[00:18] using vmbuilder inside a private network… should —proxy pointing to my apt-proxy URL be sufficient? [00:20] I have maas setup. Is there another site for maas setup? [00:24] I don't have the web interface for maas [00:24] harushimo: another one? what are you after? [00:25] I can't get it to add nodes [00:25] web interface [00:25] I'm not running a gui on the server [00:26] harushimo: what does that have to do with it? a web server doesn't require a graphical environment [00:26] how I access the http? then [00:27] https://wiki.ubuntu.com/ServerTeam/MAAS#Create_MAAS_admin_account [00:27] harushimo: the same way you access any web site [00:27] harushimo: what happens when you point your browser to http://your_server_ip_address/MAAS ? [00:28] it keeps reloading [00:28] I'm using maas on my virtual machine [00:29] I'm guessing that doesn't matter [00:29] it does [00:29] it doesn't show anything [00:30] sorry [00:30] it is importing the isos right now [00:30] i'm at that step [00:30] I'm getting ahead of myself [00:31] I just realized I need to change my default gateway [00:31] I use the ip address to access my router setup [00:31] how long does it take to import isos? [00:32] harushimo: i think it takes 4, all precise, by default. so not long [00:32] harushimo: and take your time, haste makes waste (Ent saying) [00:33] 4 hrs? [00:34] I'm guessing that is what it is [00:34] harushimo: no, 4 images [00:34] my bad [00:34] sorry [00:34] how long does 4 images take to download? [00:34] i'm guessing that is dependent on the internet connection [00:35] harushimo: bingo [00:35] thank you [00:36] this is my first time doing maas. I may fail at it [00:36] that is okay. I will learn something [00:38] harushimo: well, it's quite new. it's bound to have bugs. report back and let me know how it goes [00:42] I will definately do that [00:42] one thing I'm thinking [00:43] please tell me if this isn't a good idea [00:43] maas is use for some who has systems background [00:43] this is just me saying: Why can maas have front end [00:44] the user can click and get it done quite quickly [00:44] if it does, I want to know about it [00:44] hehe [00:44] we can probably stream the process but like you said, it is new [00:45] I will definately know about the bugs [00:45] if this a bad idea, let me know also [00:52] how many machines would need for maas? [00:52] I mean ideally [00:55] what is the difference between maas & enlightment or just doing an install of ubuntu server and then installing maas after the installation of ubuntu server? [01:05] does maas cloud cd have 32 bit? [01:05] I have the 64 bit iso [01:06] anyone? === koolhead17|afk is now known as koolhead17 [01:11] hi all [01:12] i've got an urgent problem [01:12] upgraded my server, screwed up with ip route (which should be only temporany, and reset after reboot), rebooted and can't access anymore to the server via ssh [01:13] I can access the server in rescue mode and from the syslog mysql and other startup programms are starting [01:13] maybe you have a wrong routing table in your box [01:13] so I though it's a ip route problem, but shouldn't things modified with "ip" command be temporanyt? [01:13] of course you have to clean it [01:14] i also tried ip route flush table main from the rescue mode chrotting the server ea1het_OFF [01:14] and after restarting to normal mode still doesn't work [01:15] i'm not sure the rescue mode serves this matter.... [01:15] well that's the only way i can access the server [01:15] hum.. and i understand you don's have physical access to your server.... right? [01:15] nop [01:16] the only idea i have is to create a shellscript you can configure... put it on the rc.local init script.... and that script flush all tables and set one and only one default route to your gateway [01:16] maybe it works [01:17] same command you introduce in the rescue session but inside a shellscript to automatically be executed from rc.local in the normal init run [01:18] k i'll try thanks [01:19] you're welcome [01:20] does anyone anything about maas here [01:28] Okay, so, I'm having some ANNOYING issues with nslcd NOT starting at boot-up in Ubuntu 12.04. [01:29] reading online they say that ip and route commands aren't permanent ea1het_OFF [01:29] also I tried to do as you said and it didn't work [01:30] what do you have in your interfaces file (/etc/network/interfaces) ?? [01:30] take a look at the config for incoherences... [01:31] 1sec restarting to rescue mode [01:33] ea1het_OFF, the samething I've had for months [01:34] and these are the commands i've added to rc.local: [01:34] ip route flush table main [01:34] ip addr add 46.105.109.0/24 dev eth2 [01:34] ip route add default via 46.105.109.254 [01:38] looks good [01:39] lux`: ip route ls [01:39] lux`: ip rule list [01:39] ip route flush cache [01:40] i can connect only through rescue [01:40] is the ssh daemon running in normal init? [01:40] always been [01:41] if you don't have physical access ... how are you connecting to your box and changing the init level? [01:42] through a rescue system the hoster gives me ea1het_OFF [01:42] it let me connect to the real server harddisk [01:42] so i can chroot to it [01:42] you are into a VPS ¿ [01:42] ? [01:42] anyway yeah, ssh is enabled in init level 2-5 [01:42] dedicated server [01:43] what did you tried before, when the connections stop working? [01:43] ip route del ?. [01:43] but as I said I read that it refresh every restart [01:43] shouldn't it? [01:44] it should.....yes... [01:44] did you do any kind of change to the network configuration files? [01:45] yeah actually I did, added another nic but i restarted network and everything was fine, anyway after I removed it thinking it was the problem [01:45] so now it's exactly how it has been for months [01:46] when you enter in chroot mode... can you get access to the internet from your server? [01:46] i don't want to think it's a temporary problem in your hoster.... [01:46] yes [01:46] i can [01:47] run in normal mode [01:47] can't access anything in normal mode :/ [01:47] telnet your_server_ip 22 from you pc and wait for a connection.... [01:47] won't work [01:47] connection refused [01:47] and tell me if if presents a prompt [01:48] so only two things can happen.... no ssh daemon running in the correct interface.... or no interface well configured. [01:48] i can even start ssh in the chrotted env. [01:49] and /etc/network/interfaces is like it has always been [01:49] :( [01:51] iptables -L -n [01:51] to see if there is something blocking your machine [01:52] (firewall rules i mean) [01:52] any firewall daemon configured in the box? [01:52] i even removed iptables to make sure it wasn't it [01:52] there is not so much.... [01:52] so it can't be it [01:52] execute the line above to be sure.... [01:53] or reboot your machine [01:53] command not found [01:53] i'll try to reboot with only the default nic [01:53] good.... [01:55] lux`: it is not a good idea to configure interfaces and subinterfaces together [01:56] if you are going to work with subinterfaces.... only configure subinterfaces for all networks [01:56] without configuring the physical device [01:57] about the rest...loooks good [02:00] nothing, I'll check logs to see if ssh is giving any error i guess [02:05] Hmmm [02:06] So, I've converted my nslcd init.d script into upstart scripts, and at startup it seems nslcd.conf actually IS running, but nslcd-k5start.conf isn't. Very strange... [02:06] is there a network-manager specific event? [02:13] I've trying to setup maas. Is it possible to set it up on ubuntu-server 32 bit? [02:18] can anyone help me with maas at all? [02:18] I've been at this problem for a week [02:18] the instructions are 64 b it iso [02:18] not for 32 bit [02:26] what the fuck [02:26] i even put debug to log level of sshd [02:26] and it's not showing NOTHING AT ALL [02:32] lux`: language please. what's the problem? [02:36] pmatulis: can I ask you a question [02:36] you helped a couple hours ago [02:37] is it possible to install maas on ubuntu-server 32 bit [02:37] ? [02:37] when I look at the instructions, it is for 64 bit [02:39] pmatulis, sorry i'm very frustrated, it's almost 5am here and still can't fix the server, anyway there is a discussions of like 50lines if you scroll on the top [02:39] harushimo: yes, you can [02:39] lux`: yes, i read it now [02:39] how [02:40] I've been really struggling here [02:40] there are no instructions for the 32 bit [02:40] install lsh-server now? so i'll check if there is a problem with openssh or something else [02:40] harushimo: well, get yourself a 32bit-based machine and away you go [02:41] i'm running 32 bit vm [02:41] that I know [02:41] harushimo: so what's the problem? just 'apt-get install maas' [02:41] alright [02:41] that is easy [02:41] I'll install the ubuntu-server [02:41] okay I see now [02:42] lux`: so you put in a bad route with the ip command and now you cannot ssh into the machine? [02:43] exactly, but everything done with the ip command should be temporany right pmatulis ? [02:43] so now I'm thinking that there is a problem with ssh [02:43] lux`: yes, it won't survive a reboot [02:44] and also I'm not reading from the first "faulty" reboot anything about ssh in /var/log/auth.log [02:44] not even putting the log level to "DEBUG" in sshd's config, so there is something wrong I guess [02:44] lux`: is the ssh daemon running? [02:45] how can I know, it's not visibile from the outside anyway [02:45] nor in the logs [02:45] lux`: with the ps command [02:45] I can't access the machine in normal mode [02:45] lux`: i thought you had an alternative method [02:46] yes I can connect in "rescue mode", basically I can connect to a temporany computer which got access to the main server's hdd [02:47] lux`: so you can read data on the drive? browse the filesystem? [02:47] yes [02:47] check if there is a pid file for sshd [02:48] what release are you running there? [02:48] nop no pid pmatulis [02:49] what release are you running there? [02:50] 12.04 [02:52] where did you check for the pid? [02:52] /var/run/sshd/ [02:53] try /run/sshd.pid [02:53] nothing [02:53] did you much around with /etc/ssh/sshd_config at all? [02:54] *muck [02:54] don't remember doing so, anyway now i'm installing lsh (taking so much time to create the random seed) and then if it doesn't work I'll try to purge and reinstall openssh [02:54] but the odd thing is that there is no error nor anything at all about ssh in the logs pmatulis [02:55] since the server problem [02:55] lux`: you did a 'grep sshd /var/log/syslog' and no error? [02:55] yeah and also /var/log/auth.log [02:56] well, the reason why you can't ssh into the box is because sshd is not running [02:56] auth.log will be empty of anything useful if sshd is not running [03:05] yes yes yes [03:05] changing ssh server worked [03:05] my eyes are gonna cry blood in the pain of frustation [03:05] it's 5am but it's all right [03:11] what would be the implications of having, say, 500 ssh tunnels open at once? [03:11] to remote machines [03:11] lux`: so you can log in now? [03:12] mgw, lots of CPU overhead for the encryption [03:12] yes and i don't even wanna know what's wrong with openssh [03:12] i even purged an reinstalled it [03:12] it didn't work [03:12] using dropbear now [03:12] each tunnel has to run the expensive RSA for the handshake, but then also has to keep up the AES tunnel [03:13] and i can finally go to sleep, the sun is rising... [03:13] gnite and thanks pmatulis ea1het_OFF [03:13] qman__ thanks… so it wouldn't work very well [03:13] if at all [03:13] well, it'll work if you have enough CPU [03:13] ok [03:13] any idea how much? [03:13] any CPU made in the last few years can handle that, but it's a nontrivial amount of computation [03:14] that you have to take into account when designing the system [03:14] if that's all you're doing, a modern chip can handle it [03:14] ok, but 100 would be more doable, even on a smaller system then? [03:14] lux`: good. let me know why sshd could not start. g'night [03:15] it can, but on a typical desktop, doing 100 SSH tunnels with AES256, expect between 10% and 30% CPU use [03:15] you'd have to do some tests to get actual numbers, but IME it should be in that ball park [03:16] my 200MHz K6 starts to choke at about 6 [03:17] qman__ thanks [03:17] would I need to look at ulimits to open this many? [03:18] no, by default the limits aren't set [03:18] now, those numbers also depend on the amount of traffic going over those tunnels [03:19] idle tunnels or light communication doesn't need as much CPU power as, say, a file transfer [03:20] qman__ obviously [03:20] i'm not sure i'd actually use a tunnel… maybe a straight ssh connection, but keep them open [03:21] I'm trying to optimize an application that opens and closes connetions repeatedly [03:21] my K6 can't actually keep up with an SFTP transfer, it's notably slower than my uplink [03:22] ok… but that's a K6 [03:22] ;-) [03:22] it is, just making the point that a file transfer takes more than 7 or 8 open SSH sessions [03:22] ah, understood [03:22] ok, so this is how the system currently works [03:23] if I'm working with 100 systems, it forks 10 processes (say) and opens 10 connections at a time [03:23] closes them, and and loops through the 100 servers that way [03:23] for each task [03:23] then it starts over [03:23] for the next task [03:23] that's going to be really inefficient if you have more than one task per server [03:23] is the opening/closing more intense than keeping them idel? [03:24] much more [03:24] ok, that's what I was hoping [03:24] you're talking about 2048-bit (or more) RSA per connection [03:24] b/c it's also really slow [03:24] versus keeping an AES256 open [03:25] hoping, b/c I want to make this more efficient [03:25] and that isn't 8 times more complex, it scales exponentially [03:27] you're going to want to keep the tunnels open as long as is practically possible [03:28] they do die off with flaky connections after a while, so you might want to make them no more than an hour long depending on your network and application [03:28] but that's the range you're going to want to shoot for, not seconds or minutes [03:36] qman__ I don't really need them for more than a few minutes [03:36] right now they're being opened/closed several times a minute [03:37] maybe 15 minutes if there're a lot of longer tasks [03:48] qman__ one more question… if I open a tunnel via a subprocess, will the tunnel close when the parent process dies? === wylde_ is now known as wylde === dendro-afk is now known as dendrobates [08:10] Why does byobu look bad over ssh? === Lcawte|Away is now known as Lcawte [08:45] joseph__: it does? [08:47] yea maybe a bug. there is a scolling problem had to disable the logo to fix that. but i am trying to fix this http://img600.imageshack.us/img600/401/screenshotfrom201206020.png [08:49] looks like a curses bug [08:49] file a bug :) [08:49] what should i use tmux or screen [08:49] * RoyK just uses screen [08:50] it looks even worse when i use tmux. [08:50] joseph__: but please file a bug - bugs aren't fixed unless reported [08:58] file the bug report https://bugs.launchpad.net/byobu/+bug/1007753 should of explaned it better [08:58] Launchpad bug 1007753 in byobu "menu display screwed up over ssh?" [Undecided,New] [09:00] start up a ec2 instance and run byobu. the status bar infinitty scrooling bug is already filed. [10:24] `hostname` is returning blah instead of blah.domain.tld, I've reset /etc/hostname and rebooted, but every time it reverts to just blah. [10:24] Any ideas? [10:24] vi /etc/hosts [10:24] hostname -f will probably give the full hostname [10:25] how can I use ip route or ifconfig to use another nic and not the default one for an IP? [10:26] ip addr add 192.168.1.1/24 dev ethx [10:26] Oh wait [10:26] route [10:26] and then ip route add outgoing_ip via secondary_nicip? [10:26] You want to route a specific network over another interface? [10:27] I want to use a specific NIC for a specific IP [10:27] so when i connect to the specific IP it uses the specific NIC [10:27] and not the default one [10:28] ip route add via === Lcawte is now known as Lcawte|Away [10:29] tried already, gave me no such process error [10:29] RTNETLINK answers: No such process === Lcawte|Away is now known as Lcawte [10:29] Jeeves_: I've tried setting the fqdn into /etc/hostname, but it gets overwritten on reboot. It's a proxmox virtualised machine, so maybe there's something going on there, not sure... [10:29] /etc/hostname is being overwritten? [10:29] In Ubuntu? [10:31] try /etc/init.d/hostname.sh start after editting === ea1het_OFF is now known as ea1het [13:08] mgw, yes, the tunnel will close when the parent process dies === dendrobates is now known as dendro-afk [15:50] hi, is there someone who can help me finding the way to multiarch on linux? I want to run 32bit game server on 64bit ubuntu [16:29] I installed maas and it gave me my web interface [16:29] when I try firefox, i can't use it [16:30] I'm using this on my virtual machine === Lcawte is now known as Lcawte|Away [16:35] I don't understand why I can't see it on my computer === Lcawte|Away is now known as Lcawte [18:53] There we go, FINALLY resolved my fricken nslcd issues with it not starting up properly at boot. [18:54] Fixed it by converting my dnsmasq, and nslcd init.d scripts into upstart scripts. [18:55] qman__ thanks, that's what I thought [19:32] anyone know maas here [19:32] how come I can't access the web interface [19:32] I have maas installed on my virtual machine [19:33] I can't access the web interface at all [19:33] do I need to import isos first and then access the web interface === Lcawte is now known as Lcawte|Away [20:00] http://cloud.ubuntu.com/ [20:00] I'm looking at this site. I was wondering updating the address within my network [20:00] what is that referring to === Lcawte|Away is now known as Lcawte === erichammond1 is now known as erichammond === Lcawte is now known as Lcawte|Away === Lcawte|Away is now known as Lcawte [22:11] nmap pehden.dyndns-ip.com msg me results please [22:13] Does anyone know of a tool for analyzing the disk usage to indicate large files/dirs? needs to be command line or output some sort of file [22:14] right now im using du -h --max-depth=1 but thats annoying lol [22:15] ah cool Philesight might work, apparently its web based [22:15] slide: start with du -s /* and walk down the tree [22:15] yea but thats REALLY slow heh [22:18] slide: "fast" was not on your list of requirements [22:18] :) [22:18] true heh [22:52] gt5 is good btw === Lcawte is now known as Lcawte|Away [23:09] hello [23:18] !openstack