[00:02] is the xen guy here? [00:02] hey folks -- i just provisioned a new host via Ubuntu Orchestra. its online and ready for login. but... what is the user/password that Orchestra (or cobbler?) sets by default? [00:03] Azrael: try ubuntu/ubuntu .. .if you didn't change the preseeds I think thats it [00:03] shaweet [00:03] thought i tried that [00:03] worked though [00:03] thanks SpamapS [00:06] spamaps i accidently closed my irssi terminal what who was that guy who knew xen? [00:07] brianherman: it's zul [00:07] thanks [00:07] wait like from ghostbusters? [00:07] nvm [00:07] brianherman: zul, but he's out for the weekend most likely [00:07] haha [00:07] yes he's a big staypuft marshmallow man [00:07] * SpamapS hugs zul [00:10] lol [00:11] hey SpamapS ... where in orchestra/cobbler could i change the default users and passwords created on a newly deployed system? [00:17] Azrael: in the pre-seeds there is a default password hash [00:17] Azrael: kickstarts in the cobbler menu [00:26] quit [00:28] SpamapS: nice. looks like i'd just have to make my own. [00:33] hallyn: seccomp2 in oneiric does not support seccomp, in precise it should have exec support, /me needs to verify [00:34] jjohansen: thanks, precise is what i was wondering about [00:34] jjohansen: basically whether i'll be on the hook for the lxc exploit items in the blueprint :) [00:35] Daviey, hallyn: yeah we are working on the extensions, to apparmor. I should have a tree up next week sometime with the fake stack, and then I will focus on getting the mediation extensions [00:35] hallyn: :) [00:36] jjohansen: do you think you and stgraber and i should be getting together to discuss how much more coverage we'll still need then? [00:36] Daviey, hallyn: I am off monday too, can we meet tuesday [00:37] hallyn: yeah, we should talk, so things can get prioritized [00:37] jjohansen: sounds good, thx [00:39] tuesday sounds good [00:41] Azrael: there's an open bug to make those edittable [00:42] SpamapS: ahh. i think i'd just create a new profile with a new kickstart script, based on the original. [00:43] SpamapS: are newly provisioned hosts supposed to be automatically added to the Orchestra nagios installation? [00:44] Azrael: thats the dream, but I don't know if it works in 11.10 [00:45] interesting [00:45] Azrael: we've been focusing a bit on using juju for that kind of thing [00:45] probably a good idea [00:45] so [00:45] the oneiric and oneiric_juju profiles [00:45] whats the difference? [00:46] the juju profile has a late command that lets juju install itself [00:48] nice, thanks [00:49] can juju use orchestra as a provider? [00:49] ooh yes it can [00:49] very very interesting [00:49] * Azrael senses the power [00:50] would be cool if the XCP and VirtualBox providers were built in [00:50] and KVM [00:51] so i rebooted my computer and there is this new menu with xen4.1 [00:51] problem solved [00:51] in grub [00:52] :) [01:04] brianherman: woot [01:10] howdy [01:10] virt manager doesnt see xen? [01:10] ... [01:10] i can use xm though [01:32] ls [02:40] What anti-virus is recommened for running a 11.10 Apache2 server? [02:41] none [03:58] what does ubuntu rename the httpd binary to? [03:59] it doesn't, at least not from upstream debian -- /usr/sbin/apache2 [06:24] Hello [06:25] Is there some kind of bug in iptables on Ubuntu Server 11.10? I upgraded my Ubuntu-based "wireless router" to it and IP Masquerading just would not work whatsoever. So I had to re-install 11.04. [06:26] not of which I'm aware [06:26] did you check whether ip_forward remained 1? [06:27] Yes. It was also enabled in sysctl.conf [06:27] it's possible your sysctl configuration was modified in the upgrade [06:27] did you check that your rules were enabled? [06:27] with iptables -L [06:27] It worked, strangely, if I connected eth0 to a "real" wireless router. [06:27] But not with eth0 connected to the cable modem/ [06:28] However the exact setup works fine on Natty [06:29] my router is still running lucid, but I've set up masquerading on my oneric laptop a couple times as a temporary measure while working [06:29] I even tried a fresh install. [06:29] Complete format and install from scratch. [06:30] hostapd worked for Wi-Fi connection, dnsmasq worked for DHCP, so I was able to connect to it. But the packets simply were not being forwarded. [06:31] I set up iptables through Webmin to be sure it wasn't my own fault. [06:31] Oh by the way something ChatZilla just reminded me of... [06:32] do you normally run webmin? [06:32] Yes [06:32] !webmin | shane91c [06:32] shane91c: webmin is no longer supported in Debian and Ubuntu. It is not compatible with the way that Ubuntu packages handle configuration files, and is likely to cause unexpected issues with your system. [06:32] that's probably the issue [06:32] Oh, well it works 100% fine with Natty. [06:32] I guess I will leave my setup as it is for now. [06:32] yeah, but installing webmin is kind of like cutting a "warranty void if removed" sticker [06:33] we can't really help you because webmin breaks things [06:33] Well it's working fine for now, system has been running flawlessly for over a month. [06:34] If I ever feel the need to upgrade I will try without using webmin. [06:34] brb [06:45] Back [06:46] Another question, how can get my system to use my domain name as it's FQDN, as opposed to the ISP-assigned dIP-AD-RE-SS.cgocable.net [06:50] hostname -A shows that the server's FQDN is d67-193-84-52.home3.cgocable.net. I want it to be my own domain. [06:51] If I put "option domain-name "mydomain.com"" in /etc/dhcp/dhclient.conf will that work? [07:00] no [07:00] that's up to the registrar, which is most likely your ISP [07:01] if you want it to say something different you'll have to buy it as a service from your ISP, or whoever owns your IP space [07:03] PTR records are looked up as your.ip.in-addr.arpa, which is authoritative to whomever owns the IP in question [08:36] IIIiiii been fixing multi-arch bugs... aaaaalllll the live long day [08:48] hi all, I,m have'n a pxe install problems for weeks now, when running a unattended pxe install on ubuntu 11.04 server ( "tasksel tasksel/first multiselect standard" in the preseed file) on the first reboot of the fresh install all i get is a blinking cursor. But when i change the above line in the preseed file to ( tasksel tasksel/first multiselect ubuntu-desktop ) everything works fine? ( but i get a desktop install.... ) [08:49] this user has the same issue ..... but used kickstart http://ubuntuforums.org/showthread.php?t=1874027 [08:49] any help would be brill [09:03] billy_: you get a desktop install from which iso? [09:06] i use netboot and then a mirror [09:06] the heanet.ie mirror [09:16] weird.. well I'm passing out else I'd help :p [09:17] * SpamapS falls into an instant slumber [09:17] [09:20] np. thanks for you time [09:52] morning [09:52] any idea how I can list how much of a process's memory is swapped out? === sjefen6_ is now known as sjefen6 [11:06] New bug: #896551 in samba (main) "package samba 2:3.5.8~dfsg-1ubuntu2.3 failed to install/upgrade: ErrorMessage: package samba is not ready for configuration cannot configure (current status `half-installed')" [Undecided,New] https://launchpad.net/bugs/896551 [14:32] My computer sometimes hangs at startup due to 'degraded raid array' but rebooting it (sometimes a couple of times) and it comes up fine and /proc/mdstat lists the array as [UU] [14:33] /etc/mdadm/mdadm.conf contains the line: ARRAY /dev/md0 UUID=f0a0371d:12376ea7:4c4ad349:XXXXXXXX [obfuscated] [14:33] blkid on the relevant partitions returns the same UUID [14:33] But it seems to be that when /dev/sdb is detected as /dev/sdf instead then it doesn't work. [14:35] Is it an issue with grub, perhaps? [14:39] Going to try probing the issue some more. [17:28] what is the best soft to do virtualization on ubuntu server? [17:31] best is a matter of opinion, but the included/supported option is KVM [17:31] thanks!!! [18:09] hi, what do you recommend: virtualbox or kvm? [18:10] for server kvm, for just your home fun computer, virtualbox [18:11] thanks!! [18:11] :) [18:12] andygraybeal, why kvm for servers? [18:13] KVM is designed around running on a headless server with a bunch of VMs all the time [18:13] virtualbox is designed around firing up a virtual machine on your desktop to use it, then shutting it down [18:13] both can do either task, but they're better suited as I explained [18:14] thanks!!! [18:15] i run kvm on my poop at home too.. i like it [18:15] but i wouldn't recommend it to a friend just getting started with virtual machines :) [18:16] virtualbox is more about easy desktop setup, and it has fairly weak networking capabilities [18:18] KVM's a little harder to get going, but it's a more robust system and easier to make it run in the background [18:19] thanks!! === gustav- is now known as beerbro [19:13] Does anyone recomend a SNMP/MRTG tutorial for Ubuntu Server? Everything I find is outdated, the conf files even don't match. [20:01] kirkland: do you know of a way to do 'tmux --splitw command' and have the tmux call wait until command is finished? [20:37] Can someone please help me troubleshooting a connectivity issue on my 11.10 server? It's hardwired to the router, sees the router at 192.168.1.1 and can ping it, however it can not ping any other network devices nor can it make any connections to the internet (i.e. apt-get update completely fails) [20:38] kirkland: nm, think i'm all set, thx [20:47] Duvrazh so you mean it can't ping anything beyond the router or you can't ping anything on the local subnet? [20:47] both [20:47] it can only ping the router [20:48] but i can ssh into it [20:48] and the machine has 1 NIC which is connected to a switch, everything connects to the switch including router and it's an unmanaged switch? [20:49] 1 nic, router is a WRT610N with DDWRT [20:50] any firewalls enabled? [20:50] iptables -L [20:50] not that i know of, but ill check [20:50] input forward and output policies all to ACCEPT [20:51] might be worth completely disabling it just to rule it out [20:51] i.e. remove all rules [20:52] I was using webmin, can you help me with that command? [20:52] I never really used iptables before and honestly I'm not sure why they're active, had to be a dependency thing or something [20:53] see: http://www.cyberciti.biz/tips/linux-iptables-how-to-flush-all-rules.html [20:53] t/y [20:53] don't bother creating the script, just run them manually [20:54] saves time [20:54] rodger [20:55] rules flushed no errors [20:55] ping to iMac fails [20:56] what subnet mask are you using (make sure to check on both machines to ensure it's the same)? What is the IP for the mac and the machine that you're pinging from, also, does the Mac have a firewall enabled? [20:57] mac has no firewall, same subnet, mac ip = 192.168.1.40 and server is 192.168.1.42 [20:57] what's the mask? [20:58] 255.255.255.0 [20:58] both run on full auto dhcp [20:58] no virtual networks on router [20:58] hmm, and the router is also acting as the switch? [20:58] i.e. it's got a switch built into it [20:59] is this a wired network or wireless? [20:59] wired, both [21:00] have you got another switch that is just a switch? [21:01] it's a linksys wireless router but yes it's acting as the switch. the only advanced network features employed on these two devices or any others present is manually defined dhcp by MAC address to enable auto dhcp on clients to fall into a certain order (so I can remember what to ping since hostnames are unreliable for me sometimes) [21:02] hmm, you said earlier you can't ping the imac, presumably the imac can't ping your computer? [21:02] ill try [21:02] mac can ping server [21:03] server still can not ping mac [21:04] maybe it's time to reinstall the os. all data is on a 5-drive raid-5…. I could always remount that [21:05] hmm, shouldn't be necessary, have you tried rebooting? [21:05] that's what caused this [21:05] i powered down every electronic in house because network was running slow. i wanted a fresh start, then i noticed this server wouldn't connect [21:06] I have another 11.10 server not connecting either and it's making think I should roll back to an LTS [21:06] have you tried using an 11.10 live cd? [21:06] if one exists... [21:07] No I have not. I don't think there is an 11.10 live, I think 11.04 or 10.10 is the latest… based solely on errors from unetbootin [21:07] a reinstall would not be unwelcomed [21:07] it would help me fix a folding@home problem [21:08] only my ps3 has been getting work units over the past week [21:08] fair enough, an LTS version would be better [21:08] all 3 ubuntu computers plus iMac have not been getting them [21:08] yeah… I'll just do that (I'm a quitter) [21:08] hehe [21:08] if you can provide SSH access I'll take a look [21:08] but otherwise yeah just reinstall an LTS release [21:11] l0n: may I PM you? [21:12] guntbert sure [21:40] hi, I'm trying to launch a virtual server (on eucalyptus) with ubuntu 11.10, but it fails to boot: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0) [21:40] any ideas on what could I do to make it run? [21:41] before that, it says: [21:41] VFS: Cannot open root device "sda1" or unknown-block(0,0) [21:41] and [21:41] Please append a correct "root=" boot option; here are the available partitions: [21:41] (but it lists no partitions after that) [21:44] so you have two OSes installed on two partitions and you want to launch a 2nd partition as a virtual? [21:45] no... [21:45] I'm running eucalyptus [21:45] I'm trying to install oneiric on a virtual machine [21:46] what vm software? [21:46] I'm running the latest Eucalyptus version in the Natty repositories [21:46] (the host machines run Natty) [21:48] oh snap. sorry, Eucalyptus is out of my limited knowledge pool. I wish you luck though. [21:51] bfreis I've also never used Eucalyptus but from the docs, it looks like it can use either xen or KVM, which are you using? [21:52] let me check [21:52] (i'm using the default, just making sure which one it is) [21:52] kvm [21:53] Apparently the device containing the filesystem is not /dev/sda1. However, I have no where I could change it, nor what else could it be... [21:54] I've simply downloaded the cloud image from the website, published it to my cloud (doing exactly as I've done on the original and my tweaked 11.04 images, which work fine), but I can't launch an instance [21:54] well, I can launch an instance, but it won't finish booting, because the kernel fails to mount the root filesystem [21:55] I don't have the slightest idea on how could I get to a GRUB menu and try to list the disks in this virtual environment... [21:55] well you can change the root device using the root=/dev/device kernel parameter but I am not sure what to change it to [21:56] Yeah, the correct device is one problem [21:56] The other is: where could I change this parameter? [21:57] hmmmm [21:57] well, you could boot off a live-cd, mount the disk, change the grub timeout to something other than 0 then set the root param [21:58] inside the image, in /boot/grub/grub.cfg, I see this: [21:58] linux /boot/vmlinuz-3.0.0-13-generic root=LABEL=cloudimg-rootfs ro console=ttyS0 [21:58] it is looking for the filesystem by label [21:58] yeah [21:58] well, the kernel somehow inferred it to be /dev/sda1 [21:58] that might just be a default that it falls back on if it can't find a partition with that label [21:59] the label is ok [21:59] the image is correctly labeled [22:00] in the virtual machines config, is the disk assigned to the image? [22:00] something like hd = blahblah.img [22:01] It must be, but I have no idea where this configuration is [22:01] All I do to launch an image is: euca-run-instances blahblahblha image-id, where image-id is an id given by the cloud system after I publish the image [22:02] (and the procedure to publish an image is what I do exactly the same on the original Natty and Oneiric images, and it works for Natty but not for Oneiric) [22:03] that's a bit tricky, have you thought about just using plain old kvm rather than kvm wrapped in something else? [22:03] can euca be used to distribute a single application as opposed to an os? [22:03] such as folding@home? [22:03] or the origami frontend? [22:04] I have no idea how to use plain kvm... and I think it would be much more difficult to manage the distribution of virtual machines on a set of many physical machines [22:04] Duvrazh, no [22:04] damn [22:04] eucalyptus is a cloud system that is supposed to work just like Amazon Web Services [22:04] oh [22:04] okay [22:05] you have a bunch of machines, you install a lot of packages in one (which you call "cloud controller"), you install the node controller packages on the machines that will host the virtual machines, you exchange ssh keys (all of this is done automatically from the Natty Server install CD), then you are almost ready to go [22:06] then you have your own Amazon Web Services EC2- and S3-like cloud [22:07] It then automatically distribute the load on the worker machines for the new instances you launch, it manages the distribution of images, etc [22:08] however, there's something broken in Eucalyptus/Oneiric that won't allow me to boot it up :/ [22:08] have you tried contacting Eucalyptus support? [22:08] It is sad not to know how to launch an instance after you spent some hours tweaking the image... [22:08] Lol [22:08] yeah [22:08] well, not the "support" [22:08] i've been to #eucalyptus [22:09] but well, it's almost useless, people won't respond there [22:09] and they couldn't help at all? [22:09] oh :/ [22:09] there's 37 people in there, no one will answer [22:09] Actually, they are quite good when they answer [22:10] some months ago, when I first installed an Eucalyptus cloud, I found a bug related to time synchronization between node controllers and the cloud controller that would simply destroy the network communication on the cloud [22:10] I asked there, in less then a week they have published an update in the ubuntu repositories fixing the bug [22:11] well, maybe there just isn't anyone around at the moment, have you tried asking at diff times of the day? [22:11] I was quite impressed acutally... but now, since 2 days ago I'm trying to talk to them without success (there's another bug bugging me, but completely unrelated to Oneiric's bug) [22:11] yeah, I will keep trying though [22:21] hey [22:21] for you who know KVM [22:21] maybe the command line executed to Eucalyptus could give some hits: [22:21] /usr/bin/kvm -S -M pc-0.14 -enable-kvm -m 626 -smp 1,sockets=1,cores=1,threads=1 -name i-4F5B08DC -uuid 4f437ae4-7c18-e624-c7ff-7cac9fa1c008 -nographic -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/i-4F5B08DC.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=readline -rtc base=utc -boot c -kernel /var/lib/eucalyptus/instances//admin/i-4F5B08DC/kernel -append root=/dev/sda1 console=ttyS0 -device lsi,id [22:21] =scsi0,bus=pci.0,addr=0x3 -drive file=/var/lib/eucalyptus/instances//admin/i-4F5B08DC/disk,if=none,id=drive-scsi0-0-0,format=raw -device scsi-disk,bus=scsi0.0,scsi-id=0,drive=drive-scsi0-0-0,id=scsi0-0-0 -netdev tap,fd=19,id=hostnet0 -device e1000,netdev=hostnet0,id=net0,mac=d0:0d:4f:5b:08:dc,bus=pci.0,addr=0x2 -chardev file,id=charserial0,path=/var/lib/eucalyptus/instances//admin/i-4F5B08DC/console.log -device isa-serial,chardev=charserial0,id=seria [22:21] l0 -usb -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x4 [22:21] got it with ps [22:22] hmm, there it says root=/dev/sda1 [22:23] bfreis: what was your question exactly? [22:23] Why can't I boot an Oneiric image on Eucalyptus? [22:24] anything from euca-get-console-output ? [22:24] VFS: [22:24] VFS: Cannot open root device "sda1" or unknown-block(0,0) [22:24] bfreis: ugh, I'm horrible with kernel stuff. ;) [22:25] I can boot a Natty image, but not an Oneiric image [22:25] Well, Natty is 2.6, Oneiric is 3.0 [22:25] (I'm not sure it should matter...) [22:25] Yeah, I know nothing about kernel stuff either [22:26] bfreis: I don't know if euca supports aki's .. but maybe try an oneiric image with a natty kernel? [22:27] oh... [22:27] you mean, take the 2.6 kernel from natty and use it on oneiric? [22:27] bfreis: unfortunately with Euca dropping to universe, the level of testing on it has gone down by an order of magnitude. :-P [22:27] yeah [22:27] I'd be happy to try the new cloud system [22:27] However the installation procedure is WAY TOO COMPLICATED! [22:28] We're kind of in this weird space right now where OpenStack is just starting to grow production readiness, and eucalyptus is stagnant.. [22:28] yeah [22:28] bfreis: you can try it with orchestra+juju .. much simpler that way. ;) [22:28] I've tried the live image, it was very easy but... it's just a play ground, I can't use it for real [22:28] Since you guys are on the cloud topic, can you make a recommendation for distributing a program across multiple machines? [22:28] Now, the problem with orchestra+juju is that I don't have 6 machines [22:29] bfreis: http://cloud.ubuntu.com/2011/10/ubuntu-cloud-deployment-with-orchestra-and-juju/ [22:29] it seems that orchestra+juju can install only one service per physical machine, which sucks, since there are services which are veeeery simple and could go together [22:29] Duvrazh: any one thread of a program will have to be on one machine at a time. If your program can do two things at once, there are about 1000 ways to get that done. ;) [22:29] bfreis: right, thats being worked on for 12.04 ... probably will land in Jan. [22:29] great! [22:30] bfreis: it works fine for giant deployments, but for the 2 - 5 server cluster.. its useless. :( [22:30] yeah hehe [22:30] It is for a test environment [22:31] I have 4 machines for it [22:31] Basically you can do 1 server, or 6 .. [22:32] Duvrazh, what do you mean by "distributing a program"? You mean, you are writing a software and you want to distributed the load across multiple machines? [22:33] * SpamapS kind of wishes he could step back in time and play with 4 machines just to solve one problem.. instead of trying to solve the 1000+ problems in Ubuntu server. :) [22:33] any software, particularly donating my spare cpu cycles, so anything from boinc to folding@home or origami. The goal would be one client to run on all the cores in my house…. instead of running 12 separate clients [22:33] oh [22:33] it would be a convenience for switching from one client to another [22:33] You want to have something like one big, fast core, made up from multiple cores? [22:34] yup [22:34] I don't think it is feasible... [22:34] I don't either, that's why I like asking [22:34] it's an anomaly in the back of my mind [22:34] lol [22:34] It can be done but I lack the programming prowess [22:34] I doubt it [22:34] I also lack the linux skills [22:34] You simply cannot have more clock speed than your real clock speed [22:35] What you can do with multiple cores is run more code at the same time (but at the nominal clock speed) [22:35] Duvrazh: thats really not how distributed computing works. :) [22:35] I am mis-speaking [22:35] look at the FAH program [22:35] it runs an instance for each core on each processor [22:35] Yeah, that's what can be done [22:35] quadcore machine = 4 pids [22:35] I have 15 machines [22:35] You can't "add up cores" [22:35] You can certainly make a program detect how many cores there are and run that many threads/processes [22:36] I total something like 76 cores [22:36] I want one client to run on them all [22:36] 76 pids [22:36] Your best bet is to write something that would connect to each machine and launch the original FAH software [22:36] yeah, you'd basically manage that the same way you manage load balanced webservers [22:37] how would i pipe back the status output though? [22:37] that part kills me [22:37] * SpamapS must go afk.. will rejoin the discussion in a bit [22:37] Actually, this should be very simple to do [22:37] what do you mean by status output? [22:37] what it prints to the console? [22:37] yes [22:37] like i said < linux newbie [22:37] well [22:38] you could try something like this [22:38] ssh to the server [22:38] launch a FAH instance with: [22:38] nohup FAH-COMMAND > fah-instance-1-stdout.log & [22:38] (launch all the instances you want) [22:39] then disconnect from ssh [22:39] when you want to see the output, you could do: [22:39] ssh machine cat fah-instance-N-stdout.log [22:39] Now, you just have to write some code to make it automatic [22:39] hmm. you're right. [22:40] hello, how i can connect using remote desktop to ubuntu server on EC2? [22:40] gllera, won't ssh be enough? [22:40] Did you install a GUI on your ubuntu server? [22:40] bfreis: thanks for the food for thought [22:40] bfreis, yes but is only to know ;) [22:41] Duvrazh, you're welcome! [22:41] Duvrazh, if you ever write this code, be sure to publish it somewhere, people could benefit from it! [22:41] Hey people. I'm trying to setup a DNS server, but I need some clarification on the zone files and how it all works. [22:44] hmmmm, coming back to my boot problems, I see that some of my EC2 instances (in AWS) have /dev/sda1 as the root filesystem, and others have /dev/xvda1... What is xvda? never seen it... [22:45] xvd a is a virtual address is it not? [22:45] damn you autocorrect [22:46] I know that when using Xen I had to boot my kernel from /dev/xvda but it was a virtual image and a virtual disk [22:46] hmm [22:46] part of a retarded event of seeing how many machines I could load inside of each other to piss off an old desktop I had sitting around [22:47] (I removed the fans before that test) [22:48] PASTE: Boot Configuration [22:48] Enter Linode Configuration Profile in your Linode Manager. Change Kernel to pv-grub-x86_32 or pv-grub-x86_64, depending on installed kernel and userspace. Unless you developed a complex configuration set your root device to /dev/xvda. [22:48] from a google [22:50] New bug: #896715 in mysql-5.1 (main) "package mysql-server-5.1 5.1.54-1ubuntu4 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/896715 [22:53] * SpamapS returns [22:53] jakupl: what do you want to do with said DNS server? [22:57] SpamapS: Well, I have used the DNS benchmark utility from grc.com to find the fastest dns server, and all that is well and good, but I thought that it would be faster if I used a local dns server. [22:58] jakupl: thats fine for *local* requests, but what matters for DNS is latency. are all of your clients local? [22:58] clients? [22:58] bfreis: re your ec2 question.. EC2 is all xen.. euca is kvm. [22:59] jakupl: yes, the things that will be asking your DNS server questions. [22:59] SpamapS: yes. It will only be used locally, but I don't need local domains. I only want the dns server to handle external dns. [23:01] SpamapS: and altso, I really like DHCP, so I still want that. I have rules in my router so that the ubuntu server always has the same IP. [23:01] (router is dhcp server.) [23:04] http://serverfault.com/questions/335076/cant-launch-oneiric-x64-instance-on-eucalyptus [23:04] in case someone has a brilliant idea :D [23:06] jakupl: so your clients are.. your home machines? [23:06] jakupl: and they will just be using the dns server to query external domains? [23:07] SpamapS: exactly right [23:08] is that a bad idea? [23:08] jakupl: honestly, I get better speed just pointing all my machines at google's DNS servers [23:08] since they almost *never* have cache misses [23:09] dnsmasq is good [23:10] it basically just forwards name requests to other servers you specify (including google if you so choose) [23:10] SpamapS: hmh. That's so boring [23:10] and layers in your local stuff as well [23:10] in fact, you just add it to the hosts file [23:10] and you get local name resolution [23:10] its very lightweight [23:11] [23:11] SpamapS: however. local dns server would be faster for cached domains. [23:12] nebajoth: yeah, i've heard other people mention dnsmasq. Maybe I should look into it. [23:12]