[00:07] aaand I broke the Ubuntu install on it [00:07] *fixes* === bastidrazor is now known as bastid_raZor [01:59] anyone alive in here today? [02:04] chrislabeard, depends [02:04] I need help with figuring out how to let the server write to directories [02:05] I'm pretty sure we answer questions in here, not statements [02:05] alright. [02:06] I'm using EHCP I create a domain add files and the server does not have write access to the directory [02:06] how can I fix this ? [02:07] ehcp? [02:07] easy hosting control panel [02:09] its just a web panel [02:10] dunno anything about it [02:10] check your directory permissions [02:10] there should be isntallation documentation for that [02:10] if not, that sucks [02:12] lol k [02:13] sorry, I'm just not in the mood to learn about ehcp to really help you [02:19] Right i gotcha [02:25] man, rsync is really slow, even with copying from one drive to another... [02:25] in the same machine [02:40] anyone have vsftp running on their system? I'm getting this after trying to init.d restart [02:40] Rather than invoking init scripts through /etc/init.d, use the service(8) [02:40] utility, e.g. service vsftpd restart [02:40] Since the script you are attempting to invoke has been converted to an [02:40] Upstart job, you may also use the restart(8) utility, e.g. restart vsftpd [02:40] vsftpd start/running, process 12030 [02:40] ya, just do what it says :) [02:43] im not familiar with the restart utility. I tried typing restart vsftpd and that didn't work either. [02:43] restart utility? [02:43] I'm pretty sure it says [02:43] service vsftpd restart [02:43] yes thats what I typed., here is the response: [02:44] restart: Unknown instance: [02:44] service vsftpd start [02:44] maybe it wasn't running, so restart failed [02:44] tried that as well... [02:44] strange, I have no issue with it on my system [02:45] also tried with sudo [02:45] oh, you have to use sudo [02:47] i had vsftpd running the other day. Reformatted my system and started from scratch. I didn't have this issue before. === unreal_ is now known as unreal === bastidrazor is now known as bastid_raZor [04:07] curious how bad is the TLER issue with raid cards these days [04:07] any experiences ? [04:41] sommer: Bug #614897 [04:41] Launchpad bug 614897 in ubuntu-docs "bind9 instructions should use relative paths for files" [Undecided,New] https://launchpad.net/bugs/614897 [04:49] Apache problem on my Ubuntu 9.10 server: I'm trying to get a subdomain to point to the right DocumentRoot, but it isn't working. Here's the vhost file in question: http://pastebin.com/S4tkm9F4 What am I missing? [04:51] monit stopped monitoring one of my services and I fixed the problem. how do I let monit start remonitoring the services ? [04:52] Tohuw, there are no directories defined in that virtualhost [04:52] qman__: oh, defining the dir is required? :( [04:53] na, it's magic :) [04:53] well, i just thought it could infer it from the docroot [04:53] or something [04:54] that would be, simple_vhost, not normal vhost [04:54] ok obviously not, adding directory and testing [04:54] oh [04:56] when i define the directory, do i start from the docroot defined by that vhost, or the docroot in my apache2.conf? [04:56] directory statements use absolute paths, last I knew [04:57] at least if you start them with a / [04:57] oh [05:00] I changed the vhost file to http://pastebin.com/hhnVGksC, but still no dice :\ [05:00] the other problem I see is that the second domain name falls under the first domain name [05:01] what do you mean? [05:01] just that projects.constellationmedia.com is a subdomain of constellationmedia.com? Am I going about a subdomain setup wrongly? [05:02] I'm looking it up to confirm, but that's what I think the problem is [05:04] Thank you, I appreciate your help [05:07] how do I change a users home directory and make sure they only have access to that directory and what is below it? Im not quite following the chroot topic. [05:10] Tohuw, well, I can't find a configuration set up similarly to yours to confirm, what incorrect behavior is it showing? [05:10] if it's rendering the main site instead of the subdomain, or not rendering the subdomain at all, this is probably the case [05:11] qman__: if you go to http://projects.constellationmedia.com, you'll see it goes to the same page as http://constellationmedia.com. It should be rendering http://constellationmedia.com/projects [05:12] Tohuw, try commenting line 4 [05:13] oh, that worked [05:13] makes sense [05:13] now the question is: how do i allow www.constellationmedia.com to point to the parent domain, since the wildcard screws things up? [05:14] serveralias www.constellationmedia.com [05:14] fair enough [05:14] though to avoid problems with cookies or SSL, I'd suggest doing a redirect or mod_rewrite instead [05:14] yeah, makes sense [05:15] i'll make a .htaccess with the appropriate directives [05:30] I have a rails app located at /home/cmg/redmine/1.0-stable and symlinked to /home/cmg/public_html/projects. It's not working. http://projects.constellationmedia.com/ isn't working -- I just get a directory index. Neither is http://constellationmedia.com/projects (the same place) My vhost file: http://pastebin.com/U1K4V0d0 (using Passenger) [05:31] I don't know anything about rails, but I'd double check to make sure it's enabled in apache [05:31] and check the file association too [05:32] It is. It was working before I moved it from another directory. I updated the .yml files and environment.rb [05:32] ok [05:32] oh [05:32] wait [05:32] d'oh [05:33] forgot to chown the application directory [05:33] ah, that'll do it [05:34] actually, apparently it won't :( [05:34] hmm [05:35] I don't remember if FollowSymlinks is on by default [05:35] but if your directory listing is of the correct directory, I guess it's on [05:35] I have it on http://pastebin.com/U1K4V0d0 [05:35] yeah [05:36] why is the server iso around 600mb ? [05:37] deco, to contain as many popular packages as will fit on a CD [05:38] qman__: thanks [05:38] deco: just because the package is on the cd doesn't mean it's installed by default [05:38] Tohuw: ah ok [05:39] been wanting to setup a vps , why would somone use ubuntu-server instead of debian ? [05:39] i know there's a reason for the desktop [05:39] version of ubuntu [05:39] my main reason is tasksel [05:39] but i don't know why for the server [05:39] makes a lot of things really easy [05:39] qman__: like what for example [05:39] deco: for the same reason you would prefer ubuntu for your desktop over debian [05:39] LAMP [05:40] single checkbox for a working system [05:40] hmmm [05:40] could you do it on debian? sure [05:40] but ubuntu has a very quick and simple way to do it [05:41] ubuntu server, like its desktop counterpart, does a great job at vastly simplifying and streamlining deployment and operation without sacrificing power and configurability (or at least, not too much) [05:41] Tohuw: okay thanks [05:41] hence the old joke, "ubuntu is an African word meaning 'I can't configure Debian'." [05:42] haha [05:42] when you need a web server in ten minutes flat, ubuntu is the answer [05:42] same with the desktop version :p [05:43] i like that [05:43] oh well i should really just download the iso and test it in vbox [05:44] the only difference between the two is the default package set [05:44] I think they even use the same kernel now [05:44] desktop and server, that is [05:44] oh, and the installer [05:44] qman__: installer is ncurses right ? [05:45] server uses a menu instead of a graphical environment [05:45] yeah [05:47] qman__: as of 10.10 beta, the kernel is still -server as opposed to -generic, the default in Ubuntu Desktop. [05:47] Though the versions are the same, at least currently. (2.6.35-14) [05:47] guess they decided not to merge the two, then [05:48] it was in the plan at one point or another [05:48] even so, it's just some minor changes [05:48] what's the advantage to unifying them? [05:49] one less kernel to bug test, patch, and maintain [05:49] fair enough. I don't know enough about the tunings between the two to comment on it, was just curious [05:49] they're not very different [05:50] but the fact that they're not the same one means extra work [05:51] one thing that does bother me is that they call the 32-bit version i386, when it requires an i686 CPU [05:51] guess it doesn't matter for most people [05:52] just me and the other two guys still running an i586 ;) [05:54] Tandy called, they have a new model to sell you. It's ready for pick-up at your local Radio Shak [05:55] harsh ;) [05:55] I would replace it, but every time I go to do it, one of the newer machines fails [05:55] it just won't die, can't argue with that [05:55] qman__: I know one of my machines has an AMD K6-2+, which was i586 in terms of instruction set. Upgraded from a Cyrix MII that was i686. [05:55] I have to go. [05:56] yeah, this is a K6 [06:00] qman__: fixed my problem! it was a simple matter of moving the RailsBaseURI directive to the parent vhost entry (the first one in the file). Why that worked, or rather why having it in the vhost entry for projects.constellationmedia.com didn't work, I don't know [06:01] Rails is very black box to me sometimes. [06:01] hah [06:01] yeah, that doesn't make a lot of sense [06:01] it's working, i'll just smile and not [06:01] s/not/nod [06:02] I wonder what clever trick I can use to get http://projects.constellationmedia.com to work (http://constellationmedia.com/projects works) [06:03] maybe put it in both? [06:07] nope, that was too easy to work [06:07] this is rails: it has to be difficult and obscure to deploy. get real, man! [06:08] related: I delight in antagonizing rails fanatics, even though I don't have (many) specific complaints or even useful knowledge of rails itself [06:10] heh [06:10] I've never used it myself, I learned PHP years ago and stuck to that [06:11] though python has caught my interest, if I ever get some time [06:57] how do I copy the partition map over from one drive to another? [07:15] WalterN: you can use `sfdisk` for that. === Darkomen_ is now known as Darkomen [07:51] Im running ubuntu 10.4, but cant seem to use last or lastlog commands: last: /var/log/wtmp: No such file or directory [07:52] do i just need to create the file and it will start working? i doubt it as i would think it would automatically be created if it wasnt [09:04] I'm having an odd issue with my Ubuntu Server running Apache 2.2.11: I have several vhosts files, one of which contains , the rest contain VirtualHost directives that are domain-name specific and include ServerName entries. So if I go to something that hits the server but not any of the domains specified in the other vhost files, why don't I get the documentroot for the default vhost? It seems to go through the other [09:09] Ah, now I get it: http://www.onlamp.com/pub/a/apache/2004/01/08/apacheckbk.html [09:35] does anyone else ever have a problem where an ssh suddenly becomes very very slow, even though nothing significant has changed in terms of load on the localhost/ssh target/available bandwidth on either end? [09:36] New bug: #614939 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.41-3ubuntu12.3 failed to install/upgrade: subprocess new pre-installation script returned error exit status 1 stop: Job has already been stoped: mysql" [Undecided,New] https://launchpad.net/bugs/614939 [09:36] an ssh session, I should have said [09:37] Tohuw: I've experienced it. I never really thought much about it. [09:37] FireCrotch: it drives me insane... and I can't determine why it might be happening [09:38] Tohuw: It usually goes back to normal after a few seconds for me. I usually assume that its due to a load spike on the local machine or the remote box [09:39] FireCrotch: yeah... but the odd thing is, when it happens for me, I don't observe a load spike on either end... I was guessing some flakiness in connection on my end (not that I've observed any directly) [09:40] I assume this is over the internet, not the local network [09:40] yes [09:40] Could definitely be internet lag. The internet is a complicated beast, sometimes it just... slows down for no reason [09:44] Tohuw: thanks for the help earlier with SASL stuff might've been last night actually I finally got it all working [09:44] robertpayne: excellent :) [09:44] Tohuw: turns out using postfix + dovecot you don't need sasl2-bin at all or any of it's configured stuff.. dovecot provides it's own implementation [09:45] robertpayne: really? Interesting. I'll make a note of that... I'm setting up dovecot on a server soon (I'm used to Courier). [09:46] Tohuw: yeah. I'm trying to make a shell script now to install the base server setup ( lamp, phpmyadmin, postfix, dovecot, postfixmyadmin, ssl certs ) etc.. having trouble getting sed to replace some variables in the config files though [09:46] sed script seems to expand variables only when not inside a recursive loop like this: http://pastebin.com/UdFk2ZwP [09:46] any opinions regarding ebox vs pfSense? [09:46] robertpayne: if you're scripting that large of a deployment, you might want to consider using a "golden image" type approach instead [09:47] Tohuw: It's for my webhost's VPS I don't think I can go that route [09:47] understood [09:48] I wish though would make life a ton easier :( [09:49] Tohuw: actually you have a link with info about a golden image? I might look into it I might be able to hook it up [09:50] robertpayne: maybe cfengine, puppet or chef might be worth a look for you [09:51] robertpayne: golden image just means that you have a master image (which you maintain) and can just deploy when you need a new vps [09:52] joschi: ahh I figured as much yea heh no I have to reinstall the whole thing [09:52] joschi: they offer snapshots but you can't move them across different servers ( setting up a dev box now then will migrate to production box ) [09:55] the shell script isn't too heavy [09:55] robertpayne: that was one of my complaints about my former vps setup. I went with a cloud-based solution from the same data center, and now i can make my own images, roll them out as I please in minutes, clone them, roll out premade (by the data center) images, etc. Not to mention I can dynamically allocate resources and only pay for what I need in most regards. [09:56] Tohuw: very interesting.. what hosting solution are you using? [09:56] robertpayne: you might consider doing it more "the unix way": create several small scripts to configure the different parts. This allows for scaling and interchanging parts. You could still have a master script that includes which scripts you want to run [09:56] robertpayne: www.stormondemand.com [09:57] Tohuw: yea true.. I doubt I'll ever really have to use these that often though.. mostly just once on the dev/production box and if they ever completely fail I'd have to do it again [09:58] fair enough. sometimes, quick and dirty is completely called for [09:58] Tohuw: this sed problem is the only one still remaining.. I can't understand why it's not expand the env variable unless it's used standalone [09:59] robertpayne: I'm not sure why either... try asking #bash [10:07] robertpayne: example? [10:07] http://pastebin.com/UdFk2ZwP [10:08] the sed command replacement is always resulting as empty string.. tried several different methods [10:08] works if I don't use a env var inside the replacement [10:19] robertpayne: you're missing -i for in place editing of the files [10:19] joschi: oh right now I'm just testing.. printing it to stdout [10:19] anyways turns out on my local machine I'm running /bin/bash [10:19] server is using /bin/dash [10:20] $HOSTNAME isn't set as env for /bin/dash [10:21] robertpayne: also `set` is IMHO wrong is this context [10:21] joschi: yes that is wrong too I've tried that many different ways [10:21] robertpayne: just write USE_HOSTNAME="$HOSTNAME" [10:21] yeah [10:21] I'll have to create a second script and run it as /bin/bash [10:21] to setup the files [10:21] not sure how the rest of the script will run under bash instead of dash [10:22] just use `hostname` instead of $HOSTNAME [10:22] for dash? [10:22] er bash [10:23] both [10:23] here is what I don't get [10:23] bash supports $HOSTNAME, but not every shell does [10:23] actually.. wait a sec [10:23] maybe I do need to run it as bash.. because command line ubuntu echo $HOSTNAME works but /bin/sh is dash and doesn't so.. [10:26] yeah I just have to run it as bash.. argggg! thx [10:28] even after rebooting, i still cant run apt-get. im getting: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable) [10:29] and no, i dont have apt running, etc [10:49] MACscr, try sudo lsof /var/lib/dpkg/lock or sudo sudo lsof /var/lib/dpkg/status [10:50] blue-frog: thanks, i got it fixed. Removed the lock file. No idea why it didnt work the first time, but it did after reboot and i rm'ed it again [10:51] MACscr, there's a process looking for updates. certainly this one. [10:51] update-manager apt-get and so on [10:51] if you want to be sure use lsof. it will tell you what is using it [10:52] so give it time to finish what it has to do and then issue your commands [11:15] any LTSP gurus around? [12:54] are there only UUID's for partitions or can you get some unique identifier of the drive itself? [12:55] in other words, is there any information i can get from /dev/hda to differentiate it uniquely from /dev/hdb given that they are identical drive models? [13:01] New bug: #614981 in groovy (main) "Please sync groovy 1.7.4-1 (main) from Debian unstable (main)." [Wishlist,New] https://launchpad.net/bugs/614981 [13:05] CppIsWeird: There might be.. What do you need it for? [13:06] well i was running along the street with a big red bag full of western digital hard drives stacked in a specific order from 1 to 1 million. All the sudden a big gust of wind blew me over and the hard drives fell all over the place. [13:07] I hate that. [13:07] so now i need to put them back in the right order. [13:07] Why? [13:07] god told me i had to or the world would come to an end. [13:07] you have to help me save the world. [13:07] Why do they need to be in the right order. [13:07] *shrug* [13:09] Maybe today is the day where I stop bothering to respond to anything in this channel. [13:09] 5 years seems like a pretty good run. [13:10] * soren makes a rude gesture and wanders off [13:23] lol, wow, someone needs to get out of their cubicle more often. [13:36] CppIsWeird: You might also think about the fact that people in here have things to do, and that they're here to help people as well. [13:36] You trying to be funny isn't very helpfull (nor funny) [13:37] damn i hate grub2 [13:38] * Jeeves_ too :) [13:38] i have software raid on a dual boot machine, update-grub seems to generate a valid grub.cfg but i only see the win7 option when i reboot [13:40] so what other options are there [13:40] in terms of bootloaders [13:43] incorrect: There's also Grub. [13:43] i didn't see it in 10.04 [13:43] But i'd try to find out what is wrong with grub2 [13:44] i think its something to do with the software raid [13:44] Is grub being installed on both your disks? [13:44] well i get a grub2 menu [13:44] but it only has win7 as an options [13:44] so i pxe boot rescue, chroot in [13:44] do an update-grub [13:44] it moans about a memory leak [13:45] however the /boot/grub/grub.cfg seems to have everything in i would want [13:49] Is grub being installed on both your disks? [13:49] well i did manually force to do an install [13:49] So grub is probably on only one of your disks [13:49] does grub2 not just read the .cfg file at boot like grub2 [13:49] err grub1 [13:50] what happens if you tell the bios to boot from the other disk? [13:50] can't do that without stuffing the software raid [13:50] then all the UUID's get stuffed [13:50] huh? [13:50] uuid's don't get stuffed based on settings in your bios [13:50] the nvidia software raid is a bit weird [13:50] that's the whole idea with uuid's [13:50] they just did [13:51] Ah, it's not software raid [13:51] it's hostraid [13:51] ah [13:51] which sucks, usually [13:51] And usually creates more issues than it solves :) [13:52] well i tried disabling it and installing but that caused more problems [13:52] Why? [13:55] are there only UUID's for partitions or can you get some unique identifier of the drive itself? is there any information i can get from /dev/hda to differentiate it uniquely from /dev/hdb given that they are identical drive models? [13:56] got it, it needed a manual on both drives [13:59] oh thats is awesome dmraid didn't start so the /dev/mapper devices it wants to use aren't there at boot [13:59] sigh [13:59] hmm what to do [14:05] fine reinstall using mdadm [14:05] but i bet it won't find the windows install [14:14] are self-signed certificates any less secure than using one signed by a CA? [14:14] robertpayne: no, but they can't be verified - security is the same if you trust the certificate [14:15] the difference is you have to verify it manually, like you should be doing your ssh keys :) [14:15] RoyK: ok just wanted to make sure.. yeah it's for mail server [14:15] no it's not the same still [14:15] cause you can't verify the revoke status [14:16] well, for a private setup, it won't matter [14:19] blah patdk-lap sorry missed anything past "cause you can't verify the revoke status" === robertpayne_ is now known as robertpayne [14:21] hi guys how to check the ubuntu version is being usede by my server [14:21] any idea..? [14:21] cat /etc/lsb-release [14:33] How can I create a second qcow2 image for my windows guest? I did do "qemu-img create -f qcow2 win2k3-ext.qcow2 20GB" and edit the .xml, but when I start windows again i'm not able to format the new disk. The disk in Disk managment is only 1MB? [14:55] any package that has a bit of a gui for iptables on server edition? [15:45] robertpayne: any of the firewall gui programs can be X11 forwarded over ssh [15:46] papertigers: ahh didn't realize that thx [15:46] robertpayne: if you do ssh -Y user@server [15:46] then you could run something like firewallbuilder [15:51] damn it i think i will buy a hardware raid controller [15:51] incorrect: and use it to do JBOD and then do mdadm? :P [15:57] mdadm is the pain [15:58] i just put a 3ware 8 disk raid10 setup in my server, much nicer than mdadm [16:07] incorrect: I love mdadm [16:07] I have a rocketraid 2320 hardware card in mine [16:07] using mdadm for raid6 [16:08] i have 8 disks 5.4TB usable [16:39] papertigers, i guess you only use it for a file server not a vm host [16:43] Hello, is this the right place to discuss server documentation bugs? [16:49] incorrect: I use it for storave for mv VM's as well, linux kvm [16:49] kvm server is seperate, but is trunk linked point to point ip and mounted over nfs for storage [17:24] papertigers: ping [19:58] I'm getting "fuse: failed to open /dev/fuse: Permission denied" when I try to use sshfs [19:58] is there an alternative to sshfs? [20:19] caution: adduser $USER fuse [20:19] i think [20:19] I've tried that [20:19] I want to make a login script that all users run on logon. Where is the file where i need to link script? :) [20:19] caution: have you logged out and in? [20:20] yes, `id` shows the group [20:20] I also changed the group of /dev/fuse to fuse [20:20] caution: what permisions are on /dev/fuse [20:20] crwxrwxrwx 1 root fuse 10, 229 2010-08-08 19:33 /dev/fuse [20:22] what are my alternatives? I only want to securely share a file across two computers [20:24] fuse probably isn't supported by my host or something [20:28] caution: nfs over vpn should be pretty secure [20:29] or something with unison or rsync over ssh [20:32] nfs also uses fuse? [20:37] top - 21:37:11 up 9 days, 10:08, 2 users, load average: 24.20, 24.14, 24.10 [20:37] caution: nfs is usually in kernel [20:37] no fuse there [20:38] trying to install nfs-kernel-server tried to install fuse [21:22] caution, did you use --no-install-recommends? [21:28] no [23:50] Does anyone have any good suggestions for a password management solution for ubuntu-server? I have been mounting a encrypted image of a filesystem onto /mnt. Its not very elegant - Any suggestions? [23:54] steve__: I'm unclear on what password management has to do with mounting a filesystem ... can you elaborate?