[00:24] anyone using ubuntu on ec2? i'm trying to decide which version to use [00:25] also-- anyone know where i can get the kernel source w/ any patches used to create ubuntu's ec2 kernels? [00:25] can't seem to find that anywhere... [00:30] Are there any good NAS server management software packaged for Ubuntu? [00:45] bnjmn: What do you mean "which version to use"? [00:45] bnjmn: The EC2 kernel source may or may not be available, but it isn't much use since you can't build your own kernels for EC2. [00:51] New bug: #288379 in virt-manager (universe) "4th network extremely slow/broken" [Undecided,New] https://launchpad.net/bugs/288379 [00:52] I'm trying to install Ubuntu 9.04 server onto a computer that has no CD drive, via a USB stick. The installer does not like this, and wants me to give it a CD drive. Google is not pointing me in the right direction. Does anyone know how to do this? [00:58] i want to build my own kernel module [01:09] if i could just get the config that would be great [01:11] New bug: #335809 in kvm (main) "KVM activates (GNOME) Screensaver when inside SDL window" [Low,Fix released] https://launchpad.net/bugs/335809 [01:21] New bug: #362550 in kvm (universe) "ignores -no-shutdown" [Low,Fix released] https://launchpad.net/bugs/362550 [01:21] New bug: #365569 in kvm (universe) "KVM segfaults with a specific disk image." [Low,Invalid] https://launchpad.net/bugs/365569 [01:21] New bug: #370270 in kvm (universe) "ttyS0 pass through in guest running in KVM doesn't work and eventually hangs" [Low,Fix released] https://launchpad.net/bugs/370270 [01:26] New bug: #390278 in virtinst (universe) "Windows 2000 server won't install on kvm84/hardy" [Undecided,New] https://launchpad.net/bugs/390278 [01:28] bnjmn: Yep, kernel modules are buildable for EC2. What EC2 kernel are you using? [01:37] hi [01:38] i'm looking for some help with ubuntu server 9.04 "minimal virtual machine" option [02:10] soooo.... something happened to Samba so it doesn't start on my server at startup. It doesn't fail out, it stops the startup process so the server doesn't start. [02:20] lukehasnoname: echo /etc/rc?.d/???samba [02:55] anyone around? (with lib mono exp & apache) [02:56] New bug: #333000 in kvm (universe) "qemu/kvm incorrecty enters fullscreen when using nvidia twinview" [Wishlist,Incomplete] https://launchpad.net/bugs/333000 [02:56] New bug: #353995 in kvm "BLTs appear broken - moving blocks on the remote leaves cruft" [Wishlist,Fix released] https://launchpad.net/bugs/353995 [02:56] New bug: #393430 in kvm "kvm: please prefer ALSA over OSS" [Wishlist,Fix released] https://launchpad.net/bugs/393430 === monteith is now known as monteith_afk [04:11] hello all, i have 'git' up and running on my Ubuntu 9.04 server perfectly. Only issue is i'd like to start the daemon automatically on each reboot. Perhaps i'm just searching for the wrong thing, but i'm not finding where to add that into Ubuntu's normal boot\startup process, any ideas? [04:17] If there's a daemon, there should already be an init script. [04:17] Which package provides the daemon? [04:18] git itself. You can launch it as a daemon, via: sudo -u git git daemon --base-path=/home/git/repositories/ , i just need that to launch when i startup. [04:18] OK, you mean "git-core". [04:19] That has no init script, so you will need to write one. [04:19] It is probably simplest to put it in /etc/rc.local, which runs at the end of the boot process. [04:19] You should also use start-stop-daemon rather than sudo -u git. [04:23] awesome, found a couple things on it so far, let me give it a whirl. (Good thing for test virtual servers, have a feeling i'm gonna get this wrong before right :). ) [04:24] You can run /etc/rc.local by hand until you get it riht [04:24] *right [04:52] cool, so i got that working w/ rc.local. Now while i investigate the actual 'init script' i did notice it launches the process as user 'root' i definitely want to change that to user 'git.' Any ideas? My searches are coming up with a lot about 'fork' or Perl function, not just a generic way to launch /any/ process as another user. [04:55] n/m, got it. Awesome thanks for the help :) I'll continue to read this tutorial on init scripts. Ultimately that's what i want. [05:00] altf2o: you need to pass --user to start-stop-daemon [05:00] (--run-as?) [05:01] ahh, perfect... [05:03] just curious, what does everyone use for ftp server and why [05:03] software wise [05:04] android60, openssh-server [05:04] because ftp is insecure [05:06] well, I am just looking to use the ftp on my lan, of course i use winscp and openssh to do over internet transfers :) [05:06] its b.c some of the programs i have will connect either via samba or ftp, and id rather avoid samba [05:06] over a lan, smb or nfs are better choices [05:06] For anonymous read access, HTTP beats FTP. [05:07] For uploads, use SFTP (OpenSSH) [05:07] and yes, read-only, http is great [05:07] i also have apache because as a hobby i do websites with php etc [05:07] Both in terms of security and protocol design, FTP has serious limitations that make it unsuitable for use in production. [05:07] i just finished setting up Samba on my little test server, i was amazed how easy it was. I changed 1 line, added my shares, added libpam-smbpass, rebooted, done. [05:08] ftp was originally designed before the concept of firewalls existed [05:08] as such it's completely obsolete and difficult to work with [05:09] altf2o: what did libpam-smbpass do? [05:09] libpam-smbpass makes samba authentication less difficult [05:09] android60: Samba keeps it's own db of user\pass's. That PAM module syncs Samba w/ local Linux user\pass's. [05:10] I've tried samba but I was having user permission problems. and I don't like having to add a user to samba access, i wish i could just create a user on the system and have it automatically allow them access to specified shares [05:10] well then, [05:10] So if i add user 'steve' instead of having to then do: smbpasswd .... to give him a Samba password, it updates it for me. I maintain one, much easier. [05:10] thats exactly what i want [05:10] i better write it down before i forget [05:11] is there any setup for libpam-smbpass or once installed its good to go [05:11] android60: what i've done 2 things. If i'm creating private shares, i add it to my home directory, limit access to myself. If i need a public share for multiple people i create a group, add them to that group, create the directory i want to share and assign it to that group. I usually: chmod 0775 the folder and just use "read only = user, list here" for folks i want to be 'read only' on it. [05:12] samba's magic homes share is incredibly useful [05:12] you can usually just use that, plus a public share [05:12] android60: yep, should be good to go. One thing i noticed is it won't update for me until that user 'logs in' even: su steve , did it. Can't figure out how to manually say "please re-sync" [05:12] agrees w/ homes share. [05:13] ya, see im mostly using mine as a nas, so all relevant files and folders are not on the system hard drive. that was another question i had [05:13] was is there a way to have the home directories point to like /media/MainStorage1/Documents/usernames_here [05:13] b.c i would like encrypted home directories [05:13] just not on my 40gb system drive [05:14] you can create a separate /home partition [05:14] ya, i just kinda didnt want to do that [05:14] that's the usual way to do that, or use ecryptfs for encrypted homes [05:14] i have 2 1TB drives, one is for archive via cron job rsync [05:14] you could use symbolic links, but be aware that if the target is ever missing, it will break logins [05:15] ya thats what i was trying to avoid also [05:15] ubuntu server is great for my needs, and the system its on is only like sempron 1250+ and 1 gb ram [05:16] i think i might up ram alittle bit once i get mythtv backend going [05:16] what is probably ideal for what you want is to have a separate /home partition, and use ecryptfs for homes [05:16] putting the /home partition on a separate hard drive [05:17] ya, i might do that down the line. im not in a hurry to do that right now [05:19] does widnows support nfs? like if i want to map a network drive? [05:19] no [05:19] didnt think so [05:19] windows uses SMB [05:19] which is what samba does [05:20] there are NFS clients for windows, but I've never used them, and from what I've read, they're rough [05:20] ya same here [05:20] i didnt know if vista/win7 supported it or what [05:20] nfs is generally something to be avoided also, unless you're really taking care to set up a secured nix enviroment [05:21] oh, i wasnt aware of that [05:21] unless you're using an nfsv4 with kerberos setup, it's pretty insecure [05:22] and it uses RPC, which is somewhat painful to deal with [05:22] never really heard of any of that, but ill take your word for it [05:22] it's another one of those really old protocols that just weren't built in today's internet [05:22] samba's not much better, but on a LAN you're fine [05:22] it's also pretty much the only way to integrate well with windows [05:23] ya [05:26] Kerberos looks really nice, and not to bad to setup. That's sort of my next project LDAP + Kerberos [05:28] openssh is the swiss army knife of secure protocols [05:28] shell access, tunnels, file transfer, even mounting filesystems [05:29] all on one port [05:29] all encrypted [05:29] i'm new to it, but i don't know what i'd do without it now. [06:17] I have a guy with an issue about how to use a custom 404 page on his Apache server [06:17] can anyone help me/him out? [06:17] "i run apache built into a WAMP server for my website, but i cant figure out how to set the 404 page to a custom one in the httpd.conf file for apache" [06:28] We do not support Windows. [06:29] Try apache's channel, which IIRC is #httpd. [06:31] twb: ok, thanks for the pointer :) [06:31] there's just not many people around, so I figured I'd see if anyone here was familiar with it as well [06:59] twb: I can't even log in. It stops the boot process. Is that not a bug? Should a program be able to lock a server up like that? I know I could probably go in grub and boot to single user mode, wipe Samba or take it out of the startup list, but wtf mang [08:01] hi [08:01] New bug: #432291 in php5 (main) ""Fatal error: Call to undefined function gzopen()" on php5 5.2.10.dfsg.1-2ubuntu3" [Undecided,New] https://launchpad.net/bugs/432291 [08:01] I've been trying to set up squid with NTLM auth without succeed [08:02] it seems that my main problem is that winbind is not working properly because squid is not in the same machine as samba PDC [08:02] but I'm not able to find any site which talks about this [08:02] and all tutorials I've been reading point to the same scenario: samba and squid in the same computer [08:03] now I'm able to do a wbinfo -t and wbinfo -a domain\\user%password and it works [08:03] but wbinfo -u nor wbinfo -a work [08:15] <_ruben> bah .. why do customers still request servers with php4 installed .. its been dead for years [08:17] do you still install php4? [08:17] on request? [08:19] because dead isn't defined by whether or not there are still releases occurring [08:19] people with revenue generating apps don't necessarily convert to the latest version of a language [08:23] <_ruben> henkjan: sortof .. i completely firewall the box and them fuck things up themselves [08:31] New bug: #423465 in eucalyptus (main) "euca.p12 owned by root" [Undecided,New] https://launchpad.net/bugs/423465 === cjwatson_ is now known as cjwatson [08:47] hey all [08:48] i have a standard install of 8.04 server. how can i configure ftp? [08:49] install vsftpd [08:49] i'm finding stuff about vsftpd and proftpd but what's the difference, and which is installed by default? [08:49] ah, none is installed/ [08:49] ubuntu server by default doesn't install any services [08:49] it has no open ports [08:49] i chose to install LAMP though [08:50] and apache is on by default [08:50] you don't get ftp service with that [08:50] ok [08:52] brb [08:52] then whats the difference between the different ftpds? [08:52] it's up to you [08:52] i prefere vsfptd [08:52] some like proftpd [08:52] you can't go wrong with eny of them [08:52] any [08:52] brb [08:55] how can I get a version of a packages installed? [08:55] s/get/check [08:55] dpkg -l [08:55] thanks [08:55] dpkg -l | grep package [08:55] <_ruben> dpkg -l package ;) [08:56] or that :) [08:56] apt-cache policy package [08:57] are there no suitable alternatives to ftp for accessing files? [08:57] of course there are [08:58] sftp, scp, webdav [08:58] etc... [08:58] afs :-) [08:58] does ubuntu come with netatalk client? [08:58] or is it easy to set up? [08:59] no, n/a [08:59] i'm installing netatalk anyway for my mac [08:59] jdstrand: Yes, a quick workaround would be lovely. [08:59] how can I use ubuntu server to host any website [08:59] himanshurauthan: you should install web server [08:59] himanshurauthan: you can't host other people's websites [09:00] on ubuntu server can't I do [09:00] what is this web server then [09:00] in ubuntu? [09:00] Unfortunately, unlike Debian, Ubuntu's initrd does not include a web server. [09:01] So for what purpose I can use ubuntu server [09:01] apache [09:01] break=bottom, busybox httpd FTW [09:01] himanshurauthan: sudo tasksel install lamp-server [09:02] himanshurauthan: https://help.ubuntu.com/community/ApacheMySQLPHP [09:02] ok so after installing lamp server can I host a website in that [09:02] himanshurauthan: yes [09:03] so what all other stuffs do I need if I want to host websites on a ubuntu server [09:03] do I need some special hardware also [09:04] himanshurauthan: "I want to host a website" is not really a meaningful approach. [09:04] you need a working system [09:04] hardware depends on how serious you are [09:04] Why do you want to host a website? What are you intending to host? [09:04] and how many websites you plan to have [09:05] no it is not for hosting many sites but it is about hosting my own site [09:05] Are you intending to host an intranet CMS, a family photo album, a flash-heavy bulletin board, or what? [09:05] my business site [09:05] should I go for shared server or my own dedicated server [09:06] since I want to expand it in future [09:06] OK, then the next question is: do you want to do this yourself, or do you want to hire an expert to handle it for you? [09:06] I'm a web developer basically [09:06] You're a webdev, but you don't know how to set up a webserver? [09:06] then install lamp-server [09:07] for this what type of expert I need [09:07] none [09:07] you install it [09:07] and it works [09:07] you just need to put your website on disk [09:07] is it networking or webdev [09:07] So you're already familiar with developing with technologies like PostgreSQL (or MySQL), Apache, and Python (or PHP)? [09:07] yes [09:07] is it networking or webdav? [09:07] huh? [09:08] its not networking [09:08] himanshurauthan: OK, so really all you need to do is learn how to do the underlying system administration. [09:08] its not webdev [09:08] ok [09:08] oh, how do you access the disk? [09:08] however you want - SMB/SCP/FTP/SFTP... [09:08] If you already understand how to USE stuff like Postgres, then you can probably pick up sysadmin of those services easily. [09:08] but you need to configure server for that [09:09] Installing the lamp-server task, as already recommended, will give you a basic setup that you can then tweak if/as necessary. [09:09] hm ok [09:09] basically once you install it you can just put your files there and it will work [09:09] it's not very difficult [09:09] himanshurauthan: as for hosting your own server vs. buying hosting space or a VPS, that comes down to issues like cost and who administers what. [09:10] for external purpose means not for interanet I need server right?? [09:10] that has access to the world [09:10] then I need to go for Domain name also [09:10] himanshurauthan: for a website you will also need to sort out an adequate network connection, a domain name, etc. [09:10] btw, is there a way to install i386 packages on a 64 bit OS (its a fuse filesystem) [09:11] SockPants: in general, no. It requires fundamental changes in dpkg, which are pending. There are various workarounds, however. [09:11] SockPants: I don't know what you mean by "it's a fuse filesystem" [09:11] well that's what i'm installing, afpfs-ng [09:11] there's only a .deb for 32 bit [09:11] bye [09:12] SockPants: good luck with that [09:12] >.> [09:12] how so [09:12] SockPants: why are you still running MacOS 9? [09:12] twb: whats an alternative, and i'm not [09:13] SockPants: AFP sucks. You would be better off using CIFS (Samba) and/or NFS. [09:13] i always thought cifs sucked [09:13] and i never got nfs to work [09:13] CIFS does suck. [09:13] but if afs is that bad i'm willing to try again [09:13] But AFP sucks more unless you are stuck on MacOS 9 where it's your only option. [09:13] ok [09:13] Note: AFS is not AFP. [09:13] well then i'll try nfs again [09:14] unless that's not very good either [09:14] AFS is for people who have thousands of users around the world. [09:14] yeah [09:14] i don't [09:14] If you are in a heterogeneous environment, CIFS is by far the easiest to set up [09:14] twb: the clients are only ubuntu or os x [09:14] SockPants: what do you need a client on, other than OS X, Ubuntu and Windows? [09:15] Oh sorry, I misunderstood [09:15] so nfs would work i suppose [09:15] If you are only using Ubuntu and OS X, NFS or CIFS is appropriate. [09:15] but is nfs any good [09:15] fundamentally [09:15] I hear that NFS is much easier on recent versions of OS X. [09:15] it seems all the file sharing protocols suck [09:15] YES [09:15] Basically you have to choose which kind of suck to have [09:15] cifs sucks, ftp sucks, afp sucks [09:15] hahah [09:15] ok [09:16] but nfs sucks less than those three? [09:16] what sucks most about nfs? [09:16] OS X and Ubuntu both have good CIFS and NFS clients, and Ubuntu has good servers for both of those. [09:16] ok [09:16] Probably the suckiest things about NFSv3 are its (lack of) security and its dodgy locking. [09:17] locking? [09:17] But basically all network filesystems have broken locking semantics [09:17] alright well [09:17] SockPants: as in, stopping two computers from trying to edit the same file at the same time [09:17] if it's not a severe problem [09:17] and i don't think that will happen btw [09:17] then i'll go with NFS [09:17] e.g. dodgy locking can break OpenOffice and Firefox 3, if /home is on a network filesystem. [09:18] for security i would be tunneling over ssh anyway whenever i'm outside of the lan [09:18] This happens with CIFS and NFS, and probably all other network filesystems. [09:18] SockPants: if you are outside the lan, sshfs is very good for small files [09:18] sshfs != sftp/ [09:18] + [09:18] ? [09:18] Correct. [09:19] sshfs is a way to mount SFTP as if it is part of the filesystem [09:19] ok [09:19] So programs that don't know SFTP can just read and write files normally [09:19] i'll start with nfs since i'm on the lan most of the time [09:25] what port does nfs run on [09:30] mm many [09:30] does ssh tunnel both tcp/udp? [09:31] SockPants: yes [09:31] You will not successfully tunnel NFS over ssh -w. [09:31] Not without great pain. [09:32] NFS not only uses multiple ports, it will by default negotiate additional ports during MOUNTP [09:32] And by default it refuses to countenance ports that can be bound to by unprivileged users (i.e. >= 1024). [09:37] hurrr but then how will it ever get past my firewall [09:37] i could open the default ports but it still couldnt negotiate any additional ones [09:38] SockPants: DO NOT use NFS over the WAN without using encryption (e.g. ssh, openvpn). [09:38] hence ssh [09:38] brb [09:38] If you need a network filesystem on the WAN, either use SSHFS or AFS. [09:39] NFS is only appropriate if you're inside LAN or an encrypted VPN [09:42] i might just try to set up a vpn then [09:42] what do I need to be able to use winbind in one computer to auth against a samba PDC in another computer? [09:43] though i've never used vpn so that should be interesting [09:43] but that would solve all my problems in one go i suppose [09:46] acalvo: first of all, make sure the samba side works [09:47] twb: it is working (wbinfo -u, wbinfo -g, wbinfo -a worked but no plaintext support) [09:47] acalvo: OK, I don't know what is next [09:47] I never got samba auth working right [10:41] New bug: #408947 in likewise-open5 (universe) "@reboot local crontabs do not run for likewise-open5 domain users." [Low,In progress] https://launchpad.net/bugs/408947 [10:44] Hmm [10:44] [ 392.560519] vcpu not ready for apic_round_robin [10:45] soren: How can I fix that? You're the KVM-guru, aren't you? :) [10:46] Jeeves_: Just ignore it? [10:46] soren: :) [10:46] Jeeves_: IIRC, it's just informational. [10:47] kvm migration makes the vm crash [10:47] I need more detail to form an opinion on that :) [10:47] which is kinda annoying [10:47] What detail do you want? :) [10:47] I'm currently testing convirt [10:47] You migrate a vm from one machine to another, and in the process, the vm crashes? [10:47] Oh. With convirt, you're on your own. [10:48] Why? [10:48] I have no clue about convirt. [10:48] Ok [10:48] and it doesn't use libvirt. [10:48] Well, I've never seen it work with libvirt either, so :) [10:48] ..so it's /completely/ unknown territory. [10:49] Well, that I /might/ be able to help you with, given enough detail. [10:49] just a sec, i'll install libvirt [11:01] Yo! [11:04] soren: Same thing happens :) [11:04] What do you need to know? [11:08] Jeeves_: Everything. [11:08] Jeeves_: What are you doing? Which version of ubuntu, libvirt, and kvm are you using? [11:08] Jeeves_: Where's your storage? [11:08] Jeeves_: What's the guest OS= [11:08] ? [11:09] Jeeves_: Define "crash". [11:09] soren: Right. [11:09] Running Ubuntu Jaunty on Sun Fire X2200's [11:10] Shared storage is nfs on a Ubuntu Jaunty server [11:10] Those are Intel boxes, right? [11:10] No, AMD [11:10] Ok. [11:10] Quad-Core AMD Opteron(tm) Processor 2376 [11:10] Cool. [11:10] Quad-Core AMD Opteron(tm) Processor 2354 [11:10] * soren prefers AMD [11:10] Both boxes have 8GB RAM [11:10] When migrating, the bridge interfaces become busy [11:11] So, migration is done [11:11] Right. [11:11] when it (seems to be) is finished [11:11] I can use the vnc-like interface to view [11:11] but the vm hangs [11:11] the kvm process is moved (or restarted) on the target host [11:11] and is using 100% cpu load [11:12] Fantastic. [11:12] Ok. [11:12] The NFS is mounted in the same place on both hosts? [11:12] (let's start with the basics) [11:12] it is [11:12] Ok. [11:12] 10.0.0.1:/export/vmstore /mnt/vmstore1 nfs rw,vers=3,rsize=8192,wsize=8192,namlen=255,hard,nointr,nolock,proto=tcp,timeo=4,retrans=9,sec=sys,mountaddr=10.0.0.1,mountvers=3,mountproto=tcp,addr=10.0.0.1 0 0 [11:13] Can you pastebin the contents of /var/log/libvirt/qemu/name_of_the_domain.log from both hosts? [11:14] http://pastebin.ubuntu.com/273366/ [11:15] http://pastebin.ubuntu.com/273367/ [11:19] Nothing jumps out at me. Erm.. [11:20] Can you start up the domain again and look for a few things for me? [11:20] sure [11:20] what should i look for? [11:24] /sys/devices/system/clocksource/clocksource0/current_clocksource [11:25] both tsc [11:34] Anything else? [11:39] Can you change that? [11:40] Try "echo acpi_pm | sudo tee /sys/devices/system/clocksource/clocksource0/current_clocksource" [11:43] Done [11:50] Doesn't help [11:52] Did it accept the change? [11:53] root@kvm4:/var/log/libvirt/qemu# cat /sys/devices/system/clocksource/clocksource0/current_clocksource [11:53] acpi_pm [11:53] Alright. [11:53] restarted the vm [11:53] deleted sound devices [11:56] that doesn't help either [12:18] New bug: #428650 in php5 (main) "PayPal WPP requires the PHP cURL library. (Currently using cURL Not found)" [Undecided,Confirmed] https://launchpad.net/bugs/428650 === georg is now known as kwork === georg is now known as kwork [12:58] soren: Switched to cifs storage, doesn't change a thing [12:59] Jeeves_: No, it wouldn't. [12:59] Jeeves_: I'm afraid I don't have a Jaunty rig to test this on. [13:00] soren: Need access? :) [13:00] (i'm serious though) [13:00] hi anyone here who can help me with the "minimal virtual machine" installation option in Ubuntu Server 9.04? [13:01] yonatan: Just ask your question. [13:01] We won't know unti lwe know what the question is. [13:04] towards the end of the installation you get a choice of what packages to install. I don't understand the option: "basic ubuntu server" should I select this option as a minimum? [13:04] Depends. What's the server for? [13:05] yonatan: It's kind of like asking "should I install a web server?". We have no way to know. [13:05] ok i'll elaborate [13:06] i'm looking to create a "management" virtual machine that will run Splunk, VMware vCLI and that's pretty much it [13:11] what is installed with the "basic ubuntu server"? is there a way of listing the packages? [13:14] yonatan: The safest option is to choose nothing [13:14] and apt-get install it later on [13:15] Jeeves_: Ok. Can you tell me when using apt-get is it possible to get a list of available packages? [13:15] morning [13:17] yonatan: yes. 'apt-cache search .' [13:17] But that's not what you want [13:17] you want to apt-cache search 'the stuff i need' [13:18] ok and can a similar command be used to see what is actually installed on my ubuntu server? [13:20] yonatan: dpkg -l | grep "^ii" [13:20] ok thanks Pici [13:21] Pici: That only shows packages that are installed, or have been installed [13:21] Jeeves_: Thats what he asked, isn't it? [13:21] yep I want to know what is installed so that I know if I need to get anything else [13:21] * soren takes a break [13:22] * Jeeves_ curses on kvm [13:22] might sound strange but I'm having problems installing something called VMware vCLI [13:22] and haven't found an answer anywhere [13:22] nothing concerning 'vmware' and 'problems' is strange [13:22] :) [13:24] well, perhaps the absense for problems. [13:24] ewook: Yes, that would be surprising indeed :) [13:24] henkjan: How's your redhat install on vmware going? :) [13:24] as suprising as if my grammar would be correct all the time. [13:24] ewook: :) [13:27] Jeeves_: hehe do you know how to get vCLI installed and working? [13:28] yonatan: No. And I don't want to know either === georg is now known as kinnaz [13:46] smoser: ping [13:55] soren: re workaround> ok [13:56] soren: do you think I should allow rw for the initrd? what about the kernel? [13:57] No. [13:57] Just r. [13:57] ...but you already do that, right? [13:57] I know I tested that. [13:57] jdstrand: ^ [13:59] soren: well, I'm looking at https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/431090/comments/4 [13:59] Launchpad bug 431090 in eucalyptus "libvirt apparmor profile is preventing libvirt from running eucalyptus VMs" [High,Incomplete] [14:00] jdstrand: Right. I thought that was just the console thing. [14:00] soren: Daniel says it is also the kernel and initrd [14:00] Oh... Perhaps I tested everything before we ould go into enforce mode? [14:00] soren: possibly... though -10 fixed all that [14:01] I *know* I tested with a domain that booted form an external kernel and ramdisk. [14:01] soren: anyhoo, this is just a workaround until I get virt-aa-helper to use the xml [14:01] jdstrand: Well, this was back when you first asked me to test. :) [14:01] ttx: do you have any patches you are sitting on for samba for karmic? [14:02] zul: no. [14:02] k [14:02] soren: I've added kernel, initrd, serial, console, attach-disk and detach-disk to the list of tests in https://wiki.ubuntu.com/SecurityTeam/Specifications/AppArmorLibvirtProfile. I'll get all that working [14:02] * soren hugs jdstrand [14:02] soren: I'll upload a eucalyptus specific workaround in the meantime [14:02] Thanks! [14:16] ttx, here. whats up? [14:17] smoser: bug 364916 [14:17] Launchpad bug 364916 in eucalyptus "When running Ubuntu as an Eucalyptus VM instance, module acpiphp is not available to be loaded, complicating the use EBS volume" [Undecided,Confirmed] https://launchpad.net/bugs/364916 [14:17] smoser: is there still something to do on that precise bug ? [14:17] ttx: http://pastebin.ubuntu.com/273481/ [14:18] zul: sounds good [14:18] ttx, comment 13 is correct there [14:19] the eucalyptus page does now mention how to get acpiphp into the ramdisk [14:19] smoser: a) is done and b) is todo but has another bug opened to track it ? [14:19] but we really should be producing a kernel and ramdisk alongside the images [14:19] so yeah [14:19] you're correct. [14:19] ok, will close this one and point to the other then. [14:31] New bug: #432481 in vm-builder (universe) "please provide a man page for ubuntu-vm-builder" [Wishlist,New] https://launchpad.net/bugs/432481 [14:32] I've got a Xeon quad-core, and when I run backup's and such a single gzip process seems to slow down my entire machine [14:33] When I look at resources the gzip process isn't taking much memory, but it still slows down my entire system make it unresponsive and pokey until it's finished. [14:33] Any ideas on how to make this better? [14:34] If I had a single-core processor I'd understand but it doesn't make sense to me when I have a quad-core [14:34] If you're running Intrepid or later, use iotop to see what's going on with io. [14:34] I'm running 9.04 [14:35] I'm installing it now. I'll start backup's again to see what happens with it [14:35] soren: does 'attach-disk' work for you at all in karmic? I added to /etc/libvirt/qemu.conf 'security_driver = "none"' to make sure apparmor wasn't getting in the way (see 'virsh capabilities' after libvirt restart to be sure) and then did: [14:35] virsh attach-disk testme /var/lib/libvirt/images/mini.iso hdc --driver file --type cdrom --mode readonly [14:35] error: internal error No device with bus 'ide' and target 'hdc' [14:36] soren: am I doing something wrong? is this a regression? bug #203020 indicates it is a long-standing problem [14:36] Launchpad bug 203020 in libvirt "virsh attach-disk doesn't work" [Undecided,Confirmed] https://launchpad.net/bugs/203020 [14:37] ScottK, anything specific I'm looking for. I can see the gzip/tar runnning with iotop [14:37] I guess see if it seems to be 'a lot' of IO. [14:38] The gzip is the top process and I see tar as a second process popup then remove then popup. [14:38] Hard for me to say what that would be for your system [14:38] Under the IO column it's always 0.00, there is of course alot of disk read/write activity [14:38] Perhaps do something you know is IO intensive, like copy a file, and see how it compares. [14:39] jdstrand: It worked in Jaunty. [14:39] jdstrand: I haven't really used it in Karmic, I guess. [14:39] * soren has to go pick up his daughter.. [14:43] Scott the system gets slow when I do a normal cp as well. [14:45] It's anoying. I know there is alot of disk IO from my one disk to the other, but that shouldn't slow down process that are in memory and already running. [14:45] Like firefox for example. Why would copying things slow down a web browser, can't figure that out. [14:46] ScottK, thanks for showing me iotop though, I like it alot better than top. [14:52] <_ruben> because firefox uses the disk for cache [14:53] BTW, since you mention firefox, it's also relatively clear that this isn't the right channel for the question in any case. [14:55] It's not just firefox, it's everything on the sytem. Doing server backup's shouldn't slow down all the other processes [14:55] I do VoIP on this box and any slow downs will mess up the phone calls. [14:56] you never run Xorg on server [14:56] I'll figure it out, was just wondering if it was something obvious. Haven't delt much with quad-cores so wasn't sure if it was something I needed to configure for it to better handle resources [14:56] kinnaz: for production that's what I do, this is a dev box at my house though. [14:57] thou dunno, if i copy files i doesnot affect my box performance so bad [14:57] that everything would lag [14:57] hey [14:58] kinnaz: that's what's tripping me. I've never seen disk i/o effect running processes like this. [14:58] still having trouble with local DNS. anybody willing to help me diagnose the issue? [14:58] clusty: descriibe the problem, somebody will help if they can [14:59] after doing dist-upgrade all my image bundling problems were solved with eucalyptus-1.6, but when I try to run an instance I get the following: http://paste.ubuntu.com/273505/ [14:59] so: i am having a linux box doing bind+dhcp. I wrote a zone ffile to the best of my ability [15:00] thing is I can resolve form any computer the name of the dns server, but pc's can't resolve each other, nor can the server resolve computres [15:00] but after dist-upgrade, walrus and sc both work well [15:01] i am stuck for a week now. don't know how to proceed. I think the only way is for somebody to tak a peak at my conf files [15:02] ahh I see bug https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/430093 addresses this [15:02] Launchpad bug 430093 in eucalyptus "Eucalyptus "403 Forbidden" when trying to run instance" [High,Triaged] [15:04] clusty: Sorry, I can't help with that one, it's been awhile sense I've done those. [15:05] * kpettit installing more memory [15:07] New bug: #368718 in dbconfig-common (universe) "package dbconfig-common 1.8.40 failed to install/upgrade: o subproceso post-installation script devolveu o estado de saĆ­da de erro 1" [Undecided,Incomplete] https://launchpad.net/bugs/368718 [15:14] on my new install of karmic alpha 6 I am using XFS and was amazed how fast the filesystems were created, and it was nice to see that there wasn't going to be any fsck happening at boot [15:18] aubre, i thought xfs had a few showstoppers, like FS cannot be shrunk [15:18] this makes it really annoying for a desktop installation [15:27] well I am not planning to shrink my server file systems === SyL is now known as Guest44374 [15:27] clusty: so I don't think it will bother me too much [15:28] aubre: what I'd really like is for Oracle/Sun to give ZFS a linux-compatible license [15:31] aubre, i loved ZFS [15:31] dunno how fast it is, but is insanely easy to admin [15:32] New bug: #341003 in php5 (main) "php5 crashed with SIGSEGV in start_thread()" [Low,Incomplete] https://launchpad.net/bugs/341003 [15:34] ttx: we said yes to bug #403599 right? [15:34] Launchpad bug 403599 in php5 "Installing PHP5 performs a reload to apache instead of a restart" [Low,Confirmed] https://launchpad.net/bugs/403599 [15:35] zul: looked alright to me, and you confirmed it. Some validation testing is in order, I guess [15:38] WRT 341003, isn't a php crash a security feature? [15:39] hmmm? [15:39] are the zone files in /var/cache/bind automatically generated? [15:39] it's a windows look-a-like feature. [15:54] can i build a software raid10 set ? [15:57] hello [15:58] im root on my server but i cant remember mysql login [15:58] just google for a mysql recovery [15:58] you can easily do it [15:58] i have removed mysql and installed it but it does not ask for the password [15:58] apt-get --purge remove mysql [15:59] or you could try dpkg-reconfigure mysql-server-5.0 iirc [15:59] mysqld_safe --skip-grant-tables [16:00] Then you can login as root with no password and manually change the user information in mysql.users [16:03] i got dynamic dns to work for all the windows machines on the LAN, but linux is still resilient :( [16:04] i configured send host-name "blahh"; [16:04] on all dhcp linux clients [16:04] what else can i do [16:04] grrr [16:11] deleted the /var on a production server...restored the /var/www...and /var/logs...any reason to rebuild the server..or can i patch this up? [16:12] teddymills, depends what you had there [16:12] like /var/lib is where postgres has the database files [16:20] ttx: re bug 348990 [16:20] Launchpad bug 348990 in postfix "Deinstallation doesn't delete all files" [Low,Fix released] https://launchpad.net/bugs/348990 [16:20] ttx: did you check with lamont if it makes sense? [16:20] mathiaz: yes [16:21] mathiaz: no. I checked that it was needed for proper purging though [16:22] ttx: hm right. [16:22] ttx: I had a discussion about what should be done on purge for daemons [16:22] mathiaz: ah ? [16:23] ttx: with slangasek when I was reviewing stomp-server [16:23] ttx: deleting a user on purge wasn't suggested [16:23] ttx: just to make sure that the system uid is not reused [16:23] mathiaz: that's not what the patch does though. [16:23] ttx: agreed [16:23] ttx: I just so it in the patch [16:23] ttx: as far as /var/lib/postfix [16:24] ttx: the suggestion was that if there were files in /var/lib/package that weren't reproducable/recoverable it should never be deleted by the maintainer scripts [16:25] ttx: however it's a grey area [16:25] ttx: policy doesn't say anything about it [16:25] not sure the master.lock is really unrecoverable [16:25] ttx: right - master.lock should be removed [16:25] but it for sure prevents reinstall of a purged postfix [16:25] ttx: may be not with a rm -rf /var/lib/postfix [16:26] agreed. [16:26] ttx: as this could potentially deleted other things [16:26] mathiaz: could you bring this up to lamont ? I'm nearing the end of the day and I still have plenty of things to finalize. [16:27] ttx: sure - not urgent though [16:29] mathiaz: I plan to spend some time on basic first-user-experience / usability Eucalyptus bugs next week. Do you think you will have time to participate in that effort ? [16:29] ttx: I'll be in portland for ldapcon and linuxcon next week [16:29] arg. [16:29] ttx: so I doubt it. [16:29] ok. [16:38] ttx: can you have a eucalyptus cluster with different arches ie one being i386 and the other amd64? [16:38] i hope a RAID10 mdadm setup will improve disk performance for my vm's [16:39] zul: You mean nodes with different arches ? I don't know. nurmi or soren would know for sure. [16:39] ttx: if yes then I could probably help [16:40] zul: ok, noted. [16:42] ttx: doesn't prevent reinstall of a purged postfix [16:42] mathiaz: ^^ [17:01] pfff gave up setting dynamic DNS for localhost [17:02] wrote a zone file and that's that :( [17:03] i should have used the server preseed file for my pxe install [17:03] urg this server is going to need some apt-get --purge loving [17:10] if i try installing gcc on canonical's hardy image on ec2, it says i need to remove libc6-xen. what gives? [17:26] New bug: #432581 in libvirt (main) "[karmic] libvirt/apparmor breaks non-default serial, console, kernel and initrd" [High,In progress] https://launchpad.net/bugs/432581 [17:35] hi how do I check that the following packages are installed on my system: binutils, glibc or libc6, libxml2, perl-doc, liburi-perl ? [17:36] yonatan: dpkg -l |grep package [17:37] if it has an ii on the left, it's installed [17:37] cool thanks will try now [17:37] no problem [17:38] dpkg -l | grep binutils doesn't return anything [17:39] can you try something that you know is installed? [17:39] What's being used these days, for a distributed, replicated, elastic file system? [17:39] i just installed a clean system but added wget at the end [17:39] also you can just run "apt-get install binutils" and it will try to install it and if it's already installed it will say that [17:40] it returns ii wget [17:40] ok, then they aren't installed [17:40] ah [17:40] Looking into Eucalyptus, but it doesn't seem to have anything that does distribution, replication, and elastic growth. [17:41] yonatan: "apt-cache search package" will show you what could be installed by apt-get. [17:41] blistov: what do you mean by elastic file system? [17:41] crap, lunch, I'll be back later [17:42] ok thanks SyL [17:43] SyL: IE: I have 5TB of data between 2 physical units. Data is replicated and striped. I need to increase capacity to 10TB by adding a few more physical units. [17:43] I don't want to have to build new partitions all over. === JoshL is now known as JoshL[Away] [17:59] I'm really looking for a p2p fs, providing aggregate performance. [18:01] how can i get the kernel config for canonical's ec2 hardy image? [18:04] bnjmn: I would not recommend Canonical's EC2 Hardy image at this point. It was last published in April and has a number of serious bugs filed against it. Most of these bugs have been fixed in revision control, but a new image may not be published for a while yet. [18:18] erichammond: I read your comments about XFS after I had used it on my Eucalyptus proof of concept system and I have to say I am so far impressed with it. [18:22] i'm getting the following error message ""hiding the command line arguments : symbol lookup error: /usr/lib/perl/5.10/auto/Crypt/SSLeay/SSLeay.so: undefined symbol: Perl_Tstack_sp_ptr"" === JoshL[Away] is now known as JoshL [18:47] aubre: I'm not a filesystems expert, but I started using XFS on EC2 and it's worked well for me. [19:12] New bug: #431063 in image-store-proxy (main) "Enable tests during build" [High,In progress] https://launchpad.net/bugs/431063 === centaur5_ is now known as centaur5 [19:45] seeing a strange NIC issue. I have a server with 2 nic cards, each on a different subnet (10.1.10.xx and 10.1.11.xx). I start a ping to both interfaces, and disable the switch port that the .10 interface is going to, but it continues to ping. Then I disable the switch port that the .11 interface goes to, and they both stop pinging at the same time [19:47] the /etc/network/interfaces is setup correctly, giving each server its own ip, and the mac addresses on the cards are different in "ifconfig" [19:49] when netboot installing, what does 'basic server' install? === mushroomtwo is now known as mushroomblue [19:57] New bug: #432659 in mysql-dfsg-5.0 (main) "package mysql-server-5.0 5.1.30really5.0.75-0ubuntu10.2 failed to install/upgrade: podproces post-installation script zwr?ci? kod b??du 1" [Undecided,New] https://launchpad.net/bugs/432659 === freenode is now known as Guest35677 [20:09] Hey guys [20:09] How can I see the IO taking place so I can identify what is locking up my server [20:09] Keizer: Install and run iotop? [20:24] I'd like to see all the mail leaving my server. Is there an outgoing maillog? [20:25] /var/log/mail.log? [20:25] And that's it? [20:32] stupid fake RAID controllers [20:32] incorrect: [20:33] damn, i wish i hadn't old the installer to erase data from a partition [20:42] hey guys ... where should we report bugs against the UEC karmic image? [20:43] not about a specific package, but the AMI itself? [20:44] I have a server with two nic cards, and both have their own ip. But both are pushing traffic through the same cable. Anyone know how thats possible? [20:45] Hypnoz: it isn't, how do you know they are pushing through both IPs? [20:46] if i disable the switch port for one of the nics, it keeps pinging [20:46] and it's running at 100mbit instead of 1gig, same speed as the other port [20:46] and if i disable the switch port for the 100mbit connection, both stop pinging at the same time [20:47] then I would imagine that you have 2 IPs on 1 NIC card. [20:47] /sbin/ifconfig will show you. eth0 and eth1 will have tx and rx [20:47] smoser: ^^^ [20:47] eth0 Link encap:Ethernet HWaddr 00:21:5e:4d:c7:41 [20:47] inet addr:10.1.11.195 Bcast:10.1.11.255 Mask:255.255.255.0 [20:47] UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 [20:47] RX packets:668438 errors:0 dropped:0 overruns:0 frame:0 [20:47] TX packets:389284 errors:0 dropped:0 overruns:0 carrier:0 [20:48] collisions:0 txqueuelen:1000 [20:48] RX bytes:61820665 (58.9 MB) TX bytes:50181250 (47.8 MB) [20:48] Interrupt:16 [20:48] eth1 Link encap:Ethernet HWaddr 00:21:5e:4d:c7:42 [20:48] inet addr:10.1.10.195 Bcast:10.1.10.255 Mask:255.255.255.0 [20:48] UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 [20:48] RX packets:1694179 errors:0 dropped:0 overruns:0 frame:0 [20:48] TX packets:983370 errors:0 dropped:0 overruns:0 carrier:0 [20:48] collisions:0 txqueuelen:1000 [20:48] RX bytes:2178889683 (2.0 GB) TX bytes:102839047 (98.0 MB) [20:48] Interrupt:18 [20:48] shit [20:48] sorry [20:48] http://pastebin.com/d403b2ac9 [20:48] thought the link was in my clipboard [20:48] EtienneG, https://bugs.launchpad.net/ubuntu/+filebug?field.tags=uec-images [20:49] both interfaces look good right. both have their own IP and mac [20:50] i have 4 disks, i've set up a swap partition on each, a /boot RAID 1(x4) same for / and /var and a /srv as a raid10 [20:50] AAAAaaaargh! filebug redirection [20:51] Hypnoz: can you do "route -n" for me and put it in a pastebin? [20:51] http://pastebin.com/d6dec17b5 [20:52] have a good weekend [20:52] the 10.1.5.0 to 10.1.11.190 is an openvpn static route [20:52] EtienneG: https://bugs.edge.launchpad.net/ubuntu/+filebug?no-redirect&field.tags=uec-images [20:53] mathiaz, that is so much better [20:54] SyL, there is one weird thing, I have my default gateway set for the 11 interface in /etc/network/interfaces [20:55] but still showing 0.0.0.0 in route -n [20:55] err nevermind, its the default one at the bottom [20:56] Hypnoz: I don't know the full details, but I'm guessing the traffic is going out the vpn. [20:56] there are 2 ways you can test what's going on. unplug cables one at a time, or shut down the interfaces one at a time. [20:57] this expert installer doesn't seem too stable [21:28] erichammond, which Ubuntu ec2 image would you use at this point? [21:30] All of my production systems run on the Ubuntu 8.04 Hardy AMIs I publish on http://alestic.com (32- and 62-bit) though I sometimes use the 9.04 Jaunty as well. As far as I know, these are probably the most popular two Ubuntu images out there at this time. [21:31] I'm supporting the Canonical AMI development process with the goal of having those AMIs become the public standards. [21:32] cool. great, I will check those out. how might I also get the kernel source & config used on those so i can build my own kernel module? [21:38] so, like mdadm RAID10 is a bad idea [21:46] why [21:50] erichammond, nevermind, i see your AMI is using one of amazon's kernels, so I know where to get the stuff i need. thanks [21:51] bnjmn: http://alestic.com/2008/05/kernel-modules-2-6-21 [21:53] bnjmn: Yes, and you should be aware that using Amazon's kernel makes it not not completely "Ubuntu". It's just the most reliable way I've found to run Ubuntu on EC2 at this time. I believe better times are coming :) [21:54] bnjmn: If you need to use XFS on 2.6.21, you'll need to pick up a special xfs.ko built by Amazon. Ping me later if you can't find this. [22:01] erichammond, actually, i needed a newer squashfs of all things. i got it worked out now. [22:01] thanks [22:04] how do I make my ubunut server a print server ? i like to take print requests and send it over to the printer which is on different network [22:05] axisys: https://help.ubuntu.com/9.04/serverguide/C/cups.html [22:05] axisys: this is for jaunty though. [22:06] mathiaz: that is what i am running [22:07] it does not explain how i can take a request from another server and send it over to a printer which is on a different network.. [22:07] mathiaz: does it? may be i should read it again? [22:32] how do I send a print request from client 192.168.0.21 to a cupsd running on a remote machine 192.168.0.98 ? [22:33] right, so RAID10 is a no go [22:34] so what kernel should i use to run a kvm server? [22:44] am I correct in thinking that ec2-init in jaunty do not work when the instance is run on Eucalyptus? [22:44] looks like this is what i was looking for http://ubuntuforums.org/showpost.php?p=511007&postcount=3 === erichammond1 is now known as erichammond [23:01] EtienneG, i dont know why hat would be [23:01] smoser, I am filing a bug [23:01] hold on [23:04] smoser, if you care, it is bug #432773 [23:04] Launchpad bug 432773 in ec2-init "ec2-fetch-credentials do not work in jaunty instance being run on Eucalyptus" [Undecided,New] https://launchpad.net/bugs/432773 [23:06] is there a way to connect to a remote server, like with SSH, but be able to see/use the desktop on that server? [23:07] swoody: FreeNX [23:07] swoody: VNC can tunnel through SSH [23:07] swoody: RDP can be used, as well. [23:07] ok, I'm going to point someone in here, if you guys can help him out with that? :) [23:07] hah. [23:08] sadly, my tech support shift is over now. [23:08] haha [23:10] ok was told by swoody to come ask for help in here [23:10] :) [23:10] I have a remote server with ubuntu desktop 9.04 and am wondering if i can access the gui on it via remote [23:11] https://help.ubuntu.com/community/FreeNX [23:11] read that. [23:11] that's by far the easiest solution. [23:13] k will try ty ill probaly beback :/ [23:13] kernel does not matter for kvm? [23:33] I just installed server 9.04 but I can't get the static network configuration to work [23:33] I edited /etc/network/interfaces as the documentation says, but "/etc/init.d/networking restart" and reboots do not bring the interface up [23:33] However, "ifup eth0" works [23:34] do you have "auto lo eth0" in the /etc/network/interfaces? [23:35] what filesystem to use for a large number of small files, like /usr/src, lots of small source files and headers [23:36] blak111 - danke :) [23:36] I've found reiserfs to be good with large numbers of small files [23:36] <4kb [23:37] im having problomes installing nx does someone have time to go to a msg box and help me walk thru the steps i am new to ubuntu [23:37] linux is pretty good at storing a lot of metadata in cache so if you're doing repeated actions you'll see decent lookup performance [23:47] I want to be able to access my computer via ssh over internet, but I am worried about security. a WOL udp port 9 and 777-> 22 for ssh are the only ports I have open on my router. any suggestions? [23:48] bdmurray: hey - any news on running my bug scripts on qa.ubuntu.com? [23:49] mathiaz: hey - yes I'm actually work on that right now! ;-) [23:49] bdmurray: \o/ [23:49] bdmurray: let me know if you have any question [23:50] mathiaz: line 86 in acceptedbugs.py should be Distro Series I think [23:51] bdmurray: http://people.canonical.com/~mathiaz/buglists/acceptedbugs.ubuntu-server.latest.html [23:51] bdmurray: ^^ it's just one serie [23:52] bdmurray: or may be I'm confused by the spelling [23:52] mathiaz: okay, how about we call it Release ? [23:52] bdmurray: sure