=== TannerF is now known as [12] === [12] is now known as TannerF [00:53] hi all, i installed ubuntu server and it will not boot. just shows cursor blinking with black background === JanC_ is now known as JanC [00:53] when i boot into ubuntu server cd, i can tell it to boot from first hard disk and it works fine [00:53] tried re-installing grub, installing lilo, etc. to no avail [00:54] any ideas? [00:54] oh, i am using a 2.2TB hardware RAID array which is properly recognized by the installer and is formatted as ext4 [01:03] what package do ubuntu 10.04 server has for Qos?? I want to use it as internet gateway [01:04] doolph, I don't belive this could be achived by just installing a package. You'll have to configure from ground to top a firewall with QoS, [01:05] do you know where to start? [01:06] doolph, yes: studying iptables firewall. [01:07] doolph, you'll spend a great effort reading, studying, trying, but at the end, you'll be able to have the most 'cirurgical' QoS and other complex firewall settings. [02:38] I upgraded to Ubuntu Lucid and now when logging in, I find it takes at least 50 sec just to authenticate me to user and also a long time to authenticate me as root. Why is this? [02:42] I /did/ notice something like 60 console-kit-daemons loaded. I wonder if they load up every time I log in [02:44] yes, it's starting them up every time I login. Second, I noticed if I've already authenticated as root, it doesn't take any time to log me in. Something about entering in a password correctly is triggering the creation of all of these [03:02] Saturn2888: going on memory here, not at work to check my logs, but I had the same problem and I modified the /etc/sshd_config file, appending "UseDNS no" (unsure of correct syntax. Google it. After doing so, the logins returned to a more normal time period. [03:03] apt-get remove consolekit did not fix the problem, [03:03] Saturn2888, which authentication method are you using, PAM, LDAP, DOMAIN? [03:04] PAM I think. I didn't know you could use Domain or LDAP actually bc I'd rather prefer adding domain in :P [03:04] kkk [03:04] heh [03:05] Saturn2888, change your /etc/ssh/sshd_config file to "UseDNS no" and restart the service using sudo /etc/init.d/ssh restart [03:05] "UsePAM yes" Could I even add "UseDOMAIN yes" ? [03:05] second you should be using SSH keys to login with SSH and not passwords [03:06] https://help.ubuntu.com/community/SSH/OpenSSH/Keys [03:06] meh, it's not online, it's my home network, and I use PuTTY, and it's a bit of a pain to do the key thing. The 1 security flaw about keys is that someone gets the key and no password is needed. [03:07] is the DNS thing commented out or should I add it? Second, why didn't I have these issues in Hardy? [03:07] 1 security flaw is anyone sniffs your connect and you lose [03:07] You should add it [03:07] What do you mean about sniffing the connection? [03:07] we need to fix the obvious and lay a good base so we can further troubleshoot, right now you got two big problems [03:08] Cain&Abel [03:08] :P [03:09] login is still slow. I need to restart ssh don't I? [03:09] really? [03:09] yeah, still slow [03:09] http://pastie.org/private/xekad7aqtulkicszgvyuqa [03:10] i meant "sudo /etc/init.d/sshd restart" [03:10] wait wait. That's not the whole file [03:10] ok referesh [03:10] refresh* [03:10] it's ssh restart. sshd isn't a service in init.d [03:11] does the delay occur after you type password? [03:12] ok now we need to debug [03:12] ssh -vvv user@server [03:14] after that we will need to take a look at if you have the update motd script running [03:18] ok [03:18] http://www.walkernews.net/2009/04/06/how-to-fix-scp-and-ssh-login-prompt-is-very-slow-in-linux/ [03:18] im out [03:18] yes [03:19] after the password. Does it for root too, but does NOT do it when I do sudo su - right after having logged in as root meaning, if I do not authenticate again, I do not have the issue [03:20] bogeyd6: http://pastie.org/private/1ih8lnt1o9vaal6yuosfaa [03:22] I also keep seeing domainadmin@grubber:~$ debug2: tcpwinsz: 263536 for connection: 4 at the prompt and can't get rid of it [03:22] *** System restart required *** [03:22] *** System restart required *** [03:22] lol [03:22] LOL [03:23] bogeyd6: ? Where is that? And it does the password thing in tty [03:23] dude, "sudo shutdown -r now" [03:23] or just reboot :p [03:23] line 144 [03:23] great, well darn. I wish I'd seen that before [03:23] Wow, the great feat of Windows tech support [03:23] hopefully, just hopefully [03:24] yeah [03:24] I don't think that's it. I'm almost sure it did that after a restart [03:24] it will solve it, because your next step is to start PAM trouble shooting and changes lots of things [03:24] ? [03:27] I'm getting a keyboard [03:32] finally, now it's going down. I lost my DHCP lease in the process.... [03:34] logging in on the host machine itself was fast. [03:34] but it didn't reboot from terminal [03:34] only from tty [03:34] but I can't acces sit [03:34] access* it [03:37] aw great.. I restarted the wrong server. [03:37] that was my router. The reason i can't access it is because I'm at the grubrescue> screen. I dunno what to do. [03:41] booting now. I need to update-grub then grub install into the compact flash drive I have in there. I must've forgotten. then I'll swap out the USB drive with the Compact flash and should be okay [03:42] if you're wondering, grub2 wouldn't embed into /dev/md0 so I mad /boot on a USB, it worked bu it's so slow. I ended up moving it all to an IDE Compact Flash drive I had sitting around but must've forgotten some stueps [03:42] sounds like you got alot of work ahead of you [03:42] yeah, slow in tty1 for sure [03:43] well this should be about 1 min (after login) to fix GRUB [03:43] the rest is logging in being slow. It's not SSH-related [03:49] rebooting it. hoping this time it'll use the compact flash. Maybe it did last time and grub was like "WTF?" Hmmm. The USB isn't large enough to use anyway. [03:50] bogeyd6: yay works. Ok, back to this problem [04:18] I think it has to do with something in /etc/pam.d [04:18] Saturn2888: what are you trying to do? [04:21] login in both ssh and tty is extremely slow taking 30sec to 1min. I am trying to figure out why. I restored an old snapshot of the /etc/pam.d directory, no fix. I wonder what's causing it. Before, i noticed everytime I authenticated, I got a ton of console-kit-daemon --no-daemon processes which I used apt-get remove consolekit to kill. Then I also noticed that if I logged in as root, hit CTRL D, and did sudo su - before the tim [04:24] Saturn2888, the default motd in jaunty and newer takes some time to process [04:24] it checks system load and then checks for updates [04:28] qman__: oh ok. Where is it? I have it not to load over ssh, but that probably just stops it from being displayed [04:29] Saturn2888, /etc/motd.d [04:31] hmm only motd. [04:31] http://pastie.org/private/sty5kobyzckuhinvbx4yuq [04:32] hmm, must be the wrong path [04:32] ah [04:33] /etc/update-motd.d/ [04:33] Saturn2888: what auth methods do you want/try to use? [04:33] password? [04:33] superman? [04:33] So you aren't using, say, LDAP, or kerberos? [04:33] I dunno what you mea [04:34] This is a single-user single-host machine without a network connection? [04:34] oh oh, no no. I wish. I only found out today it was possible in SSH, didn't even know it was possible for tty [04:34] I get into it via SSH almost always. it is a single-user machine in that I am the only one using it, but some apps have their own users. [04:35] Oh, you said it's still only taking a minute to boot [04:35] http://pastie.org/private/sty5kobyzckuhinvbx4yuq [04:35] Unless this is a brand new machine, one minute is *good* [04:35] yeah, all those scripts run to generate the motd [04:35] only? no no, boot is a different issue. I'm talking about login [04:35] like "Ubuntu 10.04.1 Login: " [04:35] if any one of them takes more than a couple seconds, you're going to notice it [04:35] after typing in my pass it's like "let's go to sleep now" [04:35] Saturn2888: You mean the delay is after you enter your username and password? [04:35] and they do on every system I've installed [04:35] no [04:36] qman__: ok. So I check them all? [04:36] twb: it's the delay after entering a correct password to getting to the prompt [04:36] move the ones you don't care about out of that directory [04:36] which ones are which? Why are they there? [04:36] they generate the info that shows up in the motd [04:37] Saturn2888: then the problem is most likely in your .bashrc, .bash_profile, .profile, or other login scripts. [04:37] system status, updates, zombie processes, that sort of thing [04:37] it's a convenience feature [04:37] Checking motd is also a good idea, although I *thought* those were updated at boot time, not at login time. [04:37] not sure when it updates, but I notice with fresh installs that logging in takes longer because of it [04:38] even on fast hardware [04:38] and even when logging in long after booting has finished [04:38] Shrug. [04:39] qman__: do I restart something after doing this? [04:39] no [04:39] I'm moreso thinking it's in bash then like twb says [04:39] just log in and see if it fixes the problem [04:39] didn't fix it [04:39] well lemme move more things. [04:40] yeah, no fix from even completely moving the update-motd.d folder [04:40] ok then, must be something else [04:41] The only thing I changed in bashrc is something I changed on all my other systems, I added the colors [04:41] maybe something else went screwy. I want to try copying files from one server to this one and see what happesn [04:41] happens* [04:43] hmm. did anything change from hardy to now? [04:43] in the bash rc files [04:44] they have changed, but I don't think any of the configuration itself has changed [04:44] just the order and the comments [04:44] if your question is whether a hardy .bashrc will work on lucid without issues, the answer is yes [04:45] ok good [04:45] lemme try swapping the files now. Unless you guys can point me to some default ones [04:50] Saturn2888: put "date --rfc-3339=ns" at the top of your .bashrc [04:51] Then, when you next log in, you'll know how quickly it reaches that point [04:51] why's that? [04:51] ooh cool! [04:52] You can also try something like printf "${BASH_SOURCE[0]}:$LINENO" [04:52] I forget the exact syntax [04:53] For what? [04:55] 2010-07-25 22:53:50.296067986-05:00 if it's saying it only took 10 sec, that's a lie so this is saying it's probably def bashrc? [04:55] I had pasted it in the wrong chat :P [04:55] For when you have a dozen such timestamp lines; it prints out which file/line the timestamp belongs to. [04:58] oh I see, i want that. I dunno how to lookup the syntax for it. Never done bash scripting myself [05:00] go onto #bash and ask [05:01] haha oh yeah [05:01] 1 min [05:01] gonna swap out the file first [05:04] swapping out the .bashrc file didn't fix it either [05:04] qman__: haha, it's not fixing that way either [05:05] do you guys know what scripts run when you login? Maybe that's killing it [05:13] Saturn2888: /etc/profile.d [05:15] nothing in there [05:16] I find that hard to believe [05:16] twb: http://pastie.org/private/orjqbjlsjizbpw4h9j32pw [05:16] was there supposed to be stuff in there? [05:16] bash_completion, at least [05:17] Maybe ubuntu moved it somewhere "helpful" [05:17] hahaha === rmk_ is now known as rmk [05:17] I have some things in /home/user [05:18] http://pastie.org/private/orjqbjlsjizbpw4h9j32pw [05:25] twb: Something looks off to me: http://pastie.org/private/3plvoir17bujn2lpljrsha line 19 [05:29] I see something, the if /etc/bash_completion part. That file is in /etc === oubiwann-away is now known as oubiwann [05:29] although, this is commented out for root login [05:44] twb: Oddly enough, it gives me enough time when it's logging in to switch out the .bashrc file. [05:45] so it's whatever executes before bash [05:47] .bash_profile [05:48] ok [05:48] lemme log back in, I wanna also try changing the password [05:48] Whatever; I'm not hanging on your every word. [05:49] np np, sorry. I might wanna note ,if I type in my password wrong it does that normally [05:49] the speed that is [05:50] http://pastie.org/private/xcbagfsqv9ewytwd47bq [05:50] oops, that was my fault [05:50] does bash have anything to do with password authentication? [05:51] if not, that's the new place to look for this because authenticating this password to change it took forever [05:51] that's pam [05:52] back to pam then? [05:52] probably /etc/pam.d/common-auth [05:52] What could be the issue now? [05:52] ok [05:53] http://pastie.org/private/xcbagfsqv9ewytwd47bq [05:53] well, there you have it [05:54] you're attempting to authenticate with winbind [05:54] it's probably failing [05:54] oh ok, I am? [05:54] winbind and ldap [05:54] wait wait, so I can auth as a samba users? [05:54] what specifically do you mean? [05:55] oh, this machine is the LDAP server, now I remember. But LDAP stopped working after the upgrade. Ah ha! so I should comment it out [05:55] that's not going to solve the problem [05:55] ok [05:55] that file is now automatically generated [05:55] you need to fix or remove ldap [05:55] and winbind [05:56] ok [05:56] otherwise, next time the pam update script runs, it'll be slow again [05:56] you can comment as a temporary solution [05:56] oh [05:56] it's okay. I'll just remake my samba users again [05:56] I have no clue why it's not working anyway [05:56] winbind/ldap is a separate thing from samba [05:57] what exactly are you trying to do? [05:57] you CAN authenticate samba users via ldap/winbind, but they don't by default [05:57] or was I trying to do. Does uninstalling those fix pam? [05:57] might, though you might need to purge them [05:57] I had this as an LDAP master and my other server as an AD server which got users from LDAP on this machine [05:58] the reason it's taking so long is it's timing out attempting to authenticate against winbind or ldap, or both [05:58] one of them is gone, the pam-ldap.so use_first_pass is still there. Commenting it out [05:59] no fix though [06:00] you only commented it out in one method [06:00] common-auth [06:00] it more than likely exists in others [06:00] and when pam updates again, it'll uncomment [06:01] you need to fix/remove ldap [06:01] pam-auth-update , ran that, didn't fix it though [06:01] oh oh! fixed [06:01] I forgot to uncheck samba. weird, at least now I have more control over this [06:14] Thank you guys so much! [06:14] I'm heading out [06:43] Hi [06:44] hi lwizardl [06:44] I know Landscape is a payware type of service but does there exist something similar to it for free usage ? [06:44] Depends if you want to babysit both ends [06:45] twb, what do you mean? [06:45] landscape is a service [06:46] well what I am looking for is something that will let me manage everything from a single location [06:46] i think it's proprietary, not entirely sure. But even if it was free, they would still charge for the service. [06:46] I know I could use stuff like cpanel, ebox, webmin, etc to do most of it [06:47] i don't think landscape does what you think it does [06:47] yeah I am not looking for someone else to do the work. more a just a single program versus using shell to handle this, then that, etc [06:47] ah ok [06:47] then I must have been confused [06:47] landscape will let you manage a pool of VM's [06:47] systems etc.. [06:48] so packages etc.. on the boxes [06:48] last time I looked at landscape was around the first advertisement date on the server download site [06:48] oh [06:49] cpanel and lanscape do have some crossover in terms of functionality but i think they are fairly different beasts. [06:49] landscape is probably what you want if you ahve a number of servers to manage [06:49] and you want to administer/ provision them automagically [06:50] ok yeah I just have 1 server and 10 domains on it [06:50] you probably don't want ladnscape. [06:50] that I manage manually [06:50] yeah i don't think so [06:50] now if i had like 5-6 servers then yeah [06:51] ok [06:51] thanks [06:51] what are the main issues for you wrt to managing the server via CLI ? [06:52] is it just unfamiliarity or do you need to give other people access to things? [06:57] mase_wk, I am fine with the cli just sometimes people ask for access to this or that [06:57] and then i have to remove it etc [06:58] also would like to be able to have printable charts of usage etc [06:59] there are a few things here and there I don't know how to do but a quick google search usually helps for those [06:59] yeh i understand what you mean. i've had to do something similar with some of my servers however I basically just created an ldap server, and hooked apache/ postfix / everything else up to the ldap server for auth [07:00] and then wrote a web frontend to which ever parts of the ldap records i wanted people to be able to change [07:00] lwizardl: cfengine, puppet, chef, etc. provide a centralized mechanism for managing a network of heterogeneous systems [07:00] yeah what I was doing was more setup a temp ftp and link to those files needed. and then after the login and they got that file i remove the account [07:00] hi there, i need to control which user are accesing p2p connection from internet. Somebody have any idea about any tool? [07:01] Roxyhart0, that seems to be a firewall relationship issue. I would look into ipcop [07:03] i mean, p2p is open...i just need to control when the user are downloading stuff with copyright issues [07:03] Roxyhart0: thats pretty hard to do unless you know which files have copyright issues in advance [07:03] Roxyhart0, ah ok I thought you was trying to stop all p2p on the network [07:03] no yet [07:04] Roxyhart0, that program i listed has a option to block all p2p traffic on the network, and lots of other stuff. I haven't set one up yet but it will be done this week after my network is installed at the location [07:05] mase_wk: the other problem with landscape (apart from being proprietary) is that the client is only available on Ubuntu [07:05] twb: ah ok so it handles stuff at the application level etc.. too ? i dont use landscape ( b/c it's proprietary ) [07:05] AFAIK landscape is basically just a cfengine-type system with the server side, and its web ui, being proprietary [07:06] thanks...at the moment identify user who download material with copyright issues. It is more "after" the problem as we still are not going to block p2p [07:06] twb: ah ok. fair enough. [07:07] Roxyhart0: how do you handle encrypted p2p ? [07:08] is this for an office / corp environ or something like a wifi hotspot ? [07:08] this is a educational institution [07:08] an [07:08] k [07:09] and i dont know, but the most of the traffic p2p is not encrypted i suppose? [07:11] not really sure i guess it depends on your users. the most useful way i've managed to combat p2p in an office is by heavily shaping everything that isn't a service we actively use [07:11] and offering to download torrents at high speed on behalf of users. For those that want new distro torrents or other legal things [07:12] I don't think so but most p2p today seems to be torrents and most clients are setup to use encryption [07:12] we had alot of encrypt on our network [07:12] hence why i couldn't really stop it per se [07:13] but i could make it annoyingly slow [07:13] and those that want something quick know they can get it quick if it's legal. [07:13] the ipcop program one of my friends used when he did a bunch of network setup installs for a church community center [07:14] and set it to block stuff like kazaa/limewire/frostwire bittorrents [07:14] i will check it, i leasen about ipp2p and Dante(as proxy) as well...but doesn do that im looking to do now [07:14] only thing that still works for what he told me was stuff like megaupload etc and he just added those to the blocked sites [07:15] plus that has a option to setup cache for updates so if you have say 10 computers on the network that all need to get updated it only grabs it once from the internet and the other 9 computers grab a cached version from the firewall [07:16] mase_wk, how you can do slow the traffic if you cannt detect if it is p2p? i mean encrypted traffic? [07:16] i think its just a modded version of monowall/shorewall but has lots of nice features [07:16] Roxyhart0: i dont' detect it's p2p , i just shape everything and unshape the protocols we actively use, ssh, http, ssl etc.. [07:17] ah ok... [07:17] the shaping works well because it's not like it doesn't work. so people just assume the torrents are slow [07:17] esp when generic web access etc.. is quick [07:18] so most of the time they just give up [07:18] haha... [07:18] if they complain, they usually complain about a legal torrent, in which case i explain that it's shaped etc.. [07:18] and then download it for them [07:18] mase_wk, yup but the more tech people will try and use a proxy and see if that will bypass and get faster [07:19] lwizardl: proxy won't help in this scenario unless it tunnels the p2p over http [07:19] which means your downloading it twice [07:19] ah true [07:19] once remotely and then pushing it through the proxy [07:20] which people who run proxy's don't like either [07:20] same applies to an ssh tunnel [07:20] in both cases people who understand it realise it's easier to do it from home :) [07:20] yeah just saying if someone wants it bad enough they will try other options [07:20] yes [07:21] or look for another wifi spot and try with that instead [07:22] yeh just makign it easier to do something else is usually enough [07:23] yup and then if something happens you have tried to stop people from doing it [07:24] me I plan to offer internet lan access in my store and I want to make 100% or as close as possable to stop any p2p downloading [07:25] i read dante works with p2p [07:25] and I also want to try and block access to adult related sites which is both something that the ipcop program does and then it has more options that I can make use of also [07:26] i will have a look at ipcop [07:27] yeah the only issue I have seen is that your box you install it on will need to have atleast 3 nic cards installs [07:27] so you can set them for how trusted the network you want Red/Blue/Green [07:28] basically Internal network only, external network only, and both [07:30] so, the nic with both is going the trusted traffic? === rmk_ is now known as rmk [07:31] yeah [07:32] the adult sites can be handled by blocking dns requests to anything other than your dns servers [07:33] and has full access, external only can visit web sites like google etc, and internal only is for computers you need to do other stuff on like your backend server keep track of POS cash registers etc [07:33] or you can use a service like openDNS [07:33] and only allow dns requests to those dns servers [07:33] mase_wk, yeah I was going just display a generic "sorry these types of pages are not allowed on the network" [07:34] and then also disable the usb ports so people can't bring files from home and save them onto the computers to cause problems [07:35] I want to set the machines to do like the computers at my old college did. after you reset the computer it booted a set system and automatically removed any files a user may have left or installed [07:36] can you do that with linux? I know with AD form windows ypu can do, i mean reset policies in clients (windows) but i am not sure if it is possible with linux server, could be great [07:36] lwizardl, should be pretty straightforward if you build an image and set them to netboot it [07:36] most of the trouble in setting up a netboot system is the persistent data [07:36] qman__, yeah I think that is what they did [07:36] but since you don't want any that takes that whole bit out [07:38] in absence of adequate network resources, you could build a custom live CD [07:38] but then you'd have to burn a new CD every time you patch it [07:39] yeah and that would be a waste [07:39] Roxyhart0: depends if your clients are linux or windows. I don't believe you can give the windows clients those sorts of profiles from a openldap etc.. like you can with AD [07:39] but if you have linux / osx clients you can [07:40] I will look into that because the computers I am running as internet terminals aren't that great of machines but for just net access they work great [07:40] they are running Ubuntu 10.04 [07:41] any idea why when I try and add this to iptables it fails: http://starfleet.pastebin.com/gQtgRsfg [07:41] yes the most are windows and mac [07:41] I'm using old Compaq IPAQ desktop computers 550mhz with 256mb ram, and a 10gb hdd [07:41] #netfilter says its the kernel [07:41] i'm not sure what to look for to fix it [07:42] Roxyhart0: well the mac ones with auth against openldap. I don't / haven't used windows clients since 98 so i'm not sure how they will deal with an openLDAP server. [07:42] mase_wk, they won't [07:43] I keep 1 windows desktop around and thats just for editing commercials from my ReplayTV DVR other then that I'm fully Ubuntu only on the other machines [07:43] you need samba, and some serious hacks [07:43] qman__: ye figured as much [07:43] and even then it's only partially working at best [07:43] ugh [07:43] upgraded to lucid, it picked a resolution out of range [07:43] in the most of the suff are ok, but for apply policies...no unless there are some way that we dont know [07:44] I have 1 windows machine, 1 server, 4 terminals, 2 linux pos computer cash registers, 1 office pc, 1 soon to be firewall, and 1 backend server [07:44] and 3 computers that are soon to be installed into MAME arcade cabinets [07:45] can anyone remember where hardy stored it's getty configuration ? [07:46] Hi EvilPhoenix, look this page http://www.shorewall.net/OpenVZ.html [07:47] "if you see annoying error messages as shown below during start/restart, remove the module-init-tools package from the VE" [07:47] Roxyhart0: the issue isnt with OpenVZ, the issue is iwith the kernel image loaded on it [07:48] mase_wk, the per-TTY configs are in /etc/event.d [07:48] its not reading the modules necessary to detect ESTABLISHED,RELATED connections on iptables [07:49] qman__: thats it . thanks. don't suppose you can copy/paste your getty line from /etc/event.d/tty1 ? I am de-xenifiying a virtual machine [07:49] yeah [07:49] exec /sbin/getty 38400 tty1 [07:49] thank you [07:50] no problem [07:50] now to figure out why my video mode changed [08:05] so, for some reason [08:05] video modes that worked fine in hardy, now show 'out of range' after upgrade to lucid [08:05] if I remove the modeline it works, but obviously it's low resolution [08:06] video card is an ATi Rage XL [08:13] don't tell me I need to install grub2 to get more than an 80x25 terminal :/ [08:27] How can I detect the fastes Ubuntu mirror to set in my sources.list file? [08:37] New bug: #570456 in qemu-maemo "Unimplemented cp15 register write (c9, c12, {0, 0}) with Ubuntu OMAP image" [Medium,Fix released] https://launchpad.net/bugs/570456 [09:25] hi there, there are someway to identify users by IP address? [09:26] i mean who is using an IP? [09:27] Roxyhart08: google for host based authentication [09:27] thanks [09:29] It depends on a range of factors, like whether you have control of the network and all the machines that use it. [09:37] New bug: #609979 in mysql-5.1 (main) "aptitude install mysql-server freezes at dpkg" [Undecided,New] https://launchpad.net/bugs/609979 [10:04] Hey folks .. when is the 10.04.1 release [10:25] kim0: https://wiki.ubuntu.com/LucidReleaseSchedule looks up to date [10:26] jcastro: you're da man :) [10:55] after getting ip from dhcp server.. and only ubuntu lucid image [11:40] <_ruben> hmm, speaking of 10.04.1, my desktop vm just announced itself as 10.04.1 .. that's odd === TREllis_ is now known as TREllis [12:18] good day [12:18] how can is send mail to jigsneth@yahoo.com [12:18] how can i send mail to jigsneth@yahoo.com? [12:27] silentwhisper: assuming you have a mailaccount, in the to field of your client(webbased or other) you paste the adress. If you prefer you can write something in the subject line. In the usually bigger square you write your email and press send. Some clients require you to press send & recieve button or similar. If your question is about some more server-side stuff like postfix i prefer you alter your question [13:00] Howdy. I recently replaced the motherboard in my Ubuntu Server 10.04 install, and now the network isn't working - is there any way to get Ubuntu to redetect and reinstall the correct drivers without having to reinstall the whole OS? [13:00] Any idea why a fresh install of mysql on a fresh installation of 10.04 gets stuck when running "start mysql"? [13:08] Database: Do you have an onboard NIC? [13:11] Database: did your NIC get detected as eth1, or eth2 or something? [13:11] Database: are you sure you nic is invisible? ifconfig -a what says? [13:12] Database: Your new NIC is probably mentioned in /etc/udev/rules.d/70-persistent-net [13:13] The old one will still be mentioned there, probably [13:13] hmmm. [13:13] Okay. [13:13] Jeeves_, yes, it's onboard. [13:13] hang on, I'll go check to see if it's detected as eth1. [13:13] it's probably mapped your old NIC's mac address to eth0 [13:15] not sure what the 'official' fix is, but i usually just grab the new mac address and replace what's mapped to eth0 ... and then remove reference to the newly added device (eth1, eth2, whatever) [13:16] mattt: Removing that file and rebooting works too :) [13:17] Jeeves_: does it just recreate it? [13:17] i like that option better ... cuz when i'm adjusting that file, it's usually over KVM, or something ;) [13:19] okay, so I can just delete /etc/udev.rules.d/70-persistent-net? [13:19] I've found my new NIC is mapped to eth2, so you are right. [13:20] Database: do the needful [13:20] :) [13:20] Database: Removing it is ok. It will be recreated on the next reboot [13:20] altering it is also ok [13:20] (and might help you to understand what happens better) [13:39] okay [13:39] It works :D [13:40] Thank you all :) [13:40] woot! [14:01] New bug: #610066 in commons-io (main) "Sync commons-io 1.4-3 (main) from Debian unstable (main)" [Wishlist,Confirmed] https://launchpad.net/bugs/610066 [14:03] alright, got my laptop back [14:17] How do I mirror packages between two systems [14:17] there was some command to dump the package list [14:17] and to get it on the other system [14:17] !clone | chrismat [14:17] chrismat: To replicate your packages selection on another machine (or restore it if re-installing), you can type « aptitude --display-format '%p' search '?installed!?automatic' > ~/my-packages », move the file "my-packages" to the other machine, and there type « sudo xargs aptitude --schedule-only install < my-packages ; sudo aptitude install » - See also !automate [14:20] thanks ubottu [14:20] You're welcome ;) [14:21] !clone [14:21] To replicate your packages selection on another machine (or restore it if re-installing), you can type « aptitude --display-format '%p' search '?installed!?automatic' > ~/my-packages », move the file "my-packages" to the other machine, and there type « sudo xargs aptitude --schedule-only install < my-packages ; sudo aptitude install » - See also !automate [14:21] !clone [14:22] !automate [14:22] Ways to automate installation of Ubuntu on multiple machines are described at https://help.ubuntu.com/10.04/installation-guide/i386/automatic-install.html - See also !cloning [15:04] I have an active connection to an iscsi target. I have added an extra lun to this target, however the added volume does not show up at the server with the active connection. How can I make it recognice the change without breaking the connection? [15:10] hi all [15:10] kirkland: ping [15:11] ping you too! [15:12] I have a question about UEC [15:12] anybody home? [15:13] TuxSax: just go ahead and ask the question; if anyone here has the answer I"m sure they'll share [15:14] jiboumans: hi there [15:14] long time no see [15:14] ivoks: hey [15:14] i know you hate me :) [15:14] ivoks: hah, yeah i dodge the channel ;) [15:15] ivoks: kirkland's boarding a flight right now [15:15] ok [15:15] with a little luck he's back online in ~3-4 hours [15:15] I was thinking about setting up our own private could at our company, and I'm trying to understand how exactly can we benefit from it's use [15:15] I still don't get a simple answer to what exactly a could is [15:16] TuxSax: well that depends a lot on your needs i suppose [15:16] * cloud [15:16] let me tell you what I was thinking and you can tell me if I'm in the right direction [15:16] TuxSax: out of curiousity, did you see http://www.ubuntu.com/cloud and cloud.ubuntu.com/ yet? [15:17] we have needs for linux servers, from time to time we need to add a server, or to upgrade/reinstall an old one [15:17] Yes, but the "human readable" information I was trying to confirm doesn't exist there, a lot of buzz but not a real simple answer to what I need [15:18] I was thinking about setting up a initial setup of two or three servers, according to ubuntu documentation a basic cloud consist of one controller and one or more nodes [15:19] so I was thinking about setting up a couple of servers, and be able to add more nodes as I migrate some of the services I already run, to virtual servers, and then some of the hardware that get's free will become new nodes [15:19] TuxSax: yeah, you can set up the management components on a single machine and then add single servers as nodes [15:20] tuxsax: this talks you through it https://help.ubuntu.com/community/UEC/CDInstall [15:20] my question is, if I set up, let's say five nodes servers, does it mean they all run as a kinda big single server sharing their hardware and storage or I've got the wrong concept? [15:21] I don't have problems with the setup or the howto, I know the documents online, I just want to be sure I understand the concept [15:22] tuxsax: if you use s3 and ebs, yes you can consider the storage as one big shared pool [15:22] when they talk about running an "instance", does it mean I can run a LAMP server in one instance, and an FTP server on another instance and so on? [15:23] you have multiple servers [15:24] it's not one big server, like HPC cluster; it's pile of virtualized servers on top of couple hardware servers [15:24] is an ubuntu cloud comparable to VMWare ESX or Citrix ZenServer hypervisors? [15:24] all independet [15:24] TuxSax: UEC is actually exactly comperable to Amazons EC2 [15:24] the virtualization is one part of that [15:25] it's much more than ESX :) [15:25] what I'm trying to gain is scalability, so if I run 5 virtual servers and let's say I feel a bit of load in the system, I can add another node to share the load? [15:25] YEah, also Amazon EC2 is something I'm trying to understand, it's the same concept [15:26] tuxsax: if your system benefits from more hardware, then yes of course that works [15:26] if you add a node, you can run more virtualized servers basically [15:26] what I didn't get about Amazon was that in the howto they said "remember to stop the running instance, you're paying as long as it's running" [15:26] tuxsax: amazons billing scheme is based on the amount of hours the instance is 'up' [15:26] (idle or not) [15:27] tuxsax: with UEC, your system is in house so you don't have to pay amazon ;) [15:27] but if I run a server to server files for other people, what good can be to close the instance? [15:27] I can't imagine what service could I possibly want to run for only several hours... [15:27] tuxsax: if you're using an instance, obviously you wouldn't turn it off [15:28] anyway, I prefer to run my own cloud, I was just trying to understand Amazon's service [15:28] so, the work load of all the instances is load balanced among all the "available hardware" of nodes? [15:29] TuxSax: I still don [15:29] 't get it.. [15:29] so if my five instances are running a little slow and I add two new nodes, all of them will benefit from this "processing power" addition? [15:30] tuxsax: it's not balanced automatically, no [15:30] mmm, then? who decides on what node every instance runs? [15:31] Jeeves_: what exactly you still didn't get? [15:33] hi guys. have a weird problem. virtualbox on win7 host, guest os is ubuntu server. pinging the server on 192.168.0.40 sometimes replies from 192.168.0.41?? [15:33] TuxSax: The whole use/idea behind this Cloud stuff [15:34] Jeeves_: That's exactly what I'm trying to understand too... [15:36] TuxSax: Please, do keep on trying. Maybe I'll get the picture too, some day :) [15:36] jiboumans was trying to help me out here but he dissapeared on me... ;-) [15:37] TuxSax: sorry, i'm on calls for the next 90 mins [15:37] oh, you're on duty... [15:38] tuxsax: if you're only looking to virtualize your current hardware, you may want to look at just running KVM [15:38] tuxsax: beyond that, and the UEC/EC2 feature list is the best explenation at the moment, a personal cloud may make sense [15:40] you mean to run KVM on top of a cloud? [16:02] tuxsax: no, just kvm on the hardware [16:06] jiboumans: I think that what I actually need is http://www.beowulf.org/overview/index.html [16:09] a cluster that allows me to share the load of a few virtual servers that run on top of the cluster [16:09] and then having the option of adding processing power, if needed, by adding more nodes to the cluster [16:10] I was sure that a cloud could give me the same... [16:10] guess not... [16:11] nope [16:11] that's hpc cluster [16:11] New bug: #610103 in migrationtools (main) "package suggests editing an installed non-conffile for normal operation" [Undecided,New] https://launchpad.net/bugs/610103 [16:13] TuxSax: note that your app would have to be cluster aware [16:13] HPC? [16:13] TuxSax: if you plan to have web server, then you just need lots of hardware to set up apache load balancing [16:14] I don't actually mind what exactly to use as long as it gives me what I need [16:14] you need to scale load on multiple servers, right? [16:14] right now I have a few servers that run several tasks, not only apache [16:15] a couple of apache servers, a couple of DNS, a network tool server, a backup server, and similar [16:15] what's the load? is it network connections or cpu load? [16:15] you basically need high availability cluster with load balancing option [16:16] so I was thinking about taking one or two and start a kind of cluster or cloud and move to there those services, and then, once a service is moved from a physical server to a virtual server on the cloud/cluster === dendrobates is now known as dendro-afk [16:16] doh... [16:16] let me guess what you want :) [16:16] I could take the free hardware, reinstall and make it join the cluster/cloud as a new node, thus adding more power to the main system, then I can migrate another services to there and free up another box [16:16] you have multiple services? [16:17] right? [16:17] and so on, get the idea? [16:17] ok, let's say you have 4 servers, each 4GB of RAM and two cores [16:18] that means you have 16GB of RAM and 8 cores to use [16:18] yep, that's more or less the kind of servers I have [16:18] you can set up cloud on top of that [16:18] but instead of running 4 different ubuntu servers that each one does something [16:18] meaning you can have 8 virtual machines, each could have its own core and 2GB of RAM [16:18] (more or less) [16:19] you could have 16 servers, with each 512MB of RAM [16:19] or you could mix [16:19] ok, but what you're saying is that the load distribution is kinda static and predecided [16:19] two servers for DNS, each with 256MB of RAM [16:19] exactly [16:19] nothing would scale automaticaly [16:19] and beowulf isn't for DNS/web/etc... [16:19] it's for computing [16:19] on what stage do I have to decide? when creating a new instance? [16:20] instance = running operating system [16:20] so, you could do with it what ever you want [16:20] having cloud for just scaling apache on multiple machines is pointless overhead [16:20] is there documentation on Ubuntu-Server 10.04 for RAID... i see this page: https://help.ubuntu.com/community/Installation/SoftwareRAID <--- which is for ubuntu 9.10 -- is htis also applicable to 10.04? [16:21] think of cloud as a pool of hardware [16:21] if you have 4 hardware servers, you could set up multiple machines on top of it... 4, 8, 16, ... 256 [16:21] smoser, ping [16:22] but of course, 256 virtualized machines won't work as good as just 4, since hardware is limited [16:22] but it has to be a multiple of a total [16:22] jiboumans, here. [16:22] smoser: hi, you're tough to get a hold of [16:22] no, it can be any number :) [16:22] this morning has been a bear [16:22] TuxSax: you know what KVM is? [16:22] or vmware [16:22] smoser: sorry to hear =/ is now a good time for a call, or should we try later in the week? [16:22] yeah, just played around a bit with it on my PC [16:22] vmware manages multiple virtualized machines, right? [16:23] now is good. jiboumans [16:23] so you can have 2,3,4... virtualized machines on one server [16:23] yep [16:23] so, cloud would be the same thing with one exception [16:23] * ccheney brb, rebooting irc box [16:23] smoser: cool, join me on mumble? [16:23] y [16:23] it wouldn't be one server, but 4 or how many of them you have [16:24] but when you run an instance from the cloud, what really happens? the server runs a kvm session on one of the nodes? [16:24] exactly [16:24] well, not exatly KVM I guess [16:24] it's kvm [16:24] what runs in beneath? [16:24] kvm itself? [16:24] nice [16:24] kvm is virtualization layer [16:24] but I don't have to care much about it, right? [16:25] bellow it there's eucalyptus that decides where to run what [16:25] the cloud manager makes it happen [16:25] right [16:25] kvm is the so called "hypervisor" ? [16:25] yes [16:25] now i'm sorry, but i have to go [16:25] fine, so what the cloud actually gives me is the ease of managing? [16:25] :) [16:26] ok, thanks for your info! [16:26] it merges all your hardware [16:26] this isn't something you can achive with kvm by it self [16:26] you have pool of ram, cpus, etc... [16:27] and it has potential to automaticaly start an instance once your load is high [16:27] so it does merge the sum of all hardware, so if I move a service to a new cloud instance and I get a free server, I can add a new node and gain more power to the pool? [16:27] but i'm not that much into cloud to know if that's available now [16:27] * ccheney back [16:27] yes [16:27] cool, then it does give me what I planned! [16:28] you've been a real aid, ivoks! [16:28] now it all makes sense [16:28] 10x a lot [16:28] (i have never used cloud in my life) [16:28] :D [16:28] I'll try and tell you how it feels... ;_0 [16:29] more power = more available cpus and ram [16:29] it won't automaticaly add cpus to instances :) === dendro-afk is now known as dendrobates [16:46] Daviey: morning/late afternoon [16:47] hggdh, Hello sir, good trip back? [16:47] Daviey: fantastic, slept about 8 hours of the 10-hours flight :-) [16:47] \o/ [16:47] Daviey: you were going to upload a new euca 2.0 -- did you? [16:48] anyone want to help a newb through raid configure in 10.04? i did this back in 8.10, i got it to work without much pain. [16:48] hggdh, Hmm.. not today.. nothing urgent changed. [16:48] i can either start after i install the server or before i install the server, i got 5 drives, 4 of which need to be part of the array, and the first 1 is the boot drive. [16:49] i got 4 300gb drives, and ultimatly i'd like to get like 10 raid, w/ 500gb [16:49] Daviey: OK. I will go back to the basics, and will start trying it on Maverick, from all-in-one [16:49] AndyGraybeal, Install time is much easier IMO. [16:49] hggdh, Good idea.. i'm doing the same [16:50] Daviey: okay thank you for the insight. i'm a little confused by the partitioner program. i go to "Configure Software RAID" and it says i's gonna change the partitions on the first drive (which is not necessarily what i want, i think atleast) is this correct procedure? [16:51] i want the first drive not in the array, but the next 4 in the array [16:51] for some reason I cannot get any of my scripts that use sendmail to send mail when they are run via cron [16:51] they work perfectly when ran at the CLI [16:52] AndyGraybeal, it's a two stage process... Ignore the first drive to start with.. and create a new raid partition on the others. [16:52] then you get the raid device as a block device (ie, the installer sees it as a disk) [16:53] Daviey: thank you agian, [16:54] a basic script I am trying to test with to send the mail, if I use a proper email it works from CLI not cron, cron will run the script and do the commands within however it just wont send the mail and there is no log of an attempt in any of the /var/log/mail.* files [16:54] http://pastebin.com/ryuup4UB [16:55] 00 5 * * * bash /scripts/test >/var/log/backup/test >/dev/null 2>&1 [16:55] that is the entry in crontab [16:56] these scripts work perfectly on my gentoo server [16:58] Cubber: when running a script from crontab you need to be sure the full paths for all commands are exact [16:58] so sendmail for instance [16:58] interesting that they work fine in gentoo as is but I will edit it with the full path to sendmail and try [16:58] Daviey: \o/ we now have access to ppa.lp.net under tamarind! [16:59] Cubber: if you run the /scripts/test script and on the script it says "command bla bla bla" it should be "/path/to/command bla bla bla" [16:59] what cron program you use on gentoo? vixiecron? [17:00] yes [17:00] perhaps the difference comes from there [17:00] probably thank you for the info I will test [17:00] looks like sendmail is in /usr/sbin/sendmail [17:00] hggdh, Yes! I was going to say - but i saw they CC'd you \o/ [17:01] Makes life easier, eh? :) [17:01] indeed it does :-) [17:01] <\sh> hey hggdh... [17:01] I always put all the commands with full paths under a variable on my scripts, so when calling them I don't need to worry [17:01] \sh: cheers, how are you? [17:02] <\sh> hggdh: fine so far...a bit tired after a hard weekend of datacenter work (moved 4 racks from one city dc to another city dc) [17:02] so for sendmail I'd set a variable called MAILER=/usr/sbin/sendmail or whatever mailer is used [17:02] then later on on the script I'd use $MAILER bla bla bla [17:03] \sh: and I had a bout of insomnia last week, and netted just under 20 hours of sleep in 7 days... [17:03] ok, I'm outta here, see ya all [17:05] okay.. i got it.. it was confusing because i have to first 'partition' the drives... geesh [17:05] <\sh> hggdh: oh but looks like that I have to force myself to not sleep tonight, because I have to deploy some machines till tomorrow morning... [17:06] \sh: then... Welcome to the Sleepless Club. We are glad to have you here, and all that ;-) [17:06] <\sh> hggdh: btw...the tool I talked about is released...launchpad.net/dc2 [17:06] \sh: looking at it now [17:07] <\sh> hggdh: on my blog there are more informations...and on dc2.sourcecode.de [17:07] \sh: cool, thank you. Any intention of packaging? [17:07] <\sh> hggdh: more docs are coming this and next week when I'm back on normal sleep schedule [17:07] heh [17:09] Daviey: thank you for the hand holding. [17:09] <\sh> hggdh: it's not ready for packaging, but we are planning to finalize V1.0 and to do packaging for debian and ubuntu,we need some more deps packaged before we can do the real upload :) [17:09] \sh: no prob. I am really interested in checking it [17:10] AndyGraybeal, awesome [17:10] <\sh> hggdh: people from the FAI group already got some hints...http://michael-prokop.at/blog/2010/07/08/report-from-fai-developer-workshop-072010/ <- this is the report from the FAI developer workshop there are some infos about the tool [17:13] \sh: so I understand you will eventually deploy FAI/(DC)^2 at work? [17:13] <\sh> hggdh: it's already deployed here at my company [17:14] <\sh> hggdh: it's already working since more then a year :) [17:14] \sh: heh. Feedback from the field is good, I guess :-) [17:15] <\sh> hggdh: I already have a lot of feedback several admins from dif. companies do want to have this...they are eager to test it :) [17:16] \sh: cool. I will bring this up here then [17:43] Hey, I just installed Ubuntu Server onto an old server. Everything worked fine until I restarted. Now I'm getting this error: "I9990301 Hard disk drive boot sector error" and then "I9990305 Operating System not found [17:51] silent1mezzo: might be there's a bad sector or two on it... [17:52] is there a way to test this from Ubuntu? [17:52] silent1mezzo: boot on the cd and use badblocks [17:53] ok thanks [17:53] if the system is just installed, using _destructive_ read/write might be a good idea [17:53] just keep in mind you'll need to reinstall afterwards [17:53] whats _desctruve_read/write? [17:54] but boot on the live (desktop) cd - I don't know if badblocks is on the server cd [17:54] silent1mezzo: it overwrites data on the harddisk to check for bad sectors [17:55] non-destructive testing may not find errors that easily (or at all) [17:55] ok [17:55] thanks [17:55] but destructive doesn't mean it messes with your hardware [17:56] that is - if you have data on a drive that you want out, don't do anything like that - just get the drive out and as fast as possible and don't write to it, but then, this doesn't seem to be the case - right? [17:57] aye [17:57] this is, I've formatted my disk, installed the os and tried to boot it [17:57] how much memory do you have in the box? [17:58] 8gb [17:58] oh - that'll suffice :) [17:59] lol [17:59] what sort of drive is this? [17:59] older 3.5" sata drive...don't remember the make/model [17:59] ok - try badblocks [17:59] ok, will do [18:00] man badblocks once you're running on the live cd [18:00] aye, I'll have to download/burn the desktop cd first [18:00] k [18:02] thanks RoyK [18:02] np :) [18:03] mind, if the memory isn't ECC, it might be a good idea to run memtest86 to see if something's wrong [18:03] well, even if it _is_ ECC, if it's an old server.... [18:05] I have been trying to get simple vlan trunking working for the past 5 hours, and thus far all i have managed to do is make my wireless "not managed" [18:05] Jinxed-: :) [18:05] note: currently im using the desktop version, not server edition [18:05] I installed vlan [18:05] Jinxed-: And that's about it :) [18:05] and set net.ipv4.ip_forward = 1 [18:06] I also edited the /etc/network/interfaces configuration [18:06] add interfaces in /etc/network/interfaces like so 'auto ethX.' [18:06] And that should be it. [18:07] Here is my current config: http://paste.ubuntu.com/469425/ [18:08] a diagram of my simple setup: http://imgur.com/MJm6t [18:08] two laptops on different vlans and one laptop to route between them [18:08] what sort of switch is this? [18:09] there are several trunking protocols - does yours support 802.1q? [18:09] 'auto eth1' [18:09] that shouldn't be there, you're not configuring that one [18:09] cisco [18:09] 2960 [18:09] yeah, but cisco what? [18:10] any letters after that? [18:10] cisco has lots of different software versions [18:10] RoyK: Cisco2960 is a normal switch [18:10] the 2960 is new, and supports dot1q [18:10] ok [18:10] I already checked the config on the switch it is setup for 8021q [18:10] ok [18:10] just asking [18:10] haha no i appreciate it [18:11] Jinxed-: And what interfaces do you get on the machines? [18:11] Not sure I understand your question [18:11] on ubuntulap, what does 'ifconfig' say? [18:11] Jinxed-: What RoyK says :) [18:12] http://paste.ubuntu.com/469426/ [18:12] no vlan interfaces are up there [18:13] Nope [18:13] an interface can be in 802.1q mode or non-802.1q mode - not both [18:13] Jinxed-: And what happens if you type 'ifup vlan10' ? [18:14] imho eth1 shouldn't be up at all [18:14] perhaps that's what's blocking it [18:15] RoyK: That's true. [18:15] But it may be up [18:15] I mean, it's allowed to have a untagged interface [18:16] if an interface is in 802.1q mode, it will need vlan tagging. if you send untagged frames to a switchport defined as a trunk port, the frames will be discarded [18:16] sorry that last command got me kicked off [18:16] (OTOMH) [18:16] RoyK: That's not true. [18:16] You are allowd to configure an native vlan on a trunk [18:16] Jeeves_: what should the switch do with those packages then? send them to all vlans? [18:16] I have not seen anything after when Jeeves asked what happens if you type ifup vlan10 [18:17] RoyK: it was the disk :P I just tossed it and installed it onto a new disk...Runs fine [18:17] untagged packets will be send on the native vlan [18:17] My wireless device is now not managed again [18:17] silent1mezzo: :) [18:17] and it said [18:17] Jinxed-: You're not trying to do vlan's over your wireless interface, are you? [18:17] Set name-type for vlan subsystem should be visible in /proc/net/vlan/config [18:18] added vlan with vid == 10 to IF -:eth1:- [18:18] Jinxed-: That's ok. [18:18] nope [18:18] That means that it created eth1.10 for you [18:18] (so it works) [18:18] Jinxed---: what is eth1 - the wired or the wireless interface? [18:19] silent1mezzo: remember to remove those supermagnets from the dead drive :D [18:20] most definitely, already got a use for them [18:22] wired [18:22] New bug: #610150 in mysql-5.1 (main) "autocompletion doesn't work" [Undecided,New] https://launchpad.net/bugs/610150 [18:23] RoyK how do I tell? [18:23] J3ckyl: why do you have an IP set on eth1? [18:23] Royk, why do I? or why do you? [18:23] J3ckyl: lshw will show you the mac address assigned to each interface iirc, and ifconfig will show the mac address [18:24] J3ckyl: sorry - wrong guy [18:26] ok [18:26] checking [18:28] Jinxed---: on most laptops I've been out for, eth1 is _usually_ the wireless card [18:28] can't be sure, though - the actual numbering is defined by udev [18:29] that is - wait - the wireless shouldn't really be ethx, it should be wlanx or something [18:29] eth0 for the wired one [18:30] but then again - check /etc/udev/rules.d/70-persistent-net.rules [18:30] ok im confused what am i looking for [18:31] Jinxed---: can you pastebin /etc/udev/rules.d/70-persistent-net.rules and the 'lshw' output? [18:32] there's another potential issue I've ran across - that the GUI network manager overrides /etc/network/interfaces [18:32] the latter isn't meant to be used on a desktop [18:33] and iirc you're running this on a desk- or laptop setup, right? [18:35] yeah [18:35] ok i got back online with my ubuntu laptop [18:35] about to send links [18:35] hello [18:35] http://paste.ubuntu.com/469435/ [18:35] Jinxed-: if you can use something else than ubuntu desktop for this, it'll help a lot [18:35] http://paste.ubuntu.com/469434/ [18:35] that is what this computer is for [18:36] but it's hard to pastebin the files not on the ubuntu laptop [18:36] Jinxed---: wlan0 and eth1 has the same mac address, so that's your wireless interface [18:36] eth0 is wired [18:37] ok, so in the /etc/network/interfaces [18:37] i should change everything that is eth1 to eth0 [18:37] * RoyK has never used VLAN tagging with linux [18:37] Jinxed---: try [18:38] Jinxed---: Yes, that would work better, I think :) [18:38] Jinxed---: looks ok [18:41] hmm [18:41] ok now when i do ifup [18:41] it kills my wired connection [18:41] guess it's time to ping [18:41] Jinxed---: pastebin the config again [18:41] <\sh> RoyK: vlan tagging is easy [18:42] \sh: not my problem - Jinxed--- is the one trying :) [18:42] \sh: Yes, it it. But not if you're trying to trunk on your wireless :) [18:43] everyone keeps telling me it's easy [18:43] im not trying to do my trunk on wireless [18:43] <\sh> Jeeves_: well, I wonder why someone wants to do that [18:43] Jinxed---: It is easy. If you understand what the hell you're doing ;) [18:43] <\sh> apt-get install vlan [18:43] <\sh> vi /etc/network/interfaces [18:43] did that [18:43] <\sh> auto vlan [18:43] did that [18:43] did that [18:43] <\sh> iface vlan inet static [18:43] \sh: Barking up the wrong tree here :) [18:43] <\sh> address bla [18:43] Jinxed---: I didn't say it's easy - just pastebin the config again [18:44] <\sh> netmask foo [18:44] ok [18:44] <\sh> gateway foobar [18:44] one sec [18:44] \sh: I guess we're just too smart ;) [18:44] <\sh> vlan_raw_device [18:44] \sh: we have the docs as well [18:44] \sh: Shall we create #ubuntu-server-experts and charge admission fees? :P [18:45] <\sh> Jeeves_: why? it's written all over the googlenet...e.g. http://www.mysidenotes.com/2007/08/17/vlan-configuration-on-ubuntu-debian/ ;) [18:45] http://paste.ubuntu.com/469438/ [18:45] <\sh> direct hit..."ubuntu vlan" [18:45] Jinxed-: still eth1 [18:46] \sh: Why? To make money, duh :P [18:46] <\sh> Jinxed-: you have to s/\-/\_/g [18:46] opps [18:46] <\sh> on all vlan_raw_devices lines [18:46] old pastebin [18:46] http://paste.ubuntu.com/469442/ [18:46] current interfaces [18:46] <\sh> then you need to check your switch that it has all allowed vlans for trunking [18:47] <\sh> Jinxed-: again - instead of _ [18:47] Jinxed-: and as \sh said, s/-/_/g [18:48] :%s/-/_/g [18:48] :þ [18:48] http://paste.ubuntu.com/469446/ [18:48] current trunk setup [18:48] <\sh> http://paste.ubuntu.com/469447/ <- thats a vlan trunk interface with source based routing (or policy based routing) [18:48] Jinxed-: did you change - to _ ? [18:49] in what file [18:49] <\sh> Jinxed-: /etc/network/interfaces [18:50] <\sh> Jinxed-: and I don't see any vlan config on your switch just the already enabled vlan 1 [18:50] <\sh> which is a default [18:51] Jinxed-: sh vl [18:52] Jinxed-: conf vl .... [18:52] http://paste.ubuntu.com/469448/ [18:53] can't see any tagged ports there [18:53] that first paste was just a detailed look at interface f0/4 [18:53] if you look [18:53] f0/1 is vlan 20 [18:53] sorry [18:53] f0/1 is vlan 10 [18:53] f0/2 is vlan 20 [18:53] f0/3 is vlan 30 [18:53] sure, we can read [18:53] but I don't see any trunk ports [18:53] :/ [18:54] the trunk port doesn't show in a show vlan [18:54] they don't? [18:54] it is f0/4 [18:54] you will notice it is missing [18:54] it is subtle [18:54] anyway - did you change - to _ in /etc/network/interfaces? [18:54] <\sh> Jinxed-: on f0/4 there is no alloweded vlan [18:55] http://paste.ubuntu.com/469446/ [18:55] Trunking VLANs Enabled: ALL [18:55] RoyK, yes [18:55] ok [18:55] <\sh> oh my f...god..what switch is that? ;) [18:55] Jinxed-: Can you copy an show running config please? [18:55] yep [18:56] <\sh> with tcpdump you can even see tagged packages btw [18:56] \sh: cisco 2960 [18:56] http://paste.ubuntu.com/469450/ [18:56] Ok [18:57] only port 1 and 4 can be trunks [18:57] <\sh> Jinxed-: you know the difference between access and trunk? [18:57] yeah [18:57] I don't think you can mix access and trunk [18:57] switchport mode access will make the port on a vlan [18:57] <\sh> Jinxed-: what you configured is native vlan 10 on f0/1 and no alloweded trunked vlans [18:57] switchport mode trunk makes the port a trunk [18:58] \sh: If you configure no vlans on a trunk, it accepts all.. Right? [18:58] eh sorry im not using f0/1 [18:58] you can ignore it [18:58] <\sh> Jeeves_: not on our switches [18:58] And indeed 'switchport mode trunk' is needed [18:59] On fa0/1 and fa0/4 [18:59] <\sh> switchport mode access [18:59] <\sh> Puts the interface (access port) into permanent nontrunking mode and negotiates to convert the link into a nontrunk link. The interface becomes a nontrunk interface regardless of whether or not the neighboring interface is a trunk interface. [18:59] <\sh> http://www.cisco.com/en/US/docs/switches/lan/catalyst2960/software/release/12.2_25_see/configuration/guide/swvlan.html#wp1200245 [18:59] <\sh> table 12-14 Layer 2 interface modes [19:00] ok [19:00] <\sh> and I do think that cisco "switchport mode access" rules over "trunk" [19:00] I just got rid of all the config on f0/1 [19:00] <\sh> Jinxed-: if you want to have a default vlan for untagged packages on a trunk port, you need to set switchport trunk native vlan [19:00] Jinxed-: your linux box is on fa0/4? [19:00] current setup: http://i.imgur.com/MJm6t.png [19:01] yeah [19:01] Jinxed-: ok, pastebin /etc/network/interfaces and 'sh run' [19:01] <\sh> Jinxed-: your config needs to be : interface fastethernet 0/1 \n switchport trunk allowed vlan 10 [19:01] <\sh> e.g. [19:02] hey guys! libpam-ldap and libnss-ldap ... are they still being looked after by the server team? [19:02] \sh, by default with cisco all vlans are allowed on a trunk line [19:02] I see zul triaged my bug earlier, just wondering who I should poke ... /me whistle innocently [19:02] not me :) [19:02] <\sh> Jinxed-: yeah I see that [19:02] if I do that only vlan10 will be allowed, (not vlan 20, 30.. etc) [19:02] should i partition using GUID or MBR (this is a raid array) [19:03] EtienneG: but yes I think they are still being maintained by us...you might want to poke mathiaz [19:03] <\sh> Jinxed-: anyways, it works here with the setup I gave you on the linux side [19:03] interfaces [19:03] http://paste.ubuntu.com/469453/ [19:04] <\sh> Jinxed-: you don't need to bring up eth0 [19:04] <\sh> via auto eth0 [19:04] show run [19:04] http://paste.ubuntu.com/469455/ [19:04] Jinxed-: what does 'ifconfig' say? [19:04] \sh, ok taking it off [19:05] ifconfig [19:05] http://paste.ubuntu.com/469456/ [19:05] <\sh> Jinxed-: and you loaded the kernel module 8021q [19:05] <\sh> Jinxed-: and installed the vlan package [19:06] \sh: shouldn't that be loaded automatically when configuring a vlan? [19:06] zul, thanks zul. Our friend mathiaz is offline :( [19:06] I did sudo modprobe 8021q [19:06] and installed the vlan package [19:06] zul, but no harm, there is no rush [19:06] Jinxed-: can you ping anything now? overthe vlans? [19:07] no [19:07] Jinxed-: lsmod | grep 8021q [19:07] i can ping from the laptop to itself, and both vlans on the switch [19:07] but not the other laptop [19:07] Hey all, can someone tell me where the apt mirror prevu uses is configured? [19:07] same for the other laptop [19:07] <\sh> RoyK: to be sure it's there I'm including it in /etc/modules all the time [19:08] 8021q 22232 0 [19:08] garp 7689 1 8021q [19:08] ok [19:09] hm. config looks right [19:09] \sh: what do you think? [19:10] (also I have both firewalls off on the laptops) [19:10] is there a way to show vlans? [19:10] yeah [19:10] as in 'sh vlan' the linux way? [19:10] do you want me to show vlans on the switch? [19:10] \sh: any idea? [19:11] Jinxed-: just pastebin all config once more - it looks ok to me, but then, others may have input [19:12] ok [19:13] network interfaces: http://paste.ubuntu.com/469435/ [19:13] computer hardware: http://paste.ubuntu.com/469434/ [19:13] trunk port f0/4 info: http://paste.ubuntu.com/469446/ [19:13] cisco show vlan: http://paste.ubuntu.com/469448/ [19:13] /etc/network/interfaces: http://paste.ubuntu.com/469453/ [19:14] cisco show run: http://paste.ubuntu.com/469455/ [19:14] Jinxed-: Access Mode VLAN: 1 (default) on f0/4 [19:14] ifconfig: http://paste.ubuntu.com/469456/ [19:14] Jinxed-: what happens if you ifdown eth0? do all vlan devs go down? [19:15] or ifconfig eth0 down [19:15] all the vlans go down as in when i type ifconfig [19:15] they aren't there any more [19:16] cisco switch has link lights still [19:17] ok [19:17] just curious [19:17] probably not the right thing to do, then :) [19:18] ok, so I can ping from my laptop that is 10.1.10.11 to the vlan20 which is 10.1.20.1 [19:18] Jinxed-: switchport trunk encapsulation dot1q [19:18] tried that? [19:18] sorry laptop was 10.1.20.11 [19:19] not directly (not possible), but i double checked to make sure it was configured for 802.1q [19:20] ok, so i can ping from the laptop (10.1.20.11) to the vlan20 on ubuntulaptop (10.1.20.1), but I can't hit vlan30 (10.1.30.1) on the ubuntu laptop [19:20] Jinxed-: show version [19:20] on the cisco [19:20] so ubuntu isn't forwarding the packets [19:20] netstat -rn [19:21] Cisco IOS Software, C2960 Software (C2960-LANBASEK9-M), Version 12.2(44)SE6, RELEASE SOFTWARE (fc1) [19:22] Jinxed-: ok, so from the trunk, you can ping machines at boths VLANs? [19:22] netstat: http://paste.ubuntu.com/469465/ [19:23] yes [19:23] from ubuntu laptop i can ping both external laptops on vlan 20,30 [19:24] i can also ping the ip addresses of both vlans on the cisco device [19:24] sysctl net.ipv4.ip_forward [19:25] net.ipv4.ip_forward = 1 [19:25] so both the ubuntu laptop and switch can hit everything [19:27] check sysctl -a | grep forwa [19:27] see if packets are forwarded at the device given [19:28] _mc is just multicast - normally not needed [19:29] looks like they are mostly forwarded [19:29] Jinxed-: also, if this is on a private net (rfc1918 addresses) you will ned NAT to reach the internet [19:29] or anything != RFC 1918 [19:30] RoyK, i don't want to get online [19:30] !nat [19:30] (at least yet) [19:30] can you ping from the client to the server you're on? [19:31] not sure what you mean [19:31] explain where ping works and where it doesn't work [19:32] From the external laptops (vlan 20/30) I can ping itself, and BOTH vlan on the switch, and the associated vlan on the ubuntu-laptop(server) [19:33] I am having a problem getting the /etc/init.d/apache2 script to recognize my User= and Group= directive within my apache2.conf, so the web server continues to run as root, which is certainly not what i want. am I overlooking something? [19:33] RoyK, so in this picture [19:33] Jinxed-: and I guess you can ping the 'external' laptop from those other ones? === dendrobates is now known as dendro-afk [19:34] inveratulo: the initial apache server will always run as root [19:34] inveratulo: without that, it can't open port 80 [19:34] inveratulo: but then, all child processes are changed to the apache user [19:34] RoyK: that's fine, but the damon should fork right [19:34] Laptop 1 can ping: switch (10.1.20.254, 10.1.30.254), Ubuntu-lap vlan20(10.1.20.1). It CANT ping 10.1.30.1 or 10.1.30.11 [19:35] The switch/ubuntup laptop can ping everything [19:35] Jinxed-: then what is it the other laptops can't ping? [19:36] laptop 2 (10.1.30.11) CAN'T ping 10.1.20.11 or 10.1.20.1 [19:36] so the same things [19:37] and how is the routing table on laptop 2? [19:39] how do i display that [19:40] netstat -rn [19:41] eh ok? [19:41] i can't copy/paste the laptops aren't online [19:42] well, is that linux box their default router/gateway? [19:42] or do they have static routes to the other networks they want to access? [19:42] or is RIP or OSPF enabled somewhere? [19:44] Jinxed-: any box on IP will try to reach its networks through its default gateway - if the gateway doesn't know the network, it'll send back an ICMP network unreachable [19:44] i just made the vlan they were connected to on the switch their default gateway [19:44] is the linux box their default gateway or the switch? [19:44] the switch is a layer two switch, not a router [19:45] WWWWWWWWWWWWWWWWWWOOOOOOOOOOOOOOOOOAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAHHHHHHHHHHHHHHHHHHHHHHHHHH [19:45] works! [19:46] you sparked a thought [19:46] i just changed the laptops default gateway to the interfaces on the ubuntu laptop instead of the switch [19:47] Jinxed-: http://en.wikipedia.org/wiki/OSI_protocol [19:47] start there [19:47] a switch works at layer two, not three [19:47] haha [19:47] that is, unless you have a L3 switch, which costs a wee bit more [19:47] not this one, although i do have some [19:47] that i could have used [19:48] this was pure l2 [19:49] now to try to stream videos [19:49] with xbmc [19:49] :) [19:49] thank you thank you thank you [19:49] multicast? or unicast? [19:50] RoyK, /sh, and Jeeves_ and everyone who helped [19:50] I want to set it up so the laptops(1 and 2) could request a video from the ubuntu laptop [19:50] and stream it [19:50] well [19:50] multicast or unicast? [19:50] well its requested, so it would be unicast [19:50] ok [19:51] if doing multicast, check the IGMP settings [19:51] AFAIK most switches come with IGMP disabled [19:51] multicast would be like if I decided on the ubuntu laptop that i was going to stream [19:51] video [19:51] and started a multicast stream [19:52] multicast is used if streaming live video on a LAN [19:52] for VoD, unicast is what you'd use [19:52] for live video for < 10 machines, unicast will do ok [19:53] RoyK, have you ever done VoD [19:53] large-scale, yes [19:53] Noice [19:53] 100 concurrent views [19:53] nice* [19:53] how do you do that [19:54] from a couple of cheap boxes with some cheap ATA drives [19:54] what software did you use for the vod part [19:54] we wrote a video streaming server on top of Linux RAID-0 (with sufficient servers for failover) [19:55] just make the chunk size large enough so that the disks won't spend time seeking [19:56] you wrote a video streaming server? [19:56] you can do it with RAID-5 too if you want to - just keep lots of processes reading and have sufficient memory for the read-ahead - linux can do most of this alone [19:56] yeah [19:56] I was thinking about trying to use ssd with xbmc [19:56] I would be very interested in your work [19:57] sounds very similar to what I want to do [19:57] Jinxed-: it's simple - the problem with VoD is concurrency - lots of clients wanting to read at different places on different drives. [19:58] i take it ssd would help, but not solve the problem? [19:58] the solution to this is to have enough memory to use LARGE chunk sizes for Linux software RAID, and clients that can fail over to another server if the one they're on fails, which of course implies a good buffer on the client side [19:58] forget about the SSD [19:59] we wrote a system on which we had four cheap servers with some (at the time) large ATA drives in Linux software RAID-0 [19:59] chunk size 1MB, meaning mostly no seeking, mostly just reading linearly [20:00] the software was simple, an HTTP server [20:00] we tried with Apache but got hung up with memory issues, probably fixed by now (this was around 2003) [20:01] anyway - with a cheap box with four 120GB drives, we could sustain >80 clients, each watching a movie with a bandwidth of 4Mbps [20:02] so then you also had file sharing capabilities? [20:02] we tried at the (current) Compaq lab, and with a truckload of fast SCSI drives, we could do about 100 concurrent reads [20:02] that wans't in the picture [20:02] nice [20:02] hmm [20:02] I want to do vod [20:02] file sharing [20:02] and eventually play with asterisk [20:03] in that order [20:03] just sit down and do some calculation - with today's SATA drives, you get something like 100-120 IOPS [20:03] meaning I/O operations per second [20:03] the problem is seeking [20:03] so if you read a LOT before seeking, even $120 2TB drives are blazingly fast [20:04] we just had 120gig drives back then, today it'll be far faster [20:04] would ssd eliminate seek time [20:05] but then, you'll want a client that can read an XML file with a randomized list of available servers from which to fetch the video [20:05] random access - no seek motion [20:05] sure, but SSDs cost a LOT [20:05] yeah [20:05] perhaps ZFS with SSDs in front will help [20:05] well, it will help [20:06] I have a couple of 50TB setups on ZFS on opensolaris with SSD caching and they perform better than we need them to [20:06] haha, i think that is larger than what i need for space [20:07] well, to cut it short - most filesystems aren't made for streaming [20:08] that's why we did this, seven years ago [20:08] if you want to push the limits, go low [20:08] Do you see any major disadvantages to using [20:08] if not, use ZFS, it'll work well [20:08] xbmc [20:08] or some other software that is designed to stream videos? [20:08] I don't know that product [20:08] xbox media center [20:09] IMHO it's not the streaming software that's the problem, but the filesystem and/or storing architecture [20:09] xbmc is just a gui wrapped around mplayer -- it doesn't usually do "streaming" [20:09] Linux doesn't have anything that can compare with ZFS [20:09] it was originally developed for xbox, it now is a cross platform system that has the ability to do vod afaik [20:09] btrfs will be there some day [20:10] giovani: some day, yes, I've been following the progress for more than a year, but it still doesn't have anything like raidz [20:10] giovani: and no SSD caching [20:10] Jinxed-: how much data will you be serving? [20:10] RoyK: absolutely -- and ZFS didn't appear out of nowhere -- give it time :) [20:11] giovani: sure, I'll give it time, but meanwhile I'll stick with ZFS [20:11] are you running it on Solaris, or FreeBSD? [20:11] Well right now, not too much experimenting to try to get it to work [20:11] giovani: opensolaris [20:11] i don't even understand what zfs is [20:11] Jinxed-: how much? 1TB 10TB? [20:12] Jinxed-: http://hub.opensolaris.org/bin/download/Community+Group+zfs/docs/zfslast.pdf [20:12] you mean the overall size of the video [20:12] Jinxed-: scan through that presentation [20:12] the videos that would reside on the server [20:12] or the overall amount of data going out at one time [20:12] Jinxed-: data size - the total [20:12] 1 tb would be much closer [20:12] maybe 300-600 gb [20:13] you can do that on any system, really [20:13] down the road, possibly more [20:13] I would like to make it so I could add on if I wanted to [20:13] how many concurrent clients? [20:13] 1-50 [20:13] most likely about 4 [20:14] any system will handle that [20:14] anyway [20:14] :/ [20:14] [21:12] Jinxed-: http://hub.opensolaris.org/bin/download/Community+Group+zfs/docs/zfslast.pdf [20:14] read that, and you'll understand a few new things about storage [20:16] I am having a problem getting the /etc/init.d/apache2 script to recognize my User and Group directive within my apache2.conf, so the web server continues to run as root, which is certainly not what i want. [20:17] RoyK, looks very interesting, i don't think I have ennough background to fully appreciate it though. Im new to the whole server thing [20:19] downloading/printing/googline [20:20] googling* [20:23] RoyK, what do the additional server components of ubuntu get you if you install them? === dendro-afk is now known as dendrobates [20:51] New bug: #609743 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/609743 [21:25] RoyK, so I restarted and now when i do ifconfi [21:25] ifconfig [21:25] i don't see anything [21:25] how do i get everything back up and running [21:25] ? [21:28] whats jot running jinxed [21:29] ssureshot, so I had vlan trunking working earlier [21:29] i have vlan20,30 etc set up in the /etc/network/interface [21:30] but it doesn't show up when i do ifconfig [21:30] infact eth0 seems not to be working at all [21:41] yo Daviey, still awake ;-) [21:41] ? [21:42] hggdh, sadly :) [21:42] Jinxed-: is the 8021q module loaded? [21:42] Daviey: I figured so... I see groovy errors in the cloud-debug.log [21:43] hggdh, Interesting.... can you pastebin? [21:43] Daviey: just a sec [21:43] hggdh, Ok.. i'm not fully here.. just passing - but i will read it! :) [21:45] Daviey: I frankly do no texpect you to be fully awake at this hour ;-) http://pastebin.ubuntu.com/469518/ [21:47] Daviey, do you have problems with recent iso having syslinux crash due to low dos memory? [21:51] Daviey, er well the syslinux on recent iso when used with pxe crashes for me i meant to say === A-Tuin_ is now known as A-Tuin [21:58] 22:58:15 up 3 days, 7:57, 2 users, load average: 24.26, 24.20, 24.12 [21:59] hggdh: Ahh.. i think i know what has caused that groovy issue.. it's an upstream bug if my prediction is correct. [21:59] (introduced by refactoring) [21:59] ccheney: Interesting... not seen that yet [21:59] Daviey, ok [21:59] will experiment tommorrow [22:00] RoyK: Nice. [22:00] all 24 cores running stably [22:00] close to no system time or wio [22:01] RoyK: My record to date (obtained over ssh) http://daviey.mooo.com/wp-content/uploads/2007/10/top.png [22:02] Daviey: http://karlsbakk.net/top-24.png [22:05] niiiice [22:05] we have another 16 core working on that too [22:06] Eyjafjallajökull ash movement inversions [22:17] ccheney: which package is implementing your uec provisioning? [22:17] hggdh, uec-provisioning-* [22:17] heh [22:18] ccheney: already in Maverick? [22:18] yes [22:18] k, thanks [22:45] hey is it possable to setup email accounts to auto cc/bcc to another address ? I want the emails to stay in the correct box but I would like to have it also sent to the main email of the person in charge of that so if that person leaves we still have all the old emails [22:46] Hi guys [22:46] I have an emergency here [22:46] lwizardl: always_bcc? [22:46] yeah I want it to always do it [22:46] I add to change sda in a raid1 array, so I hot removed it with mdadm, shutdown, chnaged the disk [22:47] buit it won' boot on sdba [22:47] sdb [22:47] I only get a blinking cursor [22:48] I had a look to sdb using live cd and boot flag is correctly set on the boot partition (sdb1) [22:48] lox_: wait for it [22:48] lox_: it will bring up busybox [22:49] is it supposed to be so long ? [22:49] couple of minutes [22:49] i think it's 5 [22:49] Ok so I wait more .... [22:49] Just a ping on my previous question, can someone tell me where the apt mirror prevu uses is configured? [22:50] ivoks, thks [22:50] ivoks, basically I am hosting a site for a convention and the staff members tend to be replaced. example if sam is incharge of costumes and advertising, I would like to have costumes@ & advertisment@ bcc a copy of all emails sent to those emails be sent to her main address but also keep a copy of those in there original email box also. so if she quits we don't have her personal emails mixed in with the convention emails [22:50] always_bcc is for all mails sent and received [22:50] so, i guess that's not an option [22:51] you could setup procmail as a delivery agent [22:51] and have /etc/procmailrc [22:51] ivoks, because we originally thought about just forwarding but then they don't stay in the main boxes [22:52] they can [22:52] you can forward a copy [22:53] easiest thing to do would be with /etc/aliases [22:53] for example, if you have user steve getting the mail [22:53] and you want forward a copy to mark@gmail.com [22:53] you would add: [22:54] steve: steve mark@gmail.com [22:54] to /etc/aliases [22:54] ok [22:55] if you use postfix [22:55] I thought that it would never reach a box for the other emails and just push the mails to the new location [22:55] i'm not sure with other mtas [22:59] ivoks, the cursor has been blinking for ten minutes, still nothing else [23:01] Oh, I've figured it out. /usr/bin/prevu-init is hardcoded to override the mirror in my /etc/pbuilderrc with archive.ubuntu.com. That's terrible. [23:01] Guess I'll file a bug report shortly. [23:03] lox_: alt+f1, alt+f6? [23:03] ivoks, no go ... [23:04] ivoks, I had replace sdb the same way with no problem [23:04] ivoks, is the blinking cursor from grub ? [23:04] did you see grub menu? [23:10] ccheney: this is really something, it seems euca-* commands output changed on 2.0 [23:11] ccheney: and I am getting 100% failure now :-( [23:31] okay, i made a raid array with mdadm; i'm trying to get LVM on it. i do a: "pvcreate /dev/md0" and it returns: "Device /dev/md0 not found (or ignored by filtering)" i checked "ls -l /dev/md0" and it's there. it's my raid array. has anyone run into this before? am i doing something wrong? [23:33] AndyGraybeal: check /etc/lvm/lvm.conf, search for 'filter' [23:33] hggdh: Hmm.. something seems inconsistent - i had that stage last week... left it, and it worked again. [23:34] ivoks: it says: filter = [ "a/.*/" ] [23:35] if that's the only filter, then it's not filter at all :D [23:35] ivoks, then that' sthe filter - do yoiu know what might be happening? [23:36] no idea [23:36] Daviey: this is really not kosher [23:36] okay thank you ivoks. [23:36] /proc/mdstat [23:36] is that ok? [23:36] Daviey: my euca-run-instances stay in pending for a while, then go to tesminated [23:36] hggdh: *sigh*, that *was* working [23:37] hggdh: Feel free to keep prodding, can you email me your findings? I'm going afk now :( [23:37] Daviey: I am opening bugs on all I find ;-) [23:38] rocking! [23:38] we will chat tomorrow, go hit the bed, don't go my way last week ;-) [23:42] ivoks, no grub menu appeared [23:42] lox_: then it doesn't boot from disk at all? [23:46] ivoks: i decided i run the installer again before you wrote that, sory [23:46] haha [23:47] ivoks: the installer says the same thing about /dev/md0 ... even after i deleted and recreated the array [23:47] so, in previous life, you were a windows sysadmin? :) [23:47] ivoks, yes, yes, 10 yrs ago. [23:47] i'm still fumbling with this linux. [23:47] ivoks, i'll take your queues next time though, sorry. [23:48] you would have erase superblock before creating new md [23:48] anyway - my goal is to end up with 1 80gb boot drive with linux on it and one raid array with 4 disks, 320 each. [23:48] Daviey: apparmour, it seems [23:49] ivoks: okay - i don't know what really that means, erase the superblock, does that require a reboot? [23:49] there's nothing on the disks? [23:49] hggdh: Not suprised by that - it's bitten us a few times [23:49] ivoks, no there's nothing on the disks.. i'm just starting from scratch. [23:49] i got all my data backed up safely [23:49] AndyGraybeal: and you are in the installer? [23:49] Daviey: yes... for the record, bug 610265 [23:49] Launchpad bug 610265 in eucalyptus "cannot successfully start an instance" [Undecided,New] https://launchpad.net/bugs/610265 [23:49] ivoks: yes, it's unpacking softwares as we speak, i can restart it if you recommend. [23:50] AndyGraybeal: er... didn't you just said that it doesn't work? [23:50] i'm very confused and scared a bit by this whole manual partitioning. i've been doing this for 10 years, but that doesn't mean i'm nearly anywhere good at it. [23:50] oh, you have one partition just for / [23:50] one disk [23:51] ivoks, the plan is one disk holds all of linux and assorted softwares; but the raid array holds /home and /srv (and data for my business) [23:51] ok [23:51] then finish installation [23:52] sory for the confusion and run around. i feel like i'm never going to get good at this ... in 10 yrs.. it's been longer than tear years honestly; but we won't get into that. [23:53] i gave up a high paying job with windows to take the time to really learn linux.. and here i am :) still learning. [23:54] but it's good, it's in a real business setting. [23:54] i feel more in control that's for sure. [23:55] what ubuntu is that? [23:55] which version [23:55] it's 10.04 server [23:55] 64bit [23:55] ok [23:59] okay, it's still moving, i'm going to get a bite to eat; brb