> Italian_Plumber: i had to have a plumber come over the other day, i pierced a hole in a water pipe with a drill
[13:24] <bogeyd6> How do I disconnect an iscsi mount?
[13:25] <pmatulis> bogeyd6: logout and delete i think
[13:26] <bogeyd6> you know, i could just uninstall the open-iscsi
[13:26] <bogeyd6> i was just testing it on my laptop
[13:26] <pmatulis> bogeyd6: iscsiadm -m node -p $ISCSI_TARGET --logout; iscsiadm -m node -p $ISCSI_TARGET -o delete
[13:26] <bogeyd6> awesome
[13:30] <bogeyd6> gnome terminal needs a copy and paste shortcut
[13:30] <jpds> bogeyd6: Shift-Insert?
[13:30] <jpds> bogeyd6: middle-click ?
[13:30] <bogeyd6> middle click, nice
[13:31] <bogeyd6> you just increased my productivity
[13:31] <bogeyd6> we got these solaris machines for a SAN running 28tb online in high availability
[13:32] <bogeyd6> and if you think because you are linux expert you can admin a solaris box, you are quite wrong ol chap
[13:59] <bogeyd6> Has anyone here upgraded a server from 9.10 to 10.04 yet? Can you share with me difficulties?
[14:01] <SandGorgon> hey guys.. is it recommended to use "vm.overcommit_memory = 2" on linux for a Mysql DB server with 8GB RAM ? Our database size is about 1GB
[14:07] <pmatulis> SandGorgon: is that OOM-related?
[14:08] <bogeyd6> SandGorgon, probably not unless you are aiming for what that provides
[14:11] <bogeyd6> SandGorgon, http://www.redhat.com/magazine/001nov04/features/vm/
[14:12] <SandGorgon> I actually read while reading up on optimizing database performance on Ubuntu. It recommended that vm be set with that setting
[14:19] <silentwhisper> sir pls check if my postfix set up is ok
[14:19] <silentwhisper> http://ubuntu.pastebin.com/a8p7zku8
[14:22] <silentwhisper> hello
[14:28] <silentwhisper> please help
[14:50] <pmatulis> simplexio: what's the problem?
[15:37] <sommer> :-)
[15:47] <Nokio> Hi all, I am looking for a tool on my ubuntu 10.04 server that would allow me to know at the end of the day how much bandwidth my server generated. In the form of total download upload for the day ? is that possible?
[15:54] <bogeyd6> Nokio, two things comes to mind, cacti and mrtg
[15:54] <bogeyd6> also http://www.ubuntugeek.com/bandwidth-monitoring-tools-for-ubuntu-users.html
[15:56] <Nokio> bogeyd6, ok thanks a lot i will go read about these. Thanks and have a nice day !
[16:02] <Nokio> bogeyd6, These 2 seems to be really cool and complete. I was more looking for a tool that requires maybe lest configuration because all i need is to know for a few week how much bandwidth is generated by 1 server
[16:02] <bogeyd6> check the website then
[16:03] <bogeyd6> Nokio,  http://www.ubuntugeek.com/bandwidth-monitoring-tools-for-ubuntu-users.html
[16:03] <Nokio> bogeyd6, ok cool thanks ! have a nice day
[16:03] <bogeyd6> parle vou francais?
[16:05] <pmatulis> bogeyd6: yes
[16:06] <sherr> Nokio: also something called "vnstat" might be worth a look
[17:35] <jaraco> I have libneon 0.29.0 installed (a dependency of subversion 1.6.6). I'd like to upgrade to libneon 0.29.1 or later (and have subversion use that package). What's the best way to do this?
[17:35] <RoyK> hm... top - 18:34:41 up  3:33,  3 users,  load average: 24.74, 17.09, 11.72
[17:46] <pmatulis> ouch
[17:46]  * RoyK just installed a new  toy for the scientists at work - a dual 12-core opteron thing with 64 gigs memory
[17:47] <RoyK> so the current load of ~25 isn't so bad, really :)
[17:48] <pmatulis> what are those scientists doing?  did you ever ask?  :)
[17:49] <RoyK> pmatulis: currently running inversion jobs on this particle movement model used during the Eyjafjallajökull eruption to try to make it better
[17:50] <RoyK> stuff like that takes cpu time
[17:50] <pmatulis> gotcha
[17:50] <pmatulis> do you fine tune that beast?
[17:50] <RoyK> nah - just installed Lucid
[17:51] <RoyK> I/O is on NFS to an opensolaris box with a truckload of drives (50TB net)
[17:53] <FunnyLookinHat> I forget - should I use useradd or adduser ?
[17:54] <RoyK> useradd -m
[17:54] <RoyK> that is, -m is 'make homedir'
[17:54] <FunnyLookinHat> kk ty
[17:54] <RoyK> but usually you want that
[17:54] <RoyK> man useradd
[17:55] <RoyK> http://karlsbakk.net/top-24.png
[17:57] <FunnyLookinHat> RoyK: lol - the man page suggests using adduesr
[17:57] <RoyK> oh well :)
[17:58] <pmatulis> FunnyLookinHat: adduser is a wrapper to useradd
[17:58] <RoyK> I'm probably to used to useradd, then
[17:58] <pmatulis> FunnyLookinHat: it simplifies things, uses defaults
[17:58] <pmatulis> FunnyLookinHat: see /etc/adduser.conf
[17:59] <FunnyLookinHat> Right - simple is good.  :)
[17:59] <pmatulis> FunnyLookinHat: especially on a friday
[17:59] <pmatulis> too bad i'm working on ipsec, sniff
[18:00] <FunnyLookinHat> haha
[18:00] <FunnyLookinHat> I'm just provisioning server images - so much easier...
[18:01] <FunnyLookinHat> sudo aptitude install <insert ridiculously long list of packages>
[18:01] <FunnyLookinHat> git clone
[18:01] <FunnyLookinHat> ./configure
[18:01] <FunnyLookinHat> done.
[18:01] <pmatulis> nice
[18:01] <FunnyLookinHat> so far so good :)
[18:28] <RoyK> hi all. I have a couple of 'number crunchers', one with 2x8 cores and another with 2x12 cores, both with 64 gigs of RAM. Should Ubuntu Server (lucid) work well on these without further tuning? If not, how can I see if something is slowing things down?
[18:29] <cwillu_at_work> RoyK, as much as anything works well without tuning, I'd expect so, yes
[18:30] <cwillu_at_work> The user-space isn't terribly relevant to this, and if the kernel gives you trouble, it's trivial to run a different one
[18:30] <RoyK> the 16 core box has been running a few weeks, and except a problem where the kernel seemd to eat a truckload of memory (a reboot fixed it), the system has been stably running at 100% load on all cores. Both show minimal system and wio times (althought the 24 core box has only been running a few hours)
[18:31] <RoyK> the users (scientists) mostly run up to a few more cpu intensive processes than active cores available, so I guess the scheduler shouldn't have much problem with it
[18:35] <cwillu_at_work> slabtop is useful to investigate kernel memory leaks
[18:35] <RoyK> that didn't find anything
[18:36] <RoyK> anyway - I haven't seen it lately
[18:36] <cwillu_at_work> yep, throughput is the problem that the scheduler solves at the expense of latency, to the chagrin of desktop users everywhere :p
[18:36] <cwillu_at_work> define didn't find anything
[18:36] <RoyK> but it was 40gigs or so lost, even after the userspace processes were stopped
[18:36] <cwillu_at_work> If the kernel had allocated memory, then slabtop would show it :p
[18:37] <RoyK> I couldn't find where it was allocated, and what used to be a quick machine, turned into something from the ninetees
[18:37] <RoyK> anyway - will check more if it happens again
[18:38] <cwillu_at_work> slabtop -s a should be useful
[18:38] <RoyK> 24 cores and 64 gigs of RAM in a NOK 50k box isn't bad
[18:38] <RoyK> or USD9k
[18:39] <RoyK> bbl
[19:14] <pedahzur> Trying to do an upgrade, getting this error message for the package 'linux-server' "linux-server: Depends: linux-image-server (= 2.6.32.24.25) but 2.6.32.23.24 is to be installed" Anyone else seen it?
[19:16] <pedahzur> I do have backports enabled, but not proposed.
[19:17] <pedahzur> Disabled backports...same error.
[19:18] <talcite> hey guys, I just updated the scst wiki page for ubuntu-server. Could someone who knows infiniband take a look and tell me what they think? https://wiki.ubuntu.com/scst
[19:21] <sherr> pedahzur: I see that as well today. On these occasions, I usually just wait a day or so and the problem gets fixed.
[19:31] <pedahzur> sherr: Yeah, it usually does, but it's still a bit disconcerting that an broken update was released.
[20:26] <quakenul> hi there
[20:27] <quakenul> i'm having trouble installing 10.04 64-bit server from a usb drive and hoped someone could help
[20:27] <quakenul> i prepared the stick and the installer booted just from but then tried to check for a CD drive
[20:27] <quakenul> which doesn't exist
[20:27] <quakenul> from = fine
[20:28] <quakenul> i searched for a solution but was unable to find one
[20:32] <pedahzur> quakenul: How did you prepare the USB stick?
[20:33] <quakenul> using the universal usb installer in windows
[20:35] <pedahzur> quakenul: I did this recently using the Ubuntu "USB Startup Disk Creator."  Everything worked fine, but it wrote the boot loader to the USB drive (since it was /dev/sda, and I was installing on a machine with a RAID card, and the drive was actually /dev/cciss/c0d0) You might want to give that a try.
[20:35] <quakenul> alright then, i will give it a try, thanks!
[20:35] <pedahzur> quakenul: Not really sure what to say otherwise: it didn't have any trouble finding the "CD" for me.
[20:35] <failover> quakenul
[20:35] <failover> https://help.ubuntu.com/community/Installation/FromUSBStick
[20:35] <failover> go to Know Issues
[20:36] <failover> there is a workaround
[20:36] <quakenul> i If you get "Incorrect CD-ROM detected" error
[20:36] <quakenul> are you talking about this paragraph?
[20:36] <failover> yeap
[20:36] <quakenul> well, i checked and this option is actually already there
[20:37] <quakenul> thanks anyway :)
[20:37] <quakenul> i will give usb startup disk creator a try
[20:37] <failover> At the boot prompt type "install cdrom-detect/try-usb=true"
[20:37] <failover> this works fine for me
[20:38] <quakenul> hm okay let me check
[20:41] <SandGorgon> when I use curl -LO to download a file (for example http://pl2.php.net/get/php-5.2.13.tar.gz/from/pl.php.net/mirror), it names the file "mirror". How do I get it to name it php-5.2.13.tar.gz
[20:41] <quakenul> nope, sadly not working
[20:41] <FunnyLookinHat> Hey guys - I'm getting this error when trying to bring up NFS:  * Not starting NFS kernel daemon: no support in current kernel.
[20:41] <FunnyLookinHat> I followed the guide to a T on https://help.ubuntu.com/10.04/serverguide/C/network-file-system.html
[20:45] <pmatulis> FunnyLookinHat: well, what kind of kernel are you using?
[20:45] <FunnyLookinHat> pmatulis: stock as far as I know...  hold on
[20:46] <FunnyLookinHat> pmatulis: 2.6.33.5-rscloud     It's an image for RackspaceCloud
[20:46] <FunnyLookinHat> But I'm seeing issues that the stock kernel seems to not register as being compatible anyways...
[20:46] <pmatulis> FunnyLookinHat: so it's not a stock ubuntu kernel
[20:46] <FunnyLookinHat> I suppose not...  :(
[20:48] <pmatulis> FunnyLookinHat: grep NFSD in your kernel config file that *should* be under /boot
[20:50] <FunnyLookinHat> pmatulis: looks like the kernel improperly doesn't have a symbol that the nfs startup script looks for...  https://bugs.launchpad.net/ubuntu/+source/linux/+bug/493145
[20:50] <FunnyLookinHat> I'm trying one of the patches...  :)
[20:51] <FunnyLookinHat> better yet - I just removed the check to make sure the kernel is compatible
[20:54] <sweltman_> Is this the right forum for bugs in updater?  attempting upgrade to 10.04 server but unsuccessful.
[20:54] <d3vnu11> sweltman_: what error are you getting?
[20:55] <sweltman_> unable to update (tool states broken dependencies)
[20:56] <sweltman_> there are none that I know about
[20:56] <d3vnu11> what are the broken dependencies?  I ask because I was just having an issue with a broken dependency for the linux-image-virtual package
[20:56] <d3vnu11> i think that the us.archive.ubuntu.com mirrors have a broken Packages.gz file
[20:57] <sweltman_> D3v, to be honest, I have no idea where to find them, I didn't see any on the GUI
[20:57] <d3vnu11> did you follow a howto on upgrading?  just curious what method you used
[20:58] <sweltman_> d3vnull, I clicked on the upgrade to 10.04 LTS on software update tool.
[20:59] <sweltman_> d3v, I am using the server version only as the base.  It's also running gnome on GUI + desktop pkgs
[21:00] <d3vnu11> ah, sorry I don't know how that reports errors, so I'm not sure how much help I could be
[21:00] <sweltman_> thats ok.  thank you anyhow for asking!
[21:00] <d3vnu11> no prob
[21:01] <sweltman_> #ubuntu-bugs
[21:32] <gurudrew> Hello all. I'm running 9.10, and trying to set up a virtual server environment using KVM. I'm trying to install Windows Server 2008, for which I apparently need an X-friendly environment. So I did apt-get install gdm
[21:33] <gurudrew> Now my system goes to grub and has no option to boot, goes straight into grub shell
[21:33] <gurudrew> Inserting my 9.10 CD to recover, on my list of recovery options, when I select my partition, it tells me it's not recognized
[21:33] <gurudrew> how can I check/repair a drive from recovery console?
[21:37] <pedahzur> gurudrew: Assuming you can access the drive, 'fsck /dev/sdX' where X is the device letter should do it.  Might have to modify that for special devices such as RAID controllers, etc.
[21:38] <pedahzur> gurudrew: That is a bit odd though...installing GDM shouldn't keep you from bootin.
[21:38] <gurudrew> May have been a bad shutdown
[21:38] <gurudrew> fsck, though, thank you
[21:39] <pedahzur> oh!
[21:40] <^peanut^> pedahzur: I just told him the same thing in #ubuntu =)
[21:40] <pedahzur> gurudrew: Sorry, You'll probably need /dev/sdXN...N is the partition number.
[21:40] <gurudrew> aye
[21:40] <pedahzur> ^peanut^: Yeah, well, when we're desperate, we often scatter shot our questions. :)
[21:40] <^peanut^> pedahzur: agreed
[21:41] <^peanut^> ?
[21:41] <pedahzur>  ^peanut^: I know I have.
[21:41] <quakenul> alright, using unetbootin instead of the universal usb installer i was able to get the setup running
[21:41] <^peanut^> pedahzur: yea me to over the years
[21:41] <pedahzur> quakenul: Glad to hear it!
[21:41] <quakenul> also updated the wiki to reflect that in case someone else has the same problem
[21:41] <quakenul> thanks for your help guys
[21:42] <quakenul> see ya :)
[21:44] <gurudrew> weird
[21:45] <gurudrew> getting "device or resource busy"
[21:46] <^peanut^> gurudrew: did you umount /dev/sdX
[21:46] <gurudrew> yeah
[21:46] <gurudrew> weird
[21:46] <^peanut^> gurudrew: did you reboot in single usermod? with the -s option passed to the kernel
[21:47] <gurudrew> No, I didn't, I booted into repair mode
[21:47] <^peanut^> try: fuser -kuc /dev/sdX
[21:47] <gurudrew> When I boot directly it goes into grub shell
[21:48] <gurudrew> no fuser
[21:48] <gurudrew> only fusermount
[21:48] <gurudrew> :\
[21:48] <gurudrew> gotta go, meeting, will be back shortly
[21:48] <^peanut^> k
[22:20] <webPragmatist> whats the consensus… are private keys normally per domain or just one for apache
[22:20] <webPragmatist> seems like a pain in the but generating it per domain… but in the event your private key was confiscated you wouldn't have to pay per signed cert
[22:22] <pedahzur> Has anyone used the SMB LDAP tools in Ubuntu Server? When I run them, it barfs with: "Unable to open /etc/opt/IDEALX/smbldap-tools/smbldap.conf for reading !"  This is smbldap-tools 0.9.5-1.
[22:28] <gurudrew> back
[22:39] <gurudrew> booted with -s
[22:39] <gurudrew> getting "device or resource busy" still
[23:06] <Jordan_U> gurudrew: Do you get an error message before the grub shell with a normal boot?
[23:08] <gurudrew> no
[23:08] <gurudrew> I ditched it and started a fresh install
[23:08] <gurudrew> nothing of importance on the install
[23:08] <gurudrew> probably some underlying old issue anyway
[23:09] <gurudrew> I want to use Ubuntu as the base of a VM
[23:09] <gurudrew> 1 Ubuntu server, 1 Windows 2008 server
[23:09] <gurudrew> Is KVM my best bet?
[23:09] <corpse> Hi im running ubuntu lucid server on an htpc. I just istalled a ATI Radeon 4350. I installed the catalyst 10.6 file but when i run fglrxinfo i get: Error: unable to open display (null)  I am not quite sure what to do from here. aticonfig --list-adapters shows 0.1:00.0 ATI Radeon HD 4300/4500 Series
[23:15] <pedahzur> Ah..."problem" with ldap tools is that they aren't configured at install time. https://bugs.launchpad.net/ubuntu/+source/smbldap-tools/+bug/126038
[23:48] <EvilPhoenix> does anyone know how I can specify where iptables logs to?  by default i think it logs to syslog
[23:56] <yu> Hi how do I make it so if a user creates a public_html directory in their home, they automatically get domain.tld/~user?