=== shade__ is now known as \shade\ [00:13] http://www.boldvoices.tv meme went live on channel 5 ten minutes ago [01:08] Hi all: I'm installing a ubuntu server without any graphics devices (xen domu, using the xm console method). How do I add users to the system? (when I tried using the traditional "useradd" program, the account was pretty botched; the only way I've found so far to add user accounts was through the ubuntu gui if I wanted them to work) [01:09] kusznir, use 'adduser' [01:09] it sets all the things right for the users to work [01:10] Ok, thanks. [01:50] hello, I've been researching pci passthrough to enable NIC passthrough to one of my VM's and believe it requires vt-d support from my motherboard. I discovered that there is no vt-d support on my motherboard and am wondering if there is a way to do PCI (specifically NIC) passthrough without vt-d? the reason I am investigating this is I only have 1 static public facing IP to assign to a webserver I would like to run [01:54] hi all [01:58] I just changed my IP on my ubuntu-server and I get the following error... "apache2: Could bo reliably determine the server's fully qualified domain name, using 208.83.212.218 for ServerName" any idea of what I broke?? [02:45] Yosi, you need to set up your hosts file correctly [02:45] it'll work as-is, but if you want the error to go away, the domain name you're using in apache needs to resolve to your IP [02:55] qman, thanks. i think that is whats going on...some guy in #linux was saying the issue was my httpd.conf file, but in ubuntu that file is empty [02:56] qman: to fix th error, what file in apache would contain that domain name? [02:58] hi, I am unable to install/remove/upgrade mysql-server <-> http://pastebin.com/9pQjQivc (Ubuntu 10.04) It crashed/locked up during an upgrade process and since then aptitude fails on everything [03:01] hi all, I just have a quick q. im looking into starting and setting up a personal apache server on my 10.04 desktop. Does anyone have any advice or good guides for a beginner? [03:05] tomswartz07: if you already have the comp installed, and you just want apache on there [03:06] Yosi, /etc/apache2/sites-available/[yoursiteconfig] [03:07] amstan: I have 10.04 desktop installed, plain vanilla. ive found some sites hosted from ubuntu, but no decent guides for beginners. [03:07] a plain jane web site server is fine by me [03:08] what other services could you run from a home server like this? [03:08] tomswartz07: just install apache2 libapache2-mod-php5 [03:08] tomswartz07: umm.. you probably want mysql too, so install that, and php5-mysql [03:08] you can do all of this with synaptic [03:08] when you're done you put your website in /var/www [03:09] amstan: and thats it? what does mysql do, exactly? [03:10] tomswartz07: ok... hmm [03:10] Yosi: the value for ServerName needs to resolve to an address. set up /etc/hosts accordingly [03:10] tomswartz07: what exactly do you want from this website? [03:10] tomswartz07: if it's for a personal website, and you're just playing with html, then you don't need mysql and maybe not even php [03:11] tomswartz07: so.. apache is the server, php a scripting language that makes all the pages dynamic [03:11] just a place to host a page or two, even better if i could have file hosting. [03:11] tomswartz07: and mysql is a place to put all the data that php has to play with, stuff like user accounts and so on, it's easier to use mysql for some stuff than files, but you need to program that retrieval [03:13] tomswartz07: i'm not exactly sure where i can get you some guides about this, i kind of passed this stage already [03:13] if you have any questions just ask me [03:13] amstan: ah, i see. Thats a bit too rich for my blood just yet. i dont think ill be having users and logins, etc. Would just plain apache be right for my needs? [03:13] im looking to just do basic hosting, web pages and files [03:14] tomswartz07: if you don't want to touch programming with php, just plain apache is fine [03:14] tomswartz07: apache will provide all the things that makes your computer serve a static page/file [03:15] amstan: excellent. Ill see what else I could dig up. I just didnt know what to look for [03:16] amstan: thanks for the help! ill give a shout if i have anything else! :D [03:16] tomswartz07: you could try some html tutorials [03:16] tomswartz07: if you you're not already familiar with it [03:16] i have the html down pat, i just dont know how to get it hosted myself for free haha [03:17] tomswartz07: did you ever port forward/ do you know what that means? [03:17] this is a waste of time distro. [03:18] amstan: yeah, i use it for ssh pretty often. i read that apache just uses port 80 [03:18] tomswartz07: yep, you got a domain too, right? [03:18] no-ip at least? [03:19] through dyndns, i have the router set up to update with the IP, if thats what you mean [03:19] yep [03:19] tomswartz07: well.. you're pretty set then [03:20] amstan: woot woot. thanks again! ill see what i could do! [03:35] mathiaz: ping can you review the wording and make suggestion, this is the email being sent out for the sru process http://paste.ubuntu.com/451416/ [03:41] how can I fix this file error? "dovecot: IMAP(scott@moseley.ca): Invalid data in file /mnt/raid/www_root/mail/moseley.ca/scott/.Sent/dovecot-uidlist" [03:49] pmatulis: I just checked /etc/hosts and there are two entries there, the first is 127.0.0.1 localhost and the seond line is 218.83.212.218 cs1 (cs1 is the host name of the machine) [03:51] Yosi, you need the FQDN also on that second line [03:51] hey all, might not be in the right place for this, but let me know. i am interested in setting up a linux box in my house to do dns, authentication (ldap), posibly NAS, etc. Ideally, want to run on a small atom machine. Any easy to use Small Business distros you knwo of? maybe one with a nice web UI? [03:51] based on ubuntu [03:52] qman: would the fqdn not just be the hostname, but the hostname.domain [03:52] ? [03:52] that is what an FQDN is [03:53] qman: webmin put that in there.. its so odd... [03:54] !webmin [03:54] webmin is no longer supported in Debian and Ubuntu. It is not compatible with the way that Ubuntu packages handle configuration files, and is likely to cause unexpected issues with your system. See !ebox instead. [03:57] qman: would having a fqdn in there fix my apache2 error? [03:58] yes [03:59] as long as it's the FQDN used in the apache configuration [03:59] ServerName needs to be resolvable [03:59] whatever that happens to be [04:00] qman: i looked in the directory u mentioned in apache2, but i didn't see a ServerName, all I saw was default [04:00] Yosi, default is a configuration file [04:00] ServerName is specified inside it [04:03] qman: I just checked, I' read the file, and there is no ServerName field specified in the file just says VirtuaHost *.80 [04:08] Yosi, and idea on my dovecot issue? [04:15] exit [04:16] zul: line4: add the list of SRU criteria [04:59] having qemu/kvm issues on 10.04, specifically, error: unable to connect to '/var/run/libvirt/libvirt-sock': Connection refused === macd_eee is now known as macd [05:01] macd: ls -l that file [05:02] srwxrwxrwx, I set it 777 to test, added root user to libvirtd [05:03] OK, first undo every "test" you did, then try again. [05:05] Sure, and I get the same results from virsh -c qemu:///system list (FYI adding the root user to the libvirtd group is in the qemu documentation) [05:07] Now, tell me the full line for ls -l on that file. [05:08] I'm surprised the *qemu* documentation mentions libvirt at all. [05:08] srw-r-x--- 1 root libvirtd 0 2010-06-17 22:36 /var/run/libvirt/libvirt-sock [05:08] I should have been more specific, the ubuntu/qemu documentation [05:08] macd: URL> [05:09] https://help.ubuntu.com/community/KVM/Installation [05:10] Well, first of all "community" indicates that it's not Ubuntu's documentation, it's Ubuntu users' documentation -- which can be outright wrong. [05:10] Im thinking we could focus on the problem at hand rather than the documentation that led me there, I'll remove root from the group [05:11] I'm surprised root need to be in any group, since root generally circumvents posix DACs entirely. [05:11] And logout/in, reissuing the above command virsh command yields the same results [05:12] I tend to agree at adding root to anything, it should superseed [05:12] Are you running things as root, or as a normal user? [05:12] there is no selinux either so it is definetly DAC [05:12] I am running as root. [05:12] Any reason you're running as root? [05:13] I like it. [05:13] I could login and sudo instead, I dont really see the difference [05:14] My understanding is that a large part of libvirt's job is to allow non-root users to use VMs (which require privileged access to e.g. brctl). [05:15] If the normal way of using libvirt (and Ubuntu) is to not run stuff directly as root, I suggest trying that. [05:18] I get the same results using my user, with or without sudo. [05:18] And that user was in the libvirtd group when you logged in? [05:20] added, logout/in, same result [05:29] OK, at this point I would be checking the log files and asking #virt [05:30] yeah, Im querying in @qemu as well, I appreciate your help. [05:39] twb, Jun 17 23:37:11 spare5 kernel: [ 3667.415428] kvm: disabled by bios [05:39] I thnk I figured it out ;) [05:42] Grmph. I'd have hoped it'd just use software virtualization in that case. [05:45] Im under the impression it only supports full virtualization, and not para [05:47] Well, qemu sure as shit supports software virtualization, and if Ubuntu has dropped that in it's combined qemu/kvm package, I will be Pissed Off. [06:42] New bug: #595789 in lm-sensors (universe) "after resuming from suspend, sensors -s is required" [Undecided,New] https://launchpad.net/bugs/595789 === jo-erlend__ is now known as jo-erlend [07:07] anyone using 8.04 server as kvm virtualisation host? [07:07] i can't connect guest os after install [07:08] i installed with ubuntu-vm-builder [09:00] hello, why does sudo aptitude install kvm libvirt-bin ubuntu-vm-builder bridge-utils on a fresh lucid server install [09:00] tries to install x11-common ? [09:01] no idea - some dependencies - probably some X-stuff in there somewhere - why? does it matter with a few extra libs? [09:06] lau: x11-common is just a few boring files, it's not anything significant [09:06] Most likely it is there for keyboard support. [09:09] * Ng drumrolls [09:09] monring [09:10] kirkland: I have good news and bad news. Actually I just have bad news, that package doesn't fix it, three of my four test VMs have been OOM killed. I'll comment as such on the bug. [09:10] kirkland: (at least it doesn't fix the leak I'm exposing, maybe it fixes others) [10:35] twb: thx [10:51] just create new vm sudo ubuntu-vm-builder kvm lucid [10:52] want to remove it from the system, is rm -rf ubuntu-kvm the proper way ? [10:52] or does it exist a ubuntu-vm-builder erase kvm lucid ? [10:56] lau: did you check the manpage for ubuntu-vm-builder? [11:02] twb: yes man vmbuilder but I only saw references to create a vm not to delete it (maybe I missed it?) [11:04] I don't know, I'm just triaging [11:05] twb: I also took a look at https://help.ubuntu.com/community/JeOSVMBuilder but without success [11:49] lau: Unless you used the --libvirt option, the VM is just the disk image. Delete the disk image, and the VM is no more. [12:34] hello to all. i have a small question. i am running on two ubuntu servers the apache webservice however every server has it own website. when i call the Website of the second server i get allways site not found. the domain resolves great but somehow the apache server dont return the page. i must say that on the second server i have configured thecan somebody help me to debug the problem ? site as vhost maybe that is the problem ? [12:35] i must say that on the second server i have configured the site as vhost maybe that is the problem ? can somebody help me to debug the problem ? [12:36] xperia: is apache running? [12:36] xperia: is it bound to port 80? What happens when you browse from the local host (e.g. with w3m)? [12:36] yes it is running but i check it again over the local ip adress [12:37] twb: when i call the server over the local ip adress it return "It Works" [12:38] Therefore you most likely have a firewall, routing or name resolution issue. [12:38] so apache is working. will check if the vhost is configured right. the thing is both servers are behind one static ip and on both servers bind is running great one time as master and one time as slave on the second server [12:39] hmmm will check that with the firewall [12:39] Um, it would be exceedingly peculiar for a single (ip, port) tuple to be connected to multiple hosts. [12:40] port 80 is open for the second server too in the firewall. [12:40] A network diagram showing where your masquerading, DNATs, mod-proxying and load-balancing happen would be useful. [12:43] twb: thank you a lot for your help. i have a router and behind this router the two servers are located. the router has port 80 open to both servers. if i do "dig secondserver.example.com @wanip" it resolve the domain great also if i do "dig secondserver.example.com @lanipfirstserver" and "dig secondserver.example.com @lanipsecondserver" [12:43] so bind runs on both servers great and i ca browse the sites on the first server with no problems [12:44] just the second new added server dont display the sites even the hostdomain resolve great ! [12:45] a diagram would look like this [12:45] wan ip adress static | router | 2 servers [12:45] at the moment only the first server works great [12:52] If i run a reverse webserver proxy, it means that everything will be loaded from the box behind proxy, right? Is there a purpose of having documentroot for a proxy host? [12:53] Error404NotFound, you don't necessarily have to proxy _everything_ on a particular host [12:54] i.e., your static files could be served by apache [12:54] cwillu_at_work, so i can actually load some of the files off the reverse-proxy webserve [12:54] cwillu_at_work, great... [12:55] cwillu_at_work, also if i am proxying a LAMP, do i need MySQL and PHP on the reverse proxy host? even if i am not serving any PHP content on that? [13:00] Error404NotFound, lamp is a marketing term :p [13:00] Error404NotFound, you should only ever install what you need [13:00] if you don't need mysql and php on a box, don't install it [13:01] cwillu_at_work, Nginx could be a possible alternate to Apache + mod_proxy + mod_proxy_html + mod_rewrite? [13:01] I suppose so [13:01] I tend to avoid the use of mod_rewrite myself [13:19] In lucid, gdm's default greeter has the same background as the default GNOME desktop, a bar along the bottom that looks like gnome-panel, and a login dialog in the middle that looks rather similar to the old gnome 2.10-ish "loading splash" -- especially since the list of users was empty (apparently it can't/doesn't enumerate LDAP accounts by default). [13:19] This was sufficiently confusing that it took me four reboots and a co-workers' help to realize that the system wasn't crashing in the middle of the casper "ubuntu" user's login -- it was actually the GDM theme. [13:30] Daviey: just popping in.. any questions on the groovy fix proposed by Chris ? [13:31] Daviey: looks like s/groovy/groovy-all/ in eucalyptus-java-common.links might fix it [13:31] ttx: It's not looking like the fix tbh.. I manually did the symlinks, and didn't produce joy [13:31] Just built a package wth it done in the packaging, not expecting it to work [13:31] but thought i should try [13:31] ttx: In other news, removing the battery in the laptop made the NIC start working again :/ [13:32] trying to run with groovy-all in the classpath instead of groovy doesn't fix it ? [13:33] Daviey: about the laptop -- hopefully just a tarnisient issue [13:33] transient, even [13:33] ttx: Yeah, i'm investigating groovy issue as we speak [13:34] ttx: Not currently blocked on anything.. you should be on holiday :) [13:34] hello to all. how do you setup two web servers behind one wan ip to work right ? [13:34] ok, I'll stay around for a few while reading up email -- just in case you have questions [13:34] ttx: thanks [13:35] what is the best way running two servers behind one wan ip ? [13:37] morning [13:38] Xperia [13:38] good evening :-) [13:38] yes i am listening [13:38] I'm no pro [13:38] me either :-) [13:38] But if you are trying to set up 2 servers for 2 fifferent things [13:39] You can use port forwarding on your router [13:39] to send things like mail to one server and webvisits to the other [13:39] is this what you are after or have i misenturpreted [13:39] well i am running webservers that listen on port 80 both [13:39] okies [13:39] so i have two webservers behind a router with a static wan ip [13:40] ok [13:40] and need now to get the second server to run but somehow it wont [13:40] ok [13:40] so all traffic is going to server number 1? [13:41] well in my router i opened port 80 to both servers [13:41] so both servers should be able to answer [13:41] hrmm [13:41] tricky ;/ [13:41] but at the moment only the first delever the page when i call up the hostdomain of the second server [13:42] i just get the wrong page for a host domain [13:42] http://forums.kayako.com/f74/2-web-servers-1-public-ip-address-iis-17558/ [13:42] that should be delevered by the second server [13:43] Not sure if that will help [13:45] well the keyword for resolving that is as i have readed from your post "NameVirtualHost" need to study that more. thank you a lot for the help. your post descibe my problem exactly [13:48] xperia, forwarding 80 to two different boxes won't work [13:49] why not if i may ask ? [13:49] normally it should work with no prob when vhost works [13:50] vhost is different than this though [13:50] with vhosting, you've got one server answering to two different domain names [13:51] in order for your router to figure out which server it's supposed to send the packets to, it would have to provide exactly that [13:51] setting up your server to forward one vhost to the other server is probably the best way to do it, if you want port 80 to work for both addresses (and assuming your router is fairly simple) [13:52] hmmm so if i understand you right you say me the router will probably never send any request to the second router [13:52] to the second server, right. [13:52] or it might just randomly choose :p [13:53] I'm trying to find out the differences between the server and generic kernels, but I'm getting a 404 on http://www.ubuntu.com/products/whatisubuntu/serveredition/features/kernel, any ideas? [13:53] yeahh that could be the problem. then i need in my bind server to resolve to names to the wanip adress with a specific port like wanip:8081 and then redirect all trafic for this port to the second server in the router [13:54] you can't do that though [13:54] dns has no concept of ports [13:54] i.e.: xperia.com will resolve to your ip address, but there's no way you can also include a port in that [13:54] really. i have read that bind can resolve hostnames with port number [13:55] if that really is not possible with bind how can i run then two servers behind one static ip ? [13:55] now, a different approach would be to have a vhost on server A that simply redirects to xperia.com:81 [13:55] xperia, you use a vhost, and have one server proxy through to the other [13:56] ProxyPass / http://127.0.0.10:8000/ min=10 max=20 retry=0 [13:56] ProxyPassReverse / http://127.0.0.10:8000/ [13:56] yeah this i have allready also thinked but redirects arent really good [13:56] would be the apache for it, (substituting the address as necessary) [13:56] the other approach would be to use a redirect [13:57] even so, you'll find at least some people won't be able to access the server on the non-standard port [13:57] as many firewalls / proxies block such access [13:57] hmmmm did not thinked that such a small problem would need such strange fixes like redirects :-) [13:58] I wouldn't recommend the redirect approach, as I said [13:58] the right solution is the reverse proxy forwarding [13:58] there are smaller servers that are dedicated to such use, although apache will work fine as well [13:59] okay reverse proxy forwarding is a keaword that i need to look further [14:00] i am asking really how does the big hosting companys solve that problem [14:00] note that this implies all of the traffic from the second server will go through the first server as well, but that's pretty much unavoidable [14:00] xperia, they have loadbalancers which basically do exactly this [14:01] The key words are "high availability" and "load balancing". [14:01] they have hundreds of physical servers behind one ip and resolve thousend of hostdomains [14:01] yep [14:01] Daviey: I'm going back to my vacation [14:01] But since you aren't a large corporation yourself, many of the things you find will be out of your experience/budget/feasibility range [14:01] and they have several servers that are largely identical [14:01] well thank you a lot for the info need to look really more into that [14:02] FSVO largely = blade centers :-) [14:03] myself, I just use racks and racks of beagles :) [14:06] I don't have a hard drive. I just keep twelve thousand Indian teenagers in my basement and force them to memorize numbers. -- http://bash.org/?quote=98 [14:07] ttx: have fun! [14:15] In lucid, how do I get to the grub menu (so I can start memtest86+)? [14:15] hold down the right shift key [14:16] Ugh. [14:16] twb, no, ugh is that they did it to save one second on the boot time :p [14:17] twb, you can change that in /etc/default/grub though [14:17] What do you do about HHKBs and such, which have no RHS modifiers? [14:17] ... you change it in /etc/default/grub [14:17] * twb rails at change. ALL change. [14:17] cwillu_at_work: which is great unless you want to get to grub because something is wrong with the boot process [14:18] twb, silly, you should have fixed that before your boot process failed [14:19] I'm starting to think I should install an ubuntu desktop on a scratch laptop, just so I can have experience dealing with these kinds of things before I run into them on production servers. [14:20] I'm getting a crap load of these - Jun 17 16:42:33 incipio kernel: [25879.080753] [UFW BLOCK] IN=eth0 OUT= MAC=52:54:00:63:d5:dc:54:42:49:02:8c:85:08:00 SRC=10.41.0.194 DST=10.41.0.5 LEN=40 TOS=0x00 PREC=0x00 TTL=128 ID=8990 DF PROTO=TCP SPT=65170 DPT=80 WINDOW=16227 RES=0x00 ACK FIN URGP=0 [14:20] I'm trying to find the word to express the particular mixture of amusement, dismay and arrogance I'm feeling in response to that :) [14:20] Port 80 from any to any is allowed into that server... [14:20] I wish vendors wouldn't try to make the look and feel *completely different* with every point release [14:21] (That's not a criticism of Ubuntu, but the industry in general.) [14:21] twb: indeed [14:21] Like how Finder.app has a completely different nonstandard toolbar every damn time. [14:21] twb, you really should have a test server of some sort to do testing on [14:21] production servers generally aren't the place to try things for the first time [14:21] cwillu_at_work: I do, but they're VMs [14:22] MTecknology: is is probably being denied because that packet isn't part of an existing connection (ie you were doing something, then rebooted, then the webserver sent some packets) [14:22] s/ie/eg/ [14:22] And the production servers are a customer-owned hodge-podge, so I can't easily just dedicate one of 30 identical production machines to be my test machine. [14:22] jdstrand: ooooh - that makes a lot of sense - thanks :) [14:22] so, get a cheap laptop to do your testing on :p [14:23] Yeah, I have a 701 sitting at home [14:23] Not that it would've found the RAID nonsense that I was running into on #ubuntu-server [14:23] you can use a raid on a single device, kinda sorta [14:23] True [14:24] it's hard to substitute actual experience on the hardware though :/ [14:24] my usual trick is to spend time figuring stuff out for people in the troubleshooting channels [14:24] that way I get the benefit of experience without the panic of having my own stuff broken :) [14:24] Yeah, I do that, too. [14:26] I don't think too many lucid users are doing stuff with e.g. ten-year-old SCSI DDS drives, though. [14:27] I didn't have to make that work, thankfully, but I had to transplant to 2GiB (yes, two GIG) PATA drive into another host to a new motherboard so that the lucid desktop CD wouldn't bork trying to load a framebuffer [14:28] twb, oh, you know how to pull up the real failsafe vesa x on the live cd? [14:28] hit options, then escape, and then change the boot args to single; then when the recovery menu comes up, choose failsafe x [14:28] the safe-x option in the actually boot menu doesn't do that [14:28] What I want is for "textonly" to work when I append it to the boot: prompt [14:29] I get mixed up between the various forks of casper [14:29] "single" will give you that [14:29] one curses menu, and then you can pull up a normal root prompt [14:29] Hum, I thought that was gone [14:29] nope [14:29] cwillu_at_work: erm, single *on the live CD*, before install? [14:29] yes [14:29] Huh. [14:30] just did it yesterday on a laptop that would hardlock [14:31] When I did a netinst of lucid just now, I noticed that it has come full-circle and is installing and forcing me to configure an MTA (prior to tasksel), just like Debian did way back in 2006 [14:31] I accidentally booted it without priority=low, so I assume it's the default behaviour [15:03] need to create a file server for over 10 users of which I have found windows 7 is not a good solution. I prefer linux to any windows software, but I need to know what issues i might run into using a linux machine to manage file shareing for XP and Win 7 users [15:04] limitations to samba? [15:08] I dont know any as of yet [15:09] Danawar: it doesn't do AD very well yet [15:10] i.e. you probably can't make it the master server of your hundred/thousand-seat AD-centric Microsoft-centric network [15:10] What other software is there that can do Active Directory apart from windows server? [15:11] There isn't any [15:11] ahh i see [15:11] is any one working on any? [15:11] because i think that would be a very cool addition to linux [15:11] nube: if you just want a place to dump files and you have a "hard shell, gooey insides" approach to network security, it'll be fine. [15:11] Danawar: yes, the samba people are working on it [15:12] ideal cant wait :D [15:15] twb, i have 10 users at the moment that need to work off of one file server because i can't trust them to grab the files they are working on, put them on their local machine, and then when done put them back on the file server [15:15] twb, there will be more users and win 7 has a 10 connection limit [15:16] twb, thus now that my network has grown the win 7 machine that is sharing all the files used my users fails [15:17] twb, I think that an ubuntu server edition machine should be plenty sufficent for the purpose, but I need to research the possible issues before attempting to implement that option [15:18] nube: samba will be fine for that [15:18] the only need is for all my users to work off of files on one machine... [15:19] nube: do you have a scratch machine you can use as a test? [15:19] If you do, just spend an hour installing ubuntu-server on it and trying to get samba up and running. [15:19] twb, yes that was my intention.. use it for awhile and see if any issues arise [15:19] If it doesn't work, you've only lost an hour. If it does, you have something concrete you can test with. [15:20] i agree... just have to sell the idea to my boss [15:20] windows server OS seems like a waste of money and super overkill for the need [15:22] It will bit you on the arse if you don't have a "go to" guy who can give you Linux support for it [15:23] twb, at least their is plenty of documentation available for *nix...espcially ubuntu whereas a win machine is about worthless as far as docs go [15:23] Shrug. [15:24] A large part of it is experience and knowing who to ask and where to look. [15:24] There's a SCO 5 (i.e. unix) host sitting on the desk here, and if it had ACTUALLY had bootstrap errors as the customer reported, I would have NO CLUE how to debug, let alone fix, it. [15:25] twb, always been able to implement it at home... have mac os X, win 7 and ubuntu working fine on one network, ssh x11 forwarding, all that good stuff [15:25] SpamapS, do you have a WIP that i could sneak a peak at ? on uec-run-instances [15:25] And I wouldn't be much better with, say, freebsd or slackware. [15:27] twb, any decently spec'd standard desktop machine should be sufficient for hardware correct? [15:27] Depends what you mean by desktpo [15:28] But for ten users, even a Pentium III would suffice [15:28] twb, 64bit dual core processor, 4 gig ram, msi platinum MB, terabyte hardrive, 650W powersupply (overkill) [15:28] ScottK: hey. I'm working on packaging a web package called racksmith (http://racksmith.net). Looking for a) someone to review my package and b) possibly some assistance. Are you able or interested or point me to someone who may be interested? [15:29] Looking [15:30] rgreening: official or third-party/ppa packaging? [15:30] twb, i meant ti power supply was overkill... other spec's plenty sufficient? [15:30] nube: yeah, sure. [15:30] nube: assuming it's not a slow (45,000?) laptop disk. [15:30] twb: looking to get it into Maverick w/ future backport to Lucid. [15:31] Interesting tool. [15:31] demo looks broken here; entries in a rack display over the menu, while the racks themselves don't [15:32] cwillu_at_work: looks OK in w3m-el, which is a nice change for a new-fangled web app :-) [15:32] it's also painfully slow :p [15:32] rgreening: It's made of php and mysql, so I'm really not the person. [15:32] nube: I meant 5400rpm (bad) vs. 7200rpm (good). [15:32] rgreening: I suspect zul might be a good person to review it once you have something. [15:33] twb, have to check but i doubt i spec'd anything under 7200.. [15:33] ScottK: thanks. It's a real pomising project and there's very few packages like it or close to it. [15:34] nube: I mean, a ten-seat NAS could be handled by an 200MHz ARM with 32MB RAM. [15:36] twb, should be simple.. install the OS, copy the files to new machine, share folder, get users to to start working off the that shared folder, right? [15:37] cwillu_at_work: BTW, for comparison, syslinux allows any shift, alt, caps or scroll lock to trigger the "gimme the bootloader menu" [15:37] nube: yep [15:37] twb, is their any reason to use the ubuntu-server edition over the ubuntu desktop version as all it is being used for is sharing files? [15:38] twb, not using it as a web server or database server or any other of these common server requirements [15:38] twb, scrolllock on would be a good one actually; you should file a feature request though [15:38] nube: yes; it'll give you more appropriate install defaults without you needing to futz anything. [15:39] nube: during install you'll be prompted about which services you want -- you should probably tick "file sharing" and "ssh / remote access", or whatever they're called [15:39] cwillu_at_work: which package provides that functionality? grub-pc? [15:40] twb, any of the grub2 packages share the same source packages; grub-pc should work [15:40] OK. I meant grub-pc as opposed to one of the other bits that's involved in bootstrapping. [15:42] cwillu_at_work: according to the changelog, either shift key will work in grub2 [15:43] Silly question: does "splash" turn plymouth on/off as it did usplash? [15:43] twb, believe so, yes [15:43] * twb makes note to check that [15:43] actually, make that an unqualified yes [15:43] whats 'plymouth' ? [15:43] Danawar, boot splash provider [15:44] Danawar: the purple thing before the purple gnome thing [15:44] twb, note also the existence of "plymouth-log", and that "splash" without "quiet" should allow you to see most entries as well as interact with the console [15:45] The plymouth program allows you to change what the ubuntu loading screen looks like? [15:45] yes [15:45] And when set to quite shows you a long list of command that are being run or just a black screen? [15:45] cwillu_at_work: at the end of the day, it's mainly about me wanting bootstrapping to be as dumb and predictable and reliable as possible. [15:46] cwillu_at_work: so while I can understand the need for grub2 "modules" and os-prober and plymouth, I'm apprehensive about them being too clever, too early. [15:46] s/apprehensive/paranoid/ [15:46] note that plymouth was been in use for a few years now [15:46] just new to ubuntu [15:47] Yeah, well. [15:47] Xenix was used for years, but that didn't make it any good [15:48] My gear usually isn't Ubuntu's core market, either. [15:48] servers of arbitary vintage, routers, diskless PXE workstations, etc. [15:48] let me rephrase: we took flack for not using it and instead inventing our own :p [15:49] cwillu_at_work: you mean usplash? [15:49] and xsplash [15:49] Well, but I could trivially disable usplash in 8.04, so I didn't care. [15:50] well, you can trivially disable plymouth too [15:50] I have yet to confirm that with my own eyes :-) [15:50] cat /etc/init/plymouth.conf [15:51] actually, that's a lie, the check is in plymouth itself [15:51] cwillu_at_work: that's irrelevant, because if the root filesystem is mounted, I'm pretty much home free [15:52] It certainly has to at least be *installed*, because upstart Depends: mountall Depends: plymouth. [15:53] So I have (presumably unused) plymouth code installed on this router, which doesn't have any kind of video output at all. [15:53] where do you see that? [15:53] upstart doesn't depend on plymouth here [15:54] cwillu_at_work: via mountall [15:54] hmm, they should have made that a recommends: [15:54] Oh, cool. "apt-cache show foo" now has a Supported: 5y field [15:54] * cwillu_at_work suggests twb file another bug :) [15:54] That is so much easier than that old python script [15:55] cwillu_at_work: fyi, "aptitude why plymouth mountall" [15:55] yes, I saw [15:55] I figured more people need to know about "aptitude why" ;-) [15:56] aptitude why twb [15:56] That won't work on your system [15:57] Huh. It doesn't work on mine, either, despite it being installed. [15:57] "Unable to find a reason to install twb." [16:27] zul: if you have a moment, I have some questions regarding packaging a php/mysql app and inserting via maintainer scripts some sql into the db... [16:28] rgreening: sure [16:29] zul: so the package I am working on it racksmith from http://racksmith.net. It has an install.php, but doesn't setup the db ahead of time, so I was going to use dbconfig to manage everything instead of using the installer. [16:30] ok thats fine [16:30] I have the sql required in install.sql, but the installer has a couple of extra dynamic bits to be inserted. [16:31] can I call mysql via maintainer scripts to insert those bits and got any example I can look at? [16:31] e.g.: "INSERT INTO `config` (`name`, `value`) VALUES('install_date', '".date("Ymd")."'),('version', '".$systemVersion."'),('ldap_auth', '0'),('ldap_server', ''),('buildingCanvasX', ''),('buildingCanvasY', ''),('ldap_basedn', ''),('ldaps_enabled', '0'),('ldap_prefix', ''),('ldap_group', '');" [16:32] this is the PHP code for some extra bit being inserted. [16:32] the install date is dynamic for example. [16:32] thoughts zul? [16:33] rgreening: sure i would like at how packages like phpmyadmin do it as well [16:36] zul: phpmyadmin doesn't insert any sql via maintainer scripts [16:36] rgreening: it was just an example but yeah what you said should be fine [16:37] zul: so, would this go in the postinst script only? The sql insert for above? [16:37] rgreening: yes [16:38] ok. I'll try it. ty zul. [16:38] zul: Can I ping you to review the package one I upload to review? [16:38] rgreening: if i have time yes [16:38] awesome. ty zul [16:53] Hi, i create a ubuntu guest, the network is using bridge, the ip is 10.10.1.1, i can access resources on network 10.10.2.x but can't ping these hosts, i got an error: "From 10.10.2.4 icmp_seq=1 Destination Host Unreachable" [16:53] Any ideas ? === JPP_ is now known as JPP [17:29] failover: filters active on ICMP? [17:37] hggdh, yeap, my bad! [17:40] failover: welcome, and glad this was a simple thing ;-) [17:41] :) [17:41] mee too! [17:41] and thanks for the help [17:55] New bug: #596010 in ntp (main) "ntpd sementation fault using NMEA driver" [Undecided,New] https://launchpad.net/bugs/596010 === _grs80 is now known as grs80 [18:07] smoser: sorry I just now got your message.. I'll push a branch up now [18:08] SpamapS, I'll let it slide this once [18:08] :) [18:08] i'm just curious. no hurry [18:08] smoser: suer. Is the EC2PRE environment variable something that is standardized across more than just cloud-utils ? [18:10] not standard anywhere. [18:10] we just want to provide some easy way to use euca- rather than ec2- or - [18:11] Ah.. I like it.. seems like there should be a .cloudrc ;) [18:13] yeah. theres that too. [18:14] i stumbled upon mr.awsome recently, which looks pretty interesting [18:26] Hmm. I have just setup an Ubuntu 10.04 LTS Server with libvirt for kvm. And I have a guest OS up and running connected to my br0 physical network bridge, but when I try to ping from the guest OS to any IP outside of the physical server it's in, I get nothing. I've flushed the firewall out to make sure that wasn't causing it, and it's still happening and I'm not sure why. [18:26] smoser: mr.awesome is another of my aliases [18:26] The guest can ping the host's IP itself, but not beyond. [18:27] yeah, but you dont spell it as cloud-cool as he does [18:27] mr AWSome [18:28] heh [18:30] Daviey: are you still around [18:30] smoser: mr AWSome <- nice - I like that :) [18:55] smoser: lp:~clint-fewbar/+junk/uec-run-instances [19:07] smoser: yeah mr.awsome lookw pretty cool... it might be useful to support it as an alternative to ec2-run-instances since it can do some cool setup tricks [19:12] jdstrand: thank you for the comment on bug 596010, I did not know about it (and just learned a bit more) [19:12] Launchpad bug 596010 in ntp "ntpd sementation fault using NMEA driver" [Low,Incomplete] https://launchpad.net/bugs/596010 [19:12] does anybody use IPSec on servers anymore? [19:15] jdstrand: kirkland: libvirt 0.8.1 in maverick succeeded in doing save/loadvm, fwiw. pretty quickly, given it was nested inside kvm :) [19:15] \0/ [19:15] nice! [19:16] jdstrand: so thx [19:16] sure thing :) [19:16] hallyn: neat! [19:16] hallyn: will you be updating that bug then? [19:16] hallyn: i look forward to your recipe in the wiki for that; that's something i need to do more of [19:16] Daviey: hiya [19:16] jdstrand: i'm going to wait until i can test a bit more (with some hw help from kirkland :) [19:16] Daviey: i'm curious if you've made any progress with groovy [19:16] cool [19:16] kirkland: uh, which recipe? for save/restore? [19:16] hallyn: jdstrand: should be today [19:17] hallyn: yeah, all things snapshotting, in general [19:17] alrighty [19:17] on my todo :) [19:18] fyi, the qrt test-libvirt.py does do save/restore, but obviously not in a way that would trigger the bug [19:18] anybody know the ec2 ami ID for our official lucid images? or better yet, do we maintain a page showing that? [19:18] hallyn: it would be cool if you could add a test to qrt, if feasible [19:18] Daviey: i'm hoping the /usr/share/groovy/embeddable/groovy-all-1.7.0.jar suggestion from nurmi helps? [19:18] SpamapS: yeah, smoser has a page [19:18] SpamapS: one sec [19:19] ahh google doth provide [19:19] http://uec-images.ubuntu.com/releases/10.04/release/ [19:19] SpamapS: http://uec-images.ubuntu.com/lucid/current/ [19:19] SpamapS: ack [19:19] SpamapS, right . there is that. [19:19] if you want to be fancy, there is [19:19] http://uec-images.ubuntu.com/query [19:19] which is more programatic abble [19:20] would be cool if uec-run-instances just did that query for you [19:21] --relese lucid --arch i386 --root instance [19:21] jdstrand: are you talking about bug 524447? [19:21] Launchpad bug 524447 in qemu-kvm "virsh save is very slow" [Wishlist,Confirmed] https://launchpad.net/bugs/524447 [19:22] jdstrand: if so, i don't see how you can do save/restore so as to "not hit the bug". unless yo usave/restore a 500k image... [19:23] hallyn: I wasn't doing timing in qrt. I was more just saying that if you have other save/restore tests, please add them [19:25] ok (and you probably are doing a small image :) i don't have an automated test at this point. i'll have to take a look at the qrt and see if i can tewak the existing test to detect the bug [19:25] SpamapS, yes, it would indeed be cool [19:25] so I guess the current test would trigger it, but it is a pretty small image [19:25] SpamapS, difficulty there, though [19:25] right now, we're [19:26] hallyn: the save/restore test is in scripts/libvirt/libvirt-apparmor.sh (to save you some time) [19:26] a.) UEC/EC2 agnostic, by just fronting tools, we know nothing about --region. we just assume its right [19:26] zul: racksmith has jquery and jquery-ui included with the source, is that an issue, or do I need to patch to pull them out and use the system packages? Thoughts? [19:26] b.) there is no published data like that ofr UEC [19:26] jdstrand: kthx [19:26] but, i really would like to do that [19:26] err... rather i'd like to have that support, it would be cool/ [19:27] kirkland, ping [19:28] rgreening: system packages please [19:28] smoser: yo [19:29] ccheney: any progress on that euca bug? [19:29] what do you want cloud-config-byobu to look like [19:29] zul: ok, I'll speak with upstream and see how easy it is to pull out. [19:29] smoser: for uec it would be pretty easy for it to publish such a list by default though. [19:29] is the only knob that you'd want to set 'enable by default' ? [19:29] kirkland, just caught up with my large email backlog and started working on updating the kernel now (back from lunch) [19:29] SpamapS, i dont think i follow [19:29] SpamapS, the issue with UEC is that those magic numbers differ for everyone's cloud [19:30] smoser: right, so we'd have to query their cloud.. and just say "if you want this to work, you must fill in these optional metadata fields" [19:30] smoser: long term, not short. :) [19:30] smoser: I get a little pie in the sky sometimes... :-P [19:31] right. yeah., and to be consistent that query mechanism from AWS [19:31] the thing that is doable *now* is naming convention [19:31] describe-images ... parse output... client side decide [19:31] heavy though and yucky [19:31] :) [19:31] Okay, I'm seriously loosing my mind here. [19:31] but, our image names are consistently named, so it can be done. [19:32] but describe-images without an ami id pulls > 1M of data on us-east-1 [19:32] smoser: yeah thats no fun, it has to be a different metadata source [19:32] Apparently my networking issue was with my bonding. Which is really odd, cause this would be the first time that ever failed. [19:33] kirkland, ^^ [19:33] smoser: hmm, what's cloud-config-byobu? [19:34] for cloud-config syntax (in cloud-init). so you can do something like [19:34] --user-data "#cloud-config\nbyobu: on-by-default" [19:34] kirkland, should i see this with lucid with just using the maverick kernel? or should i do a full maverick install for testing? [19:35] ccheney: i think full maverick install would be most accurate [19:35] kirkland, ok will do [19:35] ccheney: reproduce the problem there [19:35] ccheney: then replace just maverick's kernel with lucids; reboot; should fix problem [19:36] ccheney: then it's a binary search of the kernels between [19:36] yea :) [19:40] hmm.. is Amazon going to get mad at me for starting/terminating instances constantly? Maybe I should setup a euca here at home... [19:45] I currently have an ubuntu 9.04 kernel 2.62.28-19 server that loses network connection 4-6 times a week and the networking service has to be restarted to get connection back and I cant seem to figure out why [19:52] SpamapS, why would they get mad ? [19:52] err... wait. i mean... here , let me help you. [19:52] I'll charge you $0.03 cents for every instance you start up on my cloud [19:53] even if you shut it down immediately you pay for an hour. I think they're happy with that. [19:56] haha [19:56] good point [19:57] New bug: #596034 in apache2 (main) "Please merge apache2 2.2.15-5 (main) from debian unstable (main)" [Wishlist,Fix released] https://launchpad.net/bugs/596034 [20:02] Client.PendingVerification: This account is currently being reviewed by our team and verified as a valid new account. Please contact aws-verification@amazon.com if you have questions. [20:02] see [20:02] they got mad at me [20:03] [q] when, trying to upgrade hardy to 10.04 - the installer failed, saying that it cant work out for me, something with a missing package for LVM2 [20:05] wait , what ? [20:05] wow. how many instancess had you started ? [20:05] smoser: 3 [20:05] well [20:05] I had 3 running [20:05] yeah, thats strange. [20:05] I had started 8 total in the last hour [20:05] way to little for them to care [20:05] I had a power outage here today. Actually, I had two power outages in 15 minutes. I have /boot on a software raid1 and / on a raid5. Now, it doesn't boot, or at least, it takes a very long time. I only see a blinking cursor. Should I be worried? [20:06] when i run the image tests, i run 15 or so per ami per region [20:06] smoser: agreed. I've run 8 before too [20:06] it costs about $40 [20:06] 3 weeks ago I had about 10 going for some puppet/elastic load balancer tests [20:07] thats very strange. [20:08] [i] my system is set up: on 4 hdd's (part 1 (4x mirrored 8gb = (lvm) root ) - part2 (4x 1gb = 4gb swap) part 3 = 4x 0,99TB raid 5 /home [20:08] Though then I used the console to spawn.. [20:08] smoser: maybe they don't like my newly created key/cert [20:09] anyway I have some errands to run anyway [20:09] damnit I was so close. >: [20:09] Ichat, I've never seen the installer say "I can't work out for you. Something with a missing package for LVM2". [20:10] jo-erlend its not a clear error message - but ill try to get it as acurate as possible... [20:10] never hurts. [20:11] SpamapS, well, push what you have. [20:11] kirkland, alright, so lets say i wanted to eanble byobu by default. [20:11] how would i do such a thing for a single user [20:11] how would i do such a thing for the system ? [20:12] smoser: ok pushed.. ttyl [20:13] maybe Daviey knows that ? [20:13] step 1 inserting the disk (ok) 2 part manager (formatting as ext3 (seems fine). 3 installing linux base - neard the end the installer says (failed to install - (not failed to install %name%.deb - or package corrupt .... nope just failed to install .... [20:13] or any other byobu fan boys [20:13] smoser: sorry [20:13] smoser: there are two ways: [20:13] in a dif terminall window (f3) i find some errors about cant install package lvm2 not availible [20:13] i figured if i said "byoubu" enough times you'd come :) [20:14] smoser: a) each user can: byobu-launcher-install [20:14] smoser: b) or globally: dpkg-reconfigure byobu [20:15] smoser: does tht help? [20:15] sure. hang on just a minute [20:15] New bug: #596041 in squid (main) "/var/tmp is not cleared after squid restart" [Undecided,New] https://launchpad.net/bugs/596041 [20:18] kirkland, http://paste.ubuntu.com/451732/ [20:18] so, the goal will be that: [20:19] euca-run-instances --user-data="#cloud-config\nbyoubu_by_default: user" emi-abceefg [20:19] would do what you would think [20:19] euca-run-instances --user-data="#cloud-config\nbyoubu_by_default: system" emi-abceefg [20:19] for all users [20:23] so is there no mdadm for 10.04 amd64 server (could that be it? ) [20:23] of course there is. [20:25] I currently have an ubuntu 9.04 kernel 2.62.28-19 server that loses network connection 4-6 times a week and the networking service has to be restarted to get connection back and I cant seem to figure out why [20:25] so why is my install complainting about it not there (cant give the exact eng version of the error but it says some about it being not avail ... i used hardy to create the lvm volumes and the raid devices. they are on... [20:26] kirkland, should installing the maverick nc not automatically detect the controller? [20:26] * ccheney wonders if he needs to reboot his controller again or something [20:33] kirkland, ping! :) [20:38] gah its not running the webserver :-\ [20:39] thats probably why it can't find it [20:39] * ccheney wonders if MaxClients = 1 is the problem [20:40] anyone recommend good Offline reading for learning "Ubuntu-Server"? [20:41] MetaJake: - a few yah, but id have to look up the author first... [20:42] http://www.bol.com/nl/p/engelse-boeken/ubuntu-certified-professional-study-guide/1001004006009485/index.html [20:43] very new to ubuntu. but my goal is to learn server architecture for the sake of deploying Python driven websites. [20:43] sorry about the nl link but the title is En [20:43] lchat, gracias [20:43] MetaJake: The Ubuntu Server Guide is also available as a PDF. [20:44] There's a link in /topic that I think will get you to it. [20:46] smoser: when you run the tests on EC2, do you ever have instances that fail to start? [20:46] MetaJake: https://help.ubuntu.com/10.04/serverguide/C/serverguide.pdf [20:46] hggdh, occasionally. [20:47] i attribute it to one of 2 things: [20:47] a.) our kernel [20:47] b.) amazon's fault [20:47] steggan, scottk thank you [20:47] rarely do i have a real strong feeling that it is 'b.' [20:47] but sometimes i think their system just falls over [20:48] anyone here deploy python driven sites such as Django or Pylons on servers they have built via Ubuntu-Server? [20:48] * ccheney found out how to solve his problem after calling kirkland :) [20:48] I'm sorry if thats a... too detailed question. I'm new to all this. [20:48] steggan = steffan : \ [20:49] !tab [20:49] You can use your key for autocompletion of nicknames in IRC, as well as for completion of filenames and programs on the command line. [20:49] ubottu: thanks for reminding me ! :) [20:49] Error: I am only a bot, please don't think I'm intelligent :) [20:49] XD [20:59] I just set up my postfix/courier server, and it is all working except for the ssl/tls login. anyone know of a good tut that covers ssl setup? [20:59] I've been googling for a couple days now and I can't seem to find one [21:00] kirkland, the above syntax is supported by cloud-init 0.5.12, which will appear shortly in an archive near you. [21:00] * smoser heads out for the day [21:00] holmser: everything you need is in the postfix howto [21:01] postfix oficial documentation [21:01] I followed this tut to get my server set up: http://holmser.net/i [21:01] holmser: for courier afaik you need to concatenate a dh helman + publick key [21:03] holmser: do you already have the certificates? [21:03] I believe so... I'm double checking right now [21:04] I know I set up certificates in the tut [21:05] I'm still learning the whole mail server thing [21:05] I have my own howto: [21:05] http://tuxjm.net/docs/temporales/02Instalacion_y_Configuracion_de_Postfix_Dovecot_Mysql_con_PostfixAdmin_sobre_Ubuntu_Server_9.10.txt [21:05] it is in spanish [21:06] check the section ==Configurar parametros SSL/TLS para el cliente y servidor SMTP== [21:06] they are all the requiered parameters for postfix [21:06] New bug: #596064 in samba (main) "nmbd fails to start on boot - problem with upstart " [Undecided,New] https://launchpad.net/bugs/596064 [21:07] if you want to create your own Certificate Authority check next chapter ==Creación de Autoridad Certificadora SSL/TLS== [21:08] I hope next week I finish the full howto [21:09] I already have the content, I jus need to translate it to docbook [21:09] it is about 100 pages about mail servers... [21:09] ok, I just rechecked the tut I followed, and I set up certificates [21:10] just to be clear, when I check my email, I am only dealing with courier, correct? [21:10] holmser: yeap [21:10] if you use pop3 or imap you deal with courier [21:10] postfix handles the sending and receiving of mail, and courier handles my logins? [21:11] well postfix is a Mail Transport Agent [21:11] its function is transporting messages via SMTP [21:12] postfix uses smtpd service to listen in the TCP/25 port [21:12] and smtp (client) service for sending mail out to other domains [21:13] postfix by itself doesnt manage logins/authentiatoion [21:13] ok, so I should really be looking for courier ssl/tls tuts [21:13] well if you want to use a secure channel for sending mails you need to add ssl/tls to the smtpd service in postfix [21:14] if not you are going to send your user and password in plain over the wire [21:14] the same for pop3/imap [21:15] well right now I have postfix set up to deliver mail through gmail through ssl [21:15] holmser: by the way Im not sure if courier is 5 years supported [21:15] last time I checked it was in multiver repository [21:15] 5 years supported? [21:16] holmser: yeap, for the Long term Support for 10.04 server... [21:17] well time to launch [21:18] dovecot is faster, secure and more flexible than courier, and is the recommended option for ubuntu [21:18] with courier I think you are on your own [21:18] I migrated most of my servers to dovecot, and they all rocks [21:22] jmedina: - only one tiny flaw :S [21:23] ebox (ubuntu's favority webbased managment suite (for sbs like services) - still runs courier [21:25] not that i agree with them or even the slow development of ebox .... but its still ubuntu's (first bet - as theres still to few support for webmin /usermin - specially for sbs- likes [21:27] Ichat: please when mentioning webmin remember that it is not supported and may break your system severely, as it cannot deal with some config files in ubuntu/debian [21:28] Ichat: otoh: do you have an ubuntu system with ebox? are you content with that? [21:31] so i got past the groovy bug but now it seems my nc isn't showing up it says i can run 0/0 [21:32] ccheney: is it really up? [21:32] hggdh, hmm i think so, not completely sure [21:33] i'm going to shut it down and bring it back up and see if it works [21:33] try ssh-ing there and checking; also verify KVM is loaded [21:34] hggdh, i didn't see anything that stuck out in /var/log/eucalyptus, but not sure what is wrong [21:34] so eucalyptus-nc is running... what does euca_conf --list-nodes say? [21:36] euca_conf doesn't like me [21:36] * ccheney looks to see if he did something wrong in the howto [21:36] euca_conf likes nobody ;-) [21:37] try (on the CC) sudo euca_conf --discover-nodes [21:37] I want to test my spam filter. Where can I post my email address to make sure that I get a ton of spam? [21:37] hggdh, well it claims it can't get credentials [21:38] hggdh, http://pastebin.ubuntu.com/451768/ [21:38] ccheney: oh, you must run it as root [21:40] hggdh, oh ok [21:41] hggdh, discover-nodes says: [21:41] INFO: We expect all nodes to have eucalyptus installed in //var/lib/eucalyptus/keys for key synchronization. [21:42] so it seems like either i botched up my maverick install or there are problems with registering on it using automated method === oubiwann is now known as away === away is now known as oubiwann [21:43] seems like the keys did not get propagated [21:44] yea, going to try the package install instructions to see if that works for registering the node [21:47] should there be a eucalyptus-nc-publication on the cloud controller in the 2 system setup? [21:47] or is that to just run on the nc? [21:48] the problem is that there is not really a way of propagating the euca keys without manual intervention of some sort [21:48] at least, none that I know of [21:49] hggdh, kvm is not running on the nc btw [21:50] hggdh, i tried running: sudo -u eucalyptus ssh-copy-id -i ~eucalyptus/.ssh/id_rsa.pub eucalyptus@ which did not help [21:50] it did seem to do something but didn't help make the node show up [21:50] ccheney: run kvm-ok -- what does it say? [21:50] hggdh, says its good [21:51] hggdh, this is the same two boxes i have been previously using on lucid and did maverick reinstall for them to test the kernel bug [21:52] maybe i should just pull the old a1 image down instead of using my synced current iso, it may have less trouble :-\ [21:53] ccheney: the euca keys are stored under /var/lib/eucalyptus/.ssh [21:53] hah, you already did iy [21:54] it [21:54] ccheney: er. Shouldn't the target be ~/.ssh/authorised_keys? [21:55] dammit. authorized_keys. This s and z thing catches me every time :-( [21:56] yea its there 4 times now, heh [21:57] i'm going to give up trying to debug that issue and just download alpha 1 and see if it works [21:57] if that does then we know something else bad is happening with current [22:00] heh i just noticed there is a high bug about node reg [22:01] I'm running UEC on 10.04 and when I run euca_conf --list-nodes it returns a blank line. How do I fix it? [22:03] goldins: this sounds like your node(s) did not register [22:04] hggdh: I tried running euca_conf --register_nodes and it seemed to work fine, but it still returns a blank line [22:05] Trying rsync to sync keys with "165.112.92.216"...done. [22:06] goldins: ssh into there, check if (1) eucalyptus is running; (2) run 'kvm-ok' and verify all is fine [22:07] hggdh: how do I tell if eucalyptus is running? [22:07] ps aux | grep eucalyptus [22:08] well that returns apache's threads and avahi's publish thread [22:08] is that all it takes for euca to be running? [22:08] guntbert: its not bad if it works [22:09] incidentally, kvm-ok says that kvm is disabled in my bios [22:09] witch is not allways sadly to say [22:09] i got new info on my boot problem... [22:11] Ichat: up till now I tried it twice and removed it immediately from the machine - I want a manager software that works directly on the config files of the system and not on it own ones - thats why I was asking [22:11] i build a new set of faikraids 1 4x 6gb raid 1 = /root 4x 1gb raid 10 - swap and the rest in raid 5.. here is the deal i get the same error bunt... [22:12] guntbert - i understand yah, - thats its real drawback - that and the slow development of it [22:12] I would say that this is a bug, as it warns you while installing the CC that you must have VT turned on but doesn't warn you when installing the NCs [22:14] the error i get ;s mdadm does not have a valid installeble candidate it may be outdated or for a diferent system - even though its required by a diferent package...... base-setup error code 100 [22:41] hggdh, i think my install might have gotten messed up due to the groovy bug and installing groovy may not have been enough to salvage it, in the process of installing the a1 version now [22:43] is it possible to control the bandwith on my ubuntu-gateway's NICs? so fx. dmz gets 10mbps/10mbps and LAN gets 40mbps/40mbps? [22:44] sure [22:44] DUEDAHL: sure [22:44] how? :) [22:44] you mean local traffic? [22:44] or internet? [22:44] internet [22:44] damn I only get 2mbps XDDD [22:44] haha :D [22:45] Im jelous I wont tell you [22:45] XD [22:45] :D [22:45] well it not that easy [22:45] you can use tc for that [22:45] I gess you already have iptables [22:45] yep [22:45] I prefer shorewall for easy traffic shapping [22:46] with new kernels you can get egress and ingress traffic shapping [22:46] if you you can use tc command by hand [22:46] DUEDAHL: what do you use for your firewall ruleset? [22:47] commandline [22:47] i think im gonna try tc.. [22:47] is shorewall gui? [22:47] nop [22:48] it its file based [22:48] ok [22:48] you write rules and then shorewall will create iptables, ip, and tc rules [22:48] goldins: if kvm is disable on your node... are you usaing kvm for the virtualisaiton? [22:48] ok cool [22:48] ccheney: yes, start fresh, probably a good idea [22:59] Hello, I am having a problem with Ubuntu 10.04 server. When I transfer a large over the network to the server hard locks during the transfer. This happens over ssh, and samba. The destination directory is an ext4 2tb hardware raid10 array. When I create a large file locally on the machine it does not crash. Memtest reports no errors after 3 passes. Does anyone have any ideas? I have a few... [22:59] ...logs. I was able to set up a netconsole, and saw the output of a kernel oops (http://paste.ubuntu.com/451797/) on my logging machine. I also grabbed all of the other logs off of the machine after rebooting it. [23:02] Rigorm0rtis: good! Now, please open a bug on this -- make sure you add in the OOPS text in toto you captured [23:02] hggdh: Where do I open the bug? [23:03] on https://bugs.launchpad.net; you must have an account there [23:07] I have an account. How do I create a bug report? [23:09] probably a good idea is to run 'ubuntu-bug -f --save= linux' on the server, and save & move the resulting crash file to a desktop, where you can then [23:10] move to a desktop and run 'ubuntu-bug -c ' [23:10] Rigorm0rtis: then manually attach your OOPS log (from the serial console) [23:12] hggdh, ok reinstall fixed it :) [23:12] so current maverick seems to have issues probably due to groovy that aren't fixed by just installing the old version from the ppa [23:13] * ccheney now gets to testing the kernel issue [23:14] So, is that ubuntu-bug program a part of the bug sumbission system? [23:17] ccheney: hey [23:17] kirkland, hi [23:17] ccheney: Daviey confirmed that updating that symlink allows maverick's euca to start [23:17] ccheney: i just forwarded you a mail [23:18] kirkland, ok, i tried installing the current maverick and then using the groovy from ppa but it refused to ever let me register the node, i am not sure why though [23:18] kirkland, so i reinstalled back to alpha 1 and that worked for me [23:20] upgrading from alpha 1 iso install to current might work though with the change listed in the email [23:21] Rigorm0rtis: yes, it is the best way to report -- in your case, it will collect a lot of data (logs, etc) that will help triage [23:22] hggdh: Okay, sounds good. I'm assuming at some point it will give me a chance to attach that netconsole output? [23:23] Rigorm0rtis: after the bug is filed, you will have to manually add it in, as an attachment [23:23] ccheney: fresh maverick install, sudo ln -sf /usr/share/java/groovy-all.jar /usr/share/eucalyptus/ [23:23] groovy.jar [23:23] (line break fail) [23:23] Daviey, ok, will try that out later after doing the kernel testing [23:23] ccheney: you rock. [23:25] Daviey, on the kernel issue did you just need to revert the kernel on the nc or both? [23:25] hggdh: All right, I think I'm getting it now. Unfortunately, I won't be able to access the troublesome box until Monday. Thanks for you help. [23:27] ccheney: just the main cloud [23:27] you don't even need an nc to test tbh [23:27] Daviey, ah ok [23:27] jiboumans: Ping [23:27] Daviey: pong [23:29] What's the "Install minimal virtual machine" mode entail? [23:30] I'm trying google but my google-fu is apparently weak... [23:45] ok this is weird [23:45] i installed 2.6.33 and now it won't give me console access anymore, just ssh [23:45] ccheney: i think that is a known bug [23:45] lol [23:46] * Daviey is starting to dislike maverick [23:46] ccheney: it's nearly middnight here.. so i'm going... have a splendid weekend. [23:46] what do you hit to actually have grub show the menu? i thought it was hold down shift but that doesn't seem to work [23:46] ccheney: Oh, could you mail me your findings, when you knock off.. [23:47] Daviey, have a good weekend :) [23:47] ok [23:47] ccheney: i thought ANY key :/ [23:47] \o [23:47] hmm ok, it seems to hate me :) [23:53] ccheney: try ESC [23:54] hggdh, ok, i tried shift again and it worked, i found out i had to hold it down after it showed the grub loading message for a few seconds it seems [23:55] what the *#$%% is wrong with ubuntu server 10.04 x64 .... it just seams like they trashed softraid rather than to 'improve' it [23:56] Hi All, I have a question.. I'm new to Apache, and I have a web template that purchased, am I better installing it into the default dir /var/www directly in this folder or better creating a sub folder for every website I want to host (if I can do that?, but I don't want users to have to type the subdirectory when they come to my domain) [23:56] i fully zerrowed my 4 disks to exactly start all over again with new disks, and still the installer give me unclear unsolvable ???? - mystery error messages about mdadm being wrong