[00:18] ok I need some quick honest advice [00:19] ok so it need not be quick but honest anyway [00:19] I just bought 8 quad core 3.4ghz 8GB servers for my new rack... [00:19] I have like 30 thin clients currently [00:19] I wanna cluster all these new servers [00:20] how many hours should be a reasonable quote to outsource it to an admin? [00:20] or is this a project I can do myself from a simple readme file? [00:21] In theory this should be like 20x the power I have now, really wanting to do something quick before I lose my drive on it lol [00:21] * tonyyarusso has utterly no idea [00:21] You can probably do it yourself, but it will take a bit more than one readme file to figure it out [00:21] I figured as much... it is possible though? [00:22] one massive computer? like rendering a 2 hour movie in 5 minutes ? [00:22] Yes. [00:23] nobody has any guesses as to what an honest quote would be? I mean if $600 is honest I will just have to accept it [00:23] There are two general approaches you could take. [00:23] ok [00:23] The first would be to declare a cluster of those 8 computers, to always work together. [00:23] k [00:23] The second would be to use "cloud computing" tools to apportion resources as needed for different things. [00:24] like a virtualization technique? [00:24] Yeah, but even more so [00:24] I think I want the first :P [00:24] sounds better [00:24] any native libraries do it? [00:25] not a plug and play type request probably [00:26] i have only found one project working on it that seemed within a hobbyist's reach [00:26] that's why I ask [00:27] hmm I wonder if the ltsp guys might know [00:30] Well, I'm just doing an apt-cache search, and some of interest includ redhat-cluster-suite, openais, gridengine-*, ltsp-cluster-*, mpich2, and related items. [00:30] The LTSP-cluster packages look particularly interesting. [00:30] wow more than google turned up [00:30] or at least google documentation searches [00:30] thank you! [00:31] np - good luck! [00:36] anyone known the command how to detect a soft raid? [00:36] mdadm [01:17] what's the easiest way of setting up an outgoing mail server on karmic server? [01:17] dragon: a satellite, or a smarthost? [01:18] twb: meaning? [01:18] If you don't understand the question, then step #1 is to understand it [01:18] twb: and you're here to complicate things. [01:18] A smarthost is capable of sending mail to arbitrary destinations. Smarthosts are usually called mail.example.net. [01:18] Typically each ISP or organization will have a single smarthost. [01:19] twb: better [01:19] You then have a bunch of "satellites". All they know how to do is send ALL their outbound mail to a designated smarthost. [01:19] So for example a home office might send all its outbound mail to mail.isp.net or smtp.gmail.com [01:19] twb: I'm setting up a smarthost, a single server sending mail directly to the outside world. [01:20] Then you probably want to install postfix, which is the default MTA on Ubuntu. [01:20] Personally I would encourage people to configure satellites where possible, since this is much simpler. [01:21] twb: I tried postfix during karmic-rc-server installation, but it failed to install. [01:21] brb [01:24] back, sorry [01:24] twb: would satellite do what I'm trying to accomplish? [01:24] twb: if so, I'd be happy to go that way. [01:24] dragon: if you have a reliable path to a smarthost, it will. [01:25] For example, if you're setting up a VPS, your VPS provider might have a smarthost [01:25] twb: does an SMTP server qualify? [01:25] That depends what the SMTP server does. [01:26] If it accepts mail from your server (to any destination), and then forwards it on, then it is a smarthost. [01:27] twb: Does Gmail's SMTP service qualify? [01:27] You can have a non-smarthost SMTP server that will accept email from anyone, but only if the mail is for users on itself. [01:27] smtp.gmail.com is a smarthost, provided you have an account there. [01:28] You might need to be careful if you have more than one local user, because I think smtp.gmail.com might rewrite the From header field to your gmail account. [01:28] mhh nice... proftp has leaks [01:30] IMO you ought to use vsftpd or not FTPd at all [01:31] why not ? [01:31] proftpd is nice with sql, but that is the issue atm :) [01:32] Because it's the only implementation that seems to prioritize security, and FTP is definitely a protocol that needs all the security-conscious implementation that you can handle. [01:32] And for authenticated FTP *uploading* I would be using SFTP (i.e. OpenSSH). [01:32] yes true [01:32] I know that, but normal FTP should do [01:46] New bug: #478827 in openldap (main) "openldap database backend back_perl has undefined symbols (aka slapd-perl back-perl)" [Undecided,New] https://launchpad.net/bugs/478827 [01:53] twb: how'd I go about setting up a satellite? [01:53] Install msmtp [01:53] Install msmtp-mta, rather [01:54] Then you basically add a couple of lines to /etc/msmtprc that say "the smarthost is foo.net, and you should claim to be bar.net" [01:54] You can also do this with postfix, of course. [01:55] MenZa: right, the /etc/update-motd.d stuff. well to me it's more of a system thing and less of a user thing [01:56] pmatulis: Well, this is a system thing, I guess... [02:05] !prefix [02:05] As you can see, this is a large channel. If you're speaking to someone in particular, please put their nickname in what you say (use !tab), or else messages get lost and it becomes confusing :) [02:06] so ubottu exists here, I see. [02:07] !hammertime [02:07] Sorry, I don't know anything about hammertime [02:07] dragon derp derp [02:07] ... [02:17] how do I change a server's FQDN? [02:18] Edit /etc/hosts and /etc/hostname, and run "hostname fred", where "fred" is the new value in /etc/hostname. [02:18] twb: thanks [02:18] If you also want to change the external FQDN that OTHER hosts see, you will need to edit DNS records somewhere [02:30] twb: a reboot is actually required in order for the kernel to be aware of the new hostname [02:31] pmatulis: really? Bleh [02:31] pmatulis: where does the kernel remember the hostname? [02:31] I thought it was only handled by hostname(8) being called early during boot [02:32] hostname(1), apparently [02:34] twb: without a reboot you could change the kernel.hostname sysctl setting [02:35] twb: i imagine that's where uname gets it [02:37] pmatulis: hostname(1) sets that, I just checked [02:37] dantaliz1ng, OMG [02:37] Hahaha, I just broke sudo [02:37] I would love to see the kvm/libvirt based upgrade tester [02:37] * poningru is the ubuntu-us-ct contact [02:37] $ sudo hostname arthur [02:37] $ cat /proc/sys/kernel/hostname --> arthur [02:37] $ sudo hostname Clio --error--> sudo: unable to resolve host arthur [02:38] and is going to start hosting upgrade testing sessions here [02:38] Fortunately I can root this box directly. [02:39] and the upgrade test I was scratching my head about [02:40] I finally came up with a procedure of dd'ing the hdd image and then booting into that using vmware to do the upgrade [02:40] how to check on which harddisk is my current directory ? [02:40] ... but seriously if you have that upgrade tester ... I would love to test it out [02:40] maxagaz, mount [02:40] and then see which one it is in [02:41] you can have issues if it is bind mounted though [02:41] New bug: #478855 in bind9 (main) "cannot maintain old config file in bind9" [Undecided,New] https://launchpad.net/bugs/478855 [02:52] New bug: #478857 in postfix (main) "package postfix (not installed) failed to install/upgrade: el subproceso script pre-installation nuevo devolvi? el c?digo de salida de error 1" [Undecided,New] https://launchpad.net/bugs/478857 [02:53] would it be better to have a 4x dual core server at 32 bit (16gb of ram), or a 2x dual core 64 bit machine (4gb of ram)? This would be for rendering with blender 3d [02:54] maxagaz: df . [02:57] jmarsden, :) [02:58] marks256: If I understand the question right, I'd say that 8 CPU cores with 16GB of RAM should be more capable than 4 CPU cores and 4GB of RAM for most tasks... I doubt blender is an exception to that. [02:59] jmarsden, that is exactly what i was thinking [03:00] jmarsden, but the 32 bit vs 64 shouldn't matter much when there is that many more cores and that much more ram? [03:01] Not *that* much. But it is an odd comparison, since any modern machine with 8 cores and 16GB is surely perfectly capable of running a 64bit OS... so why not run a 64bit OS on the beefier machine and get the best of both? :) [03:05] the processors in the beefier machine are AMD Opteron 875's @ 2.2GHz which are 32 bit processors [03:07] Um. I think the Opteron 875 is basicaly an Althon 64 with ECC RAM support... they should be 64bit capable... did you try booting it from a 64bit liveCD? [03:08] I don't even own the machine yet :) [03:09] OK, then at least do some Googling about that processor... [03:09] yep i am right now [03:09] http://www.trustedreviews.com/cpu-memory/review/2005/04/21/AMD-Dual-Core-Opteron-875/p1 is an old review that says it is 64bit for example [03:10] http://www.newegg.com/Product/Product.aspx?Item=N82E16819105079 [03:10] yep newegg says it's 64 bit too [03:10] NEAT! :) [03:10] What makes you think Opteron 875s are 32bit... ? [03:10] Right. [03:11] you konw. i'm not sure why i thought they were 32 bit :) [03:12] OK... problem solved... :) [03:13] thank you much! ;) [03:13] No problem :) [03:52] Is there a way to pipe all incoming email to a domain to a script? [04:04] biscuit_: Yes. Details depend on what MTA you are using, whether you want to use procmail, etc. But yes, it is possible. [04:05] Well I've had issues with postfix so I'm going the exim4 route [04:05] I guess I read a lot of vmail guides and such, but I just want all email sent to a domain forwarded to a script. From there I handle all the filtering, etc [04:06] OK, then I suggest you ask in #exim for info on the specifics, because I'm not very familiar with exim4 :) [04:07] Well I mean, if there's a way to do if with another MTA I'm willing to change ;) [04:10] biscuit_: you could also ask #postfix about whatever bugged you with postfix :-) [04:25] how can I configure postfix to send mail through Gmail's SMTP servers? [04:26] dpkg-reconfigure -plow postfix [04:26] Pick "satellite" and follow the prompts [04:26] twb: thanks [04:29] twb: it didn't ask about smtp authentication - smtp.gmail.com would require that.. [04:30] I don't know how to make postfix do that -- try #postfix [04:30] ok [04:30] http://ubuntu-tutorials.com/2008/11/11/relaying-postfix-smtp-via-smtpgmailcom/ may be relevant [04:31] jmarsden: thanks [04:31] dragon: earlier you asked about host renaming, I think you missed this: 13:18 If you also want to change the external FQDN that OTHER hosts see, you will need to edit DNS records somewhere [04:31] twb: ah I never got that. [04:32] hello [04:32] efremovvk: Welcome to #ubuntu-server [04:32] some problems after upgrades (9.10) [04:33] efremovvk: Please be more specific (and this *is* on 9.10 server edition, right?) [04:33] yes [04:34] i upgraded server from 9.04 to 9.10 [04:34] no problems [04:34] but system installed kernel generc-pae [04:34] after some upgrades it installs server kernel [04:35] but it is no such problem [04:35] after yesturday`s upgrade there problem with MySQL [04:36] after upgrade I reboots my machine once.. [04:37] and after Mysql say "can`t lock socket' [04:37] no any mysql porcesses I found on server [04:37] I starts tasksel [04:38] efremovvk: your native language is Russian? [04:38] but no LAMP there on my system [04:38] yes :0 [04:38] yes :) [04:38] efremovvk: you can also try #ubuntu-ru if you prefer to speak Russian [04:38] I know [04:38] efremovvk: but it is OK to ask in English here :-) [04:39] but may be it is ubuntu-server problem [04:39] i solved it [04:39] but may be some one can`t [04:39] so [04:39] there was no LAMP [04:39] It is fine to ask in English in here. Is that the full exact error message from MySQL? Can you pastebin the output of sudo service mysql start? Did you habe to delete the socket /var/run/mysqld/mysqld.sock ? [04:40] where no /var/run/mysqld/mysqld.sock [04:40] and no /var/run/mysqld/ [04:40] Ok, so how did you solve the issue? [04:41] package mysql-server was deleted from my server [04:41] i installed mysql-server [04:41] but binaries was in my system [04:41] Maybe the upgrade got confused and tried to resolve a conflict by removing mysql-server [04:42] may be [04:43] now apt-get dist-upgrade try to install postgresql [04:43] Yeah, you should use "aptitude full-upgrade" now [04:43] And check that it won't remove stuff [04:43] I think apt-get is still dumb [04:44] hm...now no asks for postgre [04:44] I usually do an "aptitude safe-upgrade" first, which should not install or remove anything, only upgrade. [04:44] Maybe Ubuntu documentation says to do it a different way -- I am too lazy to check. [04:45] A long time ago it wanted me to upgrade via a GUI, so I ignored it :-) [04:45] :) [04:46] I should use aptitude instead apt-get [04:46] kblin: i doubt your still on, but i just got back... no luck, still have no internet conncetion on my desktop [04:46] * efremovvk run away to learn aptitude [04:46] efremovvk: mostly it has the same commands as apt-get [04:49] twb: thanks. apt-get most closely to me [04:57] kblin: scratch that, it's connected, but it's not passing through DNS requests [05:01] kblin: i've gone around the issue tonight (bfcause i need to get actual work done) by setting the DNS on the clients to OpenDNS... but this is something i need to fix [05:02] (anyone else is free to chrip in) [05:07] terinjokes: Is the problem with DNS lookups happening only on desktop machines (in which case, ask in #ubuntu) or on a server (in which case, here in #ubuntu-server is the right place!)? [05:10] jmarsden: the server (the physical box *this* irssi is on) seems to be doing lookups fine... this box is the gateway for the network, and handles DHCP [05:11] OK. So the issue is that if you set /etc/resolv.conf on the desktops to point to the local Ubuntu server for DNS resolution, ... it doesn't "work"? [05:12] tshark time! [05:12] twb: My thoughts exactly :) [05:12] i can try... i just assume that was supposed to be handled by DHCP [05:12] terinjokes: DHCP doesn't do DNS resolution. [05:12] terinjokes: it tells the DHCP client where to go for DNS resolution, though. [05:12] terinjokes: Well, maybe it is... DHCP can provide a DNS server Ip address the client should use. [05:13] I would typically use dnsmasq to provide both DHCP and DNS to client workstations. [05:13] On a client desktop if you do dig @server yahoo.com does that lokup work? [05:13] ok set it to the server's address [05:14] ? I asked what the issue was and you used my question as a possible "answer" to the issue? :) [05:14] jmarsden: i might have misread [05:14] jmarsden: no... cannot resolve [05:15] "cannot resolve"? What exact error did dig @server yahoo.com give you? where server is the name or IP of your server machine... [05:15] allowing it to be handled by DHCP, cannot resolve [05:16] That does not sound like a dig error message to me. [05:16] jmarsden: i ran ping... no dig [05:16] terinjokes: sudo apt-get install dnsutils and then use dig [05:16] "Ping request could not find host yahoo.com" [05:17] jmarsden:which would be excellent... if it was a ubuntu box [05:18] terinjokes: Ok, what client OSes are you using? Can you boot a client from a Ubunut liveCD for testing? [05:18] i'm not asking for tech support on the desktop... from what i see, it's doing it's job... it's a failure on the ubuntu-server [05:18] Win7 [05:19] You/we need to determine what the client machines see when they try to do a DNS lookup... so we need to find a way to get that info. [05:19] nslookup is a DOS command you can use for that, if all your client machines run Windows. [05:21] jmarsden: Server: UnKnown (nt passed one from DHCP?) [05:21] I expect Windows got it when they stole BSD's TCP/IP stack ;-) [05:22] terinjokes: does your DHCP server provide an option 11 value in the DHCPRESPONSE? [05:22] (I think it's 11...) [05:22] terinjokes: When you just type nslookup at a DOS prompt, what does it output? [05:22] jmarsden: an interactive prompt [05:22] Before the prompt... what does it say? :) [05:22] twb: i'm running dhcp3 (likely) [05:22] jmarsden: I would normally do "nslookup google.com" all on one line. [05:23] Default server: something and then Address: something would be conventional. [05:23] twb: this way I can see what DNs sevrer windows thinks it wants to use ... [05:23] Default Server: UnKnown \n Address: fec0:0:0:fff::1 [05:23] OK, there is your problem :) [05:24] If you type in to nslookup the line set server 192.168.1.1 [05:24] or whatever your server's IP address is... and then type in on a separate line yahoo.com [05:24] does it then resolve yahoo.com for you? [05:24] my ubuntu-server is 192.168.3.1 (you mean that?) [05:24] jmarsden: or even just "dnslookup example.net 192.168.1.1" [05:24] s/^d// [05:24] terinjokes: Yes [05:25] twb: which is how *I* know the command :P [05:25] terinjokes: I am assuming the ubuntu-sevrer is running a DNS service of some kind, bind or dnsmasq or whatever, right? [05:25] It's a pity that everyone uses nslookup or dig, because host's output is much more attractive... [05:26] twb: Neither dnslookup nor host exist by default in Windows, though... nslookup does. [05:26] I'll us an existing program over a nonexistent one any time :) [05:26] jmarsden: no... i kinda wanted it to pass DNS to whatever the server was using [05:26] jmarsden: sucks, eh? [05:27] jmarsden: I was just musing [05:27] terinjokes: then you need to tell dhcpd to tell clients to use that upstream DNS server [05:27] jmarsden: im happy to install one, as i planned on it anyways [05:27] terinjokes: Ah. OK! So, what did you tell your dhcp server on ubuntu-server to give out as a DNS sevrer address to the clients? [05:27] terinjokes: they won't magically know [05:28] jmarsden: i did at one point, but getting masquerading working was a fun ordeal [05:28] hey all, i'm trying to do something a little unorthodox i think, but the end result should be nice. i use mpd to play music on my headless server, i have an amp hooked up to the local speakers. i'd like to stream the music to another room, synced as close as possible. i've tried using pulse, but i think i'm running into issues running pulse headless, as it's normally based on per-user sessions. [05:29] terinjokes: That's a confusing answer. On the Widnows machine what does ipconfig /all |find "DNS Servers" output? [05:30] jmarsden: no i do not... happen to know the conf option? [05:30] arrrghhh: the mpd part is straightforward. I can't help with pulseaudio [05:30] arrrghhh: there used to be NAS (network audio sound) for that, but I guess pulseaudio is the new/chromatic/shit replacement.. [05:30] twb, well mpd seems to work OK. i don't know what the best method of streaming sync'd music... [05:30] (I can't help with NAS either, of course...) [05:30] hrm [05:30] lol [05:30] naturally... [05:31] Oh, also, mpd doesn't link against pulseaudio, so of course it will just ignore pulse [05:31] ? [05:31] "option domain-name-servers" [05:31] terinjokes: option domain-name-servers 4.2.2.1 (or whatever the IP of the real DNS server is), in the right place in /etc/dhcp3/dhcpd,conf [05:31] The major downside of pulseaudio is that to add "plugins" you need to write code and recompile the damn thing. [05:31] arrrghhh: try xmms2 -- it's the bloatier, featureful equivalent of mpd [05:31] hrm. is there a better method? [05:32] oh. i love mpd, i was hoping to stick with it. i'll check it out, can xmms2 run headless? [05:32] aptitude install xmms2-plugin-pulse, and go from there [05:32] xmms2 is a daemon, not a GUI [05:32] ah, ok. [05:32] It's just like mpd, only bloatier [05:32] hrm [05:32] ok [05:32] e.g. its wire protocol is a binary protocol instead of simple text [05:32] and what about controlling the music from client machines? [05:33] And supports the equivalent of IMAP's IDLE as well as mpd's shitty polling, and such [05:33] arrrghhh: there are a heap of xmms2 clients -- I think xmms2 is also network-transparent [05:33] ok [05:33] (As far as the client/server part; not the server-to-speakers part) [05:33] arrrghhh: you can talk to #xmms2 about all that [05:33] i love music player minion - plugin for firefox. so long as i can find a client that works well with giant playlists, i'll be ok. mpd+mpm handle my 35,000+ playlist beautifully. [05:34] terinjokes: Did that help? Or even work? :) [05:34] arrrghhh: IME xmms2 is probably better at handling very large playlists, once you have the database cached [05:34] arrrghhh: of course, it doesn't work with NFS, due to sqlite (grr!!!) [05:35] But then if you're using firefox, you won't be able to use NFS either. [05:35] jmarsden: just renewed the windows box... working great (you guys are awesome thanks!) [05:35] NFS? i use nfs for file sharing, is there another nfs? [05:35] Cool, no problem. [05:35] All network filesystems have huge file locking issues, and sqlite (which xmms2 and firefox3 both use heavily) in turn relies heavily on file locking. [05:36] In practice you can probably just say "STFU, xmms2, I'm aware of the risks" and have no problems. [05:36] But I sure as shit wouldn't store any important data in a sqlite database (or ANY database) that's accessed over a network filesystem. [05:36] jmarsden: now... the question becomes... will it work across reboots? :D [05:37] twb: i access a MySQL database over a network [05:37] If you edited the conf file, then yes, it should be permanent. You can restart your dhcpd as a near-equivalent of a server reboot if you want to test it. [05:37] Eh, my opinion of MySQL is unprintable. [05:37] twb, ok. that's fine, i don't think that'll be used. [05:38] jmarsden: i meant the masquerading it took two days of trial and error to get working [05:38] Hmmm. Well, what did you do to make it work? [05:38] What file(s) did you edit, in particular? [05:39] terinjokes: that just means you're not very good at it :-) [05:39] jmarsden: iptables -t nat -A POSTROUTING -o wlan0 -j MASQUERADE [05:39] You ran these commands by hand in a shell, or you put them in a script, or... ? [05:39] (after flush the old rules) [05:40] terinjokes, that is not persistent, you need to either iptables-save or create a script or something [05:40] jmarsden: i didn't really mean to paste it yet [05:40] I forget what the ubuntu way is for that [05:40] qman__: i know... was about to say i *already* had that same line in a iptables load in my /etc/network/interfaces [05:41] #netfilter can help you with iptables [05:41] qman__: probably ufw :-/ [05:41] i'll jump over... thanks again guys [05:41] qman__: debian finally got iptables-persistent a few months ago [05:41] Presumably it'll sync soon [05:42] By 10.04, which is all I care about ;-) [05:42] twb: my college wifi is called "ufw" so i'm trying to advoid that package, lol [05:42] lol [05:42] twb, I just write my own script and add it to dhclient-exit-hooks.d [05:42] yea, i think the new "ubuntu" way of doing firewall changes is with ufw. [05:42] qman__: what, even on servers that don't use dhclient? [05:42] qman__: IMO it belongs in if-pre-up.d, if-up.d, or rcS.d [05:42] twb, on those I use rc.local or if-up.d [05:43] but I get my internet IP dynamically [05:43] so dhclient-exit-hooks.d it is [05:43] Why do you need that information in your iptables-restore script? [05:44] twb: me? [05:44] qman__: why do you need to know your own IP in your iptables-restore script? [05:44] twb, port forwarding [05:44] Hmph [05:45] also, I don't want to be picking up all the random garbage floating on my ISP's network [05:45] there's a lot of it [05:45] twb, so you think streaming with pulse is the best method, and xmms2 will handle it better? [05:46] I have heard an argument that port forwarding is nearly as bad as putting the firewalled host directly on the net [05:46] That you'd be MUCH better off using a simple proxy on the firewall [05:46] I'm trying to remember which one was mentioned for SMTP [05:46] I can't do mail, my ISP blocks it [05:47] The idea being that then the firewall blocks invalid layer 4(?) conversations [05:47] I've got web pages on alternate ports, VPN, SSH, and some other stuff I have to let through [05:48] You don't need to port forward SSH; you could just use -oProxyCommand with a jailed account on the firewall. [05:48] Not that I'm necessarily recommending that... [05:48] I only allow SSH to the firewall from specific hosts [05:48] internet-facing SSH is on another machine [05:49] that isn't allowed to connect to the firewall host [05:49] qman__: and yet you let anyone into the "gooey inner filling" of your network by port-forwarding SSH? [05:49] I need the access [05:49] I mean, I'm not saying that you've made a complete balls-up of it, I'm just saying that recently I have become very suspicious of port forwarding ever being the Right Thing [05:50] it's pretty well locked down, but it comes down to what I need [05:50] Particularly if your bastion isn't some embedded appliance, and can actually run something that needs a good 8MB of memory [05:51] twb, so you think streaming with pulse is the best method, and xmms2 will handle it better? [05:51] the bastion is a 200MHz box with 256MB of RAM [05:51] arrrghhh: yes [05:51] qman__: wow [05:51] it can only handle about 5 simultaneous SSH sessions [05:52] and I've taken care to make sure only the designated accounts can log in [05:52] qman__: dude, my system has 16MB of RAM and 16MB of nonvolatile storage [05:52] That was a "wow, that's big" not a "wow, that's small" [05:53] wow... that's big! [05:54] terinjokes: yeah, I deliberately got the "deluxe" unit [05:55] When there's a decent armel one with a ADSL2+ modem that can be driven by a FOSS driver, I'll upgrade to it and put emdebian down. [05:55] (This is in my imagination, where d-i supports JFFS2 and MTD.) [05:59] qman__: oh, and I wouldn't use rc.local, since that runs AFTER every service is up -- so you have a window where you have services but no firewall.) [06:00] twb, yeah, that's just a dirty hack for when a system is misbehaving [06:00] Fairy nuff [06:03] twb, excuse my ignorance, but what do i have to do to set xmms up? it's installed, looks like the daemon is running... [06:04] arrrghhh: talk to #xmms2 [06:05] sweet... [06:08] i was hoping since you suggested it.... [06:17] twb: works, thanks! [06:17] arrrghhh: I basically got too annoyed and just use mplayer with internet radio now [06:17] http://cybersource.com.au/~twb/Preferences/.bin/radio ? [06:17] Nope, too old [06:17] twb, hrm. ook... [06:18] i have a feeling i'm going to go back to mpd, but we'll see. so far i can't get any client to connect to xmms2 [06:32] twb, so do you know how to get pulse and xmms2 workin? i still think i have an issue with my pulse setup. i can't run pavucontrol or alsamixer. [06:42] I'm having problems having hardy server to recognize all 4 cores. im running a quad core. has anyone come across this issue? [06:43] spartan07, how did you concluded some cores are not being recognized? [06:44] hwinfo >/home/seg02/hardware_info.txt [06:44] shows 1 core [06:44] but cpu as intel quad core [06:44] spartan07, install the package htop and run it. Does it show four bars at the top, like 4 cores? [06:45] I also run htop and it only shows 1 cpu [06:45] lol [06:45] shows only 1 cpu. [06:45] my i7 shows 8 cpus [06:45] im guessing for the threads [06:46] spartan07, please put the result of cat /proc/cpuinfo on www.pastie.org and then paste here the URL. [06:47] http://www.pastie.org/689819 [06:49] spartan07, you pasted the whole output? [06:50] yea [06:50] #00FF14 [06:50] http://www.pastie.org/689821 [06:51] spartan07, Is your system up to date? Do a sudo aptitude update;sudo aptitude upgrade [06:51] there it is again from command prompt to finish [06:51] sudo apt-get update the same? [06:51] spartan07, what do you mean 'the same'? [06:53] the same as sudo aptitude update? [06:54] http://www.pastie.org/689823 [06:54] spartan07, you must do two commands: 1) sudo aptitude update 2) sudo aptitude upgrade [06:54] 2 packages are being held back [06:54] spartan07, you should only do command number 2 when command number 1 is finnished [06:54] yes I was just asking if aptitude and apt-get are any different [06:55] spartan07, slightly different. aptitude is smarter. [06:55] I have tried both and they seem to do the same, so I was asking because I was curious. [06:55] ahh, good to know [06:56] spartan07, is there a good reason for keeping this server running an old version of Ubuntu? [06:56] LTS [06:56] production server [06:57] other than that no [06:57] spartan07, but hardy is not the latest LTS. [06:57] ? [06:58] 8.04 is not the latest LTS [06:59] spartan07, sorry, it is. I thought Hardy was released in 2007. [06:59] maybe a problem with the 64 bit version? [07:00] im running kernel Linux 2.6.24-24-server #1 SMP Tue Jul 7 19:39:36 UTC 2009 x86_64 GNU/Linux [07:00] spartan07, we need to exclude the possibility of outdated software, mainly kernel. If you reboot with the latest kernel of Ubuntu 8.04 and the problem still exists, this will need further investigation. The right place for that is posting all the data you have about the issue on ubuntuforums.org [07:02] spartan07, does aptitude tells you there is a newer version of the kernel for you? [07:05] what command do i run to find out? [07:06] spartan07, you must do two commands: 1) sudo aptitude update 2) sudo aptitude upgrade [07:09] i posted update output on last paste bin url [07:09] 1 sec [07:10] http://www.pastie.org/689830 [07:11] The following packages have been kept back: linux-image-server linux-server [07:12] should I just sudo aptitude linux-image-server linux-server ?? [07:12] spartan07, now do a sudo aptitude full-upgrade [07:12] ahh , ok [07:12] spartan07, oops, sorry [07:12] spartan07, that's not correct. [07:12] lol done [07:12] its down loading [07:12] is there a way to cancel ? [07:13] spartan07, is it downloading only linux-image-server and linux-server? [07:13] full upgrade will take me to 9.01 correct? [07:13] full upgrade [07:13] spartan07, is it downloading several packages? If yes and you're at the download phase, you can abort with CTRL+C [07:13] can I do a cntl+c? [07:14] spartan07, if it is still downloading and not installing, yes. [07:14] done [07:15] Technically it's often safe to interrupt dpkg, too. [07:15] Get:1 http://us.archive.ubuntu.com hardy-updates/main linux-image-2.6.24-25-server 2.6.24-25.63 [17.8MB]58% [1 linux-image-2.6.24-25-server 13738342/17.8MB 77%] [07:15] I wouldn't depend on that as a strategy, though [07:15] KurtKraut: thats where it stopped. [07:16] should I just sudo aptitude linux-image-server linux-server ?? [07:17] twb, there is a newer version of linux-image for spartan07 but aptitude is not making the update after 'sudo aptitude safe-upgrade', it tells the package was kept back. What does spartan07 need to do to have an up to date kernel? [07:19] KurtKraut: that's because it'll want to remove linux-generic or linux-kernel-generic or something [07:19] Try a full-upgrade and pay attention to what it wants to remove [07:19] twb, but he wants to stick to an LTS version. [07:19] spartan07: run "aptitude full-upgrade -sy" and pastebin the output. [07:20] KurtKraut: if he hasn't fucked with sources.list, it won't upgrade out of LTS [07:20] twb, oh, okay. [07:20] spartan07, so, I was right before :P [07:20] KurtKraut: this is a known issue (to me, at least) with LTS -- the security update for linux server kernel didn't include an update to one of the metapackages [07:20] So either remove the metapackage (which is what I do), or don't apply the kernel's security update [07:21] -s will simulate a full-update and let us confirm that this is the problem. [07:21] http://www.pastie.org/689835 [07:21] twb, let me explain you his problem. He is running Hardy on a Intel Core 2 Quad and Ubuntu doesn't seem to recognize all cores. Only one. Before proceeding on an investigation, I'm telling spartan07 that it is important to have the latest kernel availuable for him [07:22] spartan07: pastebin the output of "cat /proc/cpuinfo", too. [07:22] http://www.pastie.org/689837 [07:23] spartan07: yeah, that's perfectly fine to do an "aptitude full-upgrade" [07:23] It's just whinging because the kernel's changing a minor version, which means that a "new" package is being installed. [07:25] is this core recognition a known issue? I would think many servers are multi core now a days [07:26] done, rebooting - 1sec [07:26] spartan07: I don't know. [07:27] processor : 3 [...] [07:27] model name : Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz [07:27] My 8.04 server can see four cores. [07:27] Linux plum 2.6.24-23-openvz #1 SMP Thu Apr 2 00:25:58 UTC 2009 x86_64 GNU/Linux [07:27] I suggest you ask the BIOS if it is doing anything stupid [07:27] that was my next thought "BOIS" [07:28] *BIOS [07:28] when I checked it said all 4 cores were enabled [07:28] I would not be able to check since im doing this all though ssh until morning [07:29] spartan07: check tomorrow, then [07:29] Someone will still be in here, even if I'm not [07:29] still showing 1cpu so it will have to wait till morning. [07:29] thank you for the support guys [07:29] I hear you can also use dmidecode to check BIOS settings [07:31] http://www.pastie.org/689841 [07:32] Under Processor Information it says 4 and enabled [07:32] spartan07, so after upgrade and reboot, /proc/cpuinfo still shows one core? [07:33] yea [07:34] spartan07, have you searched on ubuntuforums.org if someone had the same problem? [07:34] im kinda thinking of just trying out 9.1 64bit [07:34] yea did not find much under the server part, hardy desktop had this issue with some laptops [07:35] spartan07: you could just cherry-pick the 9.10 (it's not "9.1") kernel and see if the problem is fixed. [07:35] spartan07: that will save you a tedious and possibly useless install [07:35] in all honesty I would not know what to cherry pick [07:36] what would be the command to upgrade on the command line to 9.10? [07:37] !8.04 -> 8.10 [07:37] Sorry, I don't know anything about 8.04 - [07:37] !upgrade [07:37] For upgrading, see the instructions at https://help.ubuntu.com/community/UpgradeNotes - see also http://www.ubuntu.com/getubuntu/upgrading [07:37] !release notes [07:37] Sorry, I don't know anything about release notes [07:37] spartan07: have a look at that stuff and get back to us if you need more help [07:37] My advice regarding upgrades will be rather baroque and draconian [07:38] ok, Im on it. thank you very much for the help [07:38] lol [07:38] ok, gnight guys and thanks again [08:31] New bug: #478973 in freeradius (universe) "expiration functionality of FreeRADIUS 2.1.0 is broken" [Undecided,New] https://launchpad.net/bugs/478973 [09:19] Hi. I have a question about Ubuntu Enterprise Cloud. Imagine I have 2 machines for the local cloud. Will I be able to start an instance that uses the resources of both nodes? or must they be 2 or more to use all the resources? [09:41] Is there anyone here that have experience with using Tomcat and MySQL on ubuntu? [09:43] !anyone >sindre [09:43] Stupid ubottu! [09:43] sindre, please see my private message [09:43] Good ubottu! [09:43] The problem is that I have a tomcat web developer who have some issues connecting to the mysql db he created. We attempted to run the same application from his desktop machine, where we created a mysql user with permission to connect from his ip. this worked. So I believe there is problems with the jdbc application [09:43] sorry [09:44] He has tomcat installed on his desktop computer, which were able to run his tomcat code without trouble (connecting to the database on the ubuntu server). But when we attempt to run the same web application from the ubuntu server, the connection to mysql fails [09:46] MySQL and Tomcat from the 9.10 repositories [09:46] I also installed libmysql-java [09:47] telnet localhost 3306 from a ssh session on the server works fine. [09:47] As for me, I can't help. [09:47] Thanks for replying and helping me help others help myself anyway [09:47] :) [09:51] sindre, I don't know anything about tomcat, but it might be just a regular old mysql issue [09:51] in mysql, 'user'@'localhost' and 'user'@'1.2.3.4' are two different users [09:52] so make sure you have the right user setup [09:53] I'll check it. thanks [09:53] Goddamn mysql localhost hack [09:53] Or is it "127.0.0.1"? [09:53] Whichever one magically uses sockets [09:53] not sure, but those are two different users as well [09:53] I think localhost uses sockets [09:53] I hate that [09:54] I have 'sindre'@'localhost' (same username for the external IP I tried, and it worked fine) [09:55] did you set the password for both users? [09:55] and the permissions [09:55] yes [09:56] ok then, I see no reason why it shouldn't work [09:56] I think it's the java side of it [09:56] kind of ruled out MySQL when it worked with same MySQL installation on a different tomcat server [09:56] very well could be, like I said, I don't really know anything about tomcat [09:56] sounds like mysql is set up correctly though [09:58] I find it very difficult to google issues with tomcat. Been struggling with this since last week (I'm just a sys admin who tries to make the server work) [10:05] if a bug in launchpad has "fix committed" status, is ti completely fixed in updates? [10:05] alex88: No. [10:05] alex88: Which bug? [10:06] https://bugs.launchpad.net/bugs/446146 [10:06] Launchpad bug 446146 in linux "Several Huawei USB dongle don't work with kernel 2.6.31-12.40 (2.6.31.1 update related)" [High,Fix committed] [10:08] maybe it's fixed in some rc kernel..but i'm not so familiar wth bug system.. [10:16] soren: any clue? [10:16] alex88: It's probably in karmic-proposed. I can't see for sure. [10:17] soren: thanks, i'll try those updates [10:17] Have fun. [10:17] thanks..xD [10:17] Remember to disable karmic-proposed again once you've installed the kernel. [10:18] You may not want to just blindly accept all the stuff in -proposed. [10:18] http://pastebin.org/52079 here is a summary of my problem. [10:18] If only target-release or similar could be used for that kind of pinning :-/ [10:19] i'll do it..thanks [10:20] twb: no familiar with tomcat.. i'm sorry... [10:28] updated pastebin: http://pastebin.org/52093 [10:33] sindre: you should disable TOMCAT6_SECURITY in /etc/default/tomcat6 and see if it changes something [10:33] sindre: on Ubuntu tomcat runs by default with a security manager, while the upstream binary distribution of tomcat doesn't [10:43] ttx: I'll try it. thanks [10:45] it's commented out in the default tomcat6 configuration file [10:45] sindre: that's because it defaults to yes. [10:45] set TOMCAT6_SECURITY=no [10:45] (it's not as bad as it looks) [10:46] did it and restarting tomcat [10:46] ttx: you are a brilliant man! [10:47] sindre: I cheat. I packaged it. [10:47] thanks for giving me an easy way to install tomcat and for helping me figuring out the issue I was having. :) [10:48] sindre: If that works, it means you'll have to hack the securitymanager profiles in order to make it work with SECURITY=yes [10:48] or just drop the securitymanager altogether [10:48] since it's very coplex to get right. [10:48] It's noe exposed to the internet, so I'll be ok without security [10:48] complex, even [10:48] "security" is just about running the JVM inside a security sandbox and giving it rights to do stuff with the system [10:49] so it's another layer for defense in depth [10:49] but it's tricky to configure and the main cause of strange bugs with our packaging [10:50] Is securitymanager some Java-specific thing? [10:50] twb: yes [10:50] it's a jvm option [10:50] Right, carry on [10:50] * twb doesn't do Java [10:50] we already run tomcat6 as a specific user, so it's secure enough for most cases [10:51] twb: good idea, I shouldn't do Java either. [10:51] I was going to point half-heartedly at apparmor and make bowel-movement type noises [10:52] twb: we could replace using securitymanager by default with using apparmor by default [10:52] twb: it's so much easier to configure :) [10:53] I'm kinda suspicious of all the MACs, particularly accidentally allowing stuff [10:53] also tomcat has security manager support but upstream doesn't like it so much -- they usually talk bad words about Ubuntu for us using it by default [10:53] But also because every damn admin says "oh, selinux? I `configured' it by putting it into complain mode" [10:54] (Learn mode? Whatever. The one that doesn't enforce policy.) [10:54] twb: complex MAC is usually worse than no MAC, unless in the hands of very experienced people [10:54] twb: that's the whole story behind Ubuntu's backing of Apparmor [10:54] ttx: eh, my understanding of apparmour is that it's a slightly easier flavour of suck [10:55] twb: my experience is slightly better. [10:55] I'd like to understand NTFS' ACLs so that I can explain why they're dumb, too [10:56] We currently have some amazingly fugly layers in the security cake at some prisons [10:56] But yeah, I will grant you that apparmour is easier to get going than selinux [10:58] By 2100 I hope to have gotten around to some turnkey "default deny" apparmor policies, like I currently know how to do for packets with netfilter. [11:01] Hm. Wikipedia says TOMOYO (another path-based MAC) is in Linux proper. [11:02] Never mind, it's just using LSM like everything else. [11:16] I'm having Kerberos trouble. I can kinit, but I can't su with the Kerberos password, and I get this message: [11:17] Nov 9 06:15:14 server1 su[3889]: Libgcrypt warning: missing initialization - please fix the application [11:17] in syslog. [11:17] Oh gods, I hate krb [11:17] toddobryan: I assume your /etc/pam.d is sane-ish? [11:18] Well, I ran sudo auth-client-config -a -p kerberos_example [11:18] So if it's not, it's auth-client-config's fault. [11:18] But I looked in there and it seemed to make sense. [11:19] toddobryan: good-o [11:19] My last attempt predates auth-client-config, but I have had great success with it's lac_ldap [11:19] lac_example? [11:20] Any idea what the libgcrypt problem could be? [11:20] That seems to appear whenever I try to su... [11:21] Call me silly, but any reason you're trying with su and not login or sudo or ssh? [11:21] Not that I particularly expect those to work better [11:33] Well, I don't know what's different, but I used sudo pam-auth-update krb5 and it works, now. [11:34] So something about auth-client-config and pam-auth-update is inconsistent with each other. [11:34] But now, I have to run to school. Thanks, twb! [11:35] I feel old [11:36] I will pretend that he's a lecturer at school [11:36] Then I will not feel so old === sleeping`dragon is now known as error404notfound [13:16] New bug: #475546 in krb5 (main) "Memory leaks in version 1.6.dfsg.3~beta1-2ubuntu1.1" [Low,Incomplete] https://launchpad.net/bugs/475546 [13:36] New bug: #477011 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5 failed to install/upgrade: installed pre-removal script alt i?lemi ??k?? durumunda hata d?nd?rd?:: 1" [Low,Incomplete] https://launchpad.net/bugs/477011 [13:41] New bug: #476142 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 (not installed) failed to install/upgrade: " [Low,Incomplete] https://launchpad.net/bugs/476142 === dantaliz1ng is now known as dantalizing [13:51] New bug: #478415 in vnc (main) "login thru ssh then start vncserver causes endless create and destroy processes with no names" [Undecided,New] https://launchpad.net/bugs/478415 [14:01] New bug: #477333 in sysstat (universe) "cron job is overly noisy" [Wishlist,Confirmed] https://launchpad.net/bugs/477333 [14:05] My Ubuntu Server 8.04.3 located at http://liliput.mine.nu is serving very, very slow, even accessing from my local network [14:06] I did an ApacheBenchmark and that seems to be fast, so it must be something with the connection [14:07] when I ping the server with big packet size (like 1024 bytes) I get timeouts all the time, but 32 bytes it replies ok most of the time [14:09] But I wgetted some external big files (like a ubuntu iso) from the server through SSH and I get full download speed for my connection (10mbit), so at least downstream seems to work OK... [14:09] any help would be appreciated :) [14:21] New bug: #479185 in bind9 (main) "bind 9 won't install" [Undecided,New] https://launchpad.net/bugs/479185 [14:43] ttx: ping [15:01] zul: pong [15:02] ttx: mind if I take the openvpn merge from you? [15:02] zul: be my guest [15:23] Hello :) I have an question to ask; Can I run Plesk on an Ubuntu server? [15:29] according to http://www.parallels.com/fr/products/plesk/reqs , you can on 8.04 [15:30] Thanks [15:31] New bug: #479226 in likewise-open5 (universe) "Cannot authenticate using PolicyKit dialogs with domain credentials, prompts for local user" [Undecided,New] https://launchpad.net/bugs/479226 [15:32] Ubuntu 8.04 LTS <- plesk can run on that distro. [15:34] that's what they say [15:35] But? [15:35] okay [15:35] i don't use that, i can't tell you [15:35] It's okay [15:36] It's was just a question ;) You are using DirectAdmin, lenios? [15:42] i don't [15:42] but if i had to choose one, i'll probable check webmin first [15:44] Okay :) [15:44] how can i measure the performance of my DNS server in terms of response time [15:56] incorrect: What dns server are you running? [15:56] Some report it themselves [15:56] Jeeves_, i am not running the dns server [15:57] i am comparing dns performance from the UK / USA where ever [15:58] incorrect: Ah, ok. Well, dig reports the Query time [16:01] New bug: #479250 in mysql-dfsg-5.0 (main) "ORDER BY DESC in InnoDB not working" [Undecided,New] https://launchpad.net/bugs/479250 [16:01] what is the difference between sendmail.cf and sendmail.conf and is there a channel for more on topic conversation [16:02] Hey, that's my bug! :) [16:05] hey! [16:05] googa: I guess not much people use sendmail in here (they are wise, imho ;)) [16:06] I guess not [16:06] hey [16:06] im pretty new to the irc scene [16:07] and a bit bewildered, is there any way of searching for specific channels? [16:09] Not that I know of [16:09] googa: depends a bit on your client. /list should give you a listing of all the channels on the server though [16:10] you could try searchirc.com [16:10] googa: http://www.irchelp.org/irchelp/chanlist [16:13] ok thanks! [16:14] great found one! [16:14] oh yes dig does report the query time, i couldn't see it [16:16] ttx: why is some of the ideas in italics? [16:17] zul: see top [16:17] well duh [16:29] nxvl: Any chance you could look into Bug #413252 ? Looks very straightforward, but I think it ought to be an SRU for Karmic too. [16:29] Launchpad bug 413252 in courier "package courier-base 0.61.2-1ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 2" [High,Triaged] https://launchpad.net/bugs/413252 [16:36] hi folks, I've just installed ubuntu-server on a box and noticed upon boot thats its not loading a usb wifi adapter, what steps can I take to get it working? [16:36] (the wifi is detected and works fine in karmic desktop) [16:38] * soren calls it a day [16:43] I'd like to use the ubuntu mail stack: postfix/dovecot with procmail. Can I just redefine postfix's mailbox_command line? [16:49] <\sh> mok0, why do you want dovecot + procmail, while you could use sieve? [16:53] can anyone help me register my UEC with Landscape? The machine itself is in Landscaper, but when I try to register my cloud it says "There was an error communicating with the cloud.: [16:53] <\sh> mok0, just read http://wiki.dovecot.org/LDA/Sieve and you forget about procmail at all ;) and there is an procmail2sieve converter [16:53] I have a question regarding .htaccess on my server. This is a example of my htaccess file http://pastebin.com/d360c4244. What I'm trying to accomplish is to remove trailing /'s, redirect http://www. to just http://, prevent hotlinking, and remove php file extensions. For example http://www.mysite.com/contact.php/ would become http://mysite.com/contact. My example works except for two cases, when I want to go to a directory like my [16:53] site.com/images/ and when I try to link to my forums located at mysite.com/forums. /forums gets rewritten to /home/me/..... Any ideas? [16:53] \sh, I have a procmail setup that I've used for 10 years and I'd rather not switch [16:54] \sh does sieve work from the default setup? [16:55] <\sh> mok0, it should...I have an ISP setup with postfix + dovecot so I don't use ubuntus default setup [16:55] \sh, OK, thanx. [16:56] I will look at sieve, if I can convert my (huge) .procmail script I might go with that [16:57] <\sh> mok0, as written on this wiki page, there is a procmail2sieve converter...never used it myself, but if it can help you, try it :) [16:57] \sh, I will [16:57] Is there a better place to ask my question? [16:58] jimlovell777: Perhaps #httpd for apache questions [16:58] Pici: Ok thanks [16:59] !apache | jimlovell777 [16:59] jimlovell777: LAMP is an acronym for Linux-Apache-MySQL-PHP. However, the term is often used for setups using alternative but different software, such as Perl or Python instead of PHP, and Postgres instead of MySQL. For help with setting up LAMP on Ubuntu, see https://help.ubuntu.com/community/ApacheMySQLPHP - See also the Server CD installation process (different in Edgy+) [17:09] I have a server connected by serial cable. When it boots it never reaches the stage to ask my user name, however it does get ssh server started so I ssh into it. I'm trying to work out why it wont get to the stage of logging in, any ideas how i can? [17:10] chimp_: the start up logs are usually a good place to start; syslog, messages, dmesg [17:11] chimp_ -- so you are using something like console=ttyS0 to boot the server? [17:11] hence, you see the kernel output on the serial console, but the serial console never issues a login prompt? [17:11] yes [17:11] ok. so you don't have a getty process running for the port [17:11] syslog is doing this [17:11] Jan 1 02:29:51 ubuntu init: tty3 main process (4423) terminated with status 1 [17:11] Jan 1 02:29:51 ubuntu init: tty3 main process ended, respawning [17:11] repeatedly [17:11] still.. [17:12] Its on an arm board [17:12] huhm. [17:12] tty3 would be the third virtual console, right? [17:12] its running through tty1 to tty6 terminating and respawning [17:12] huhm :| [17:12] wonder why it's chunking on a simple virtual console [17:12] is there.. no video at all on this machine? [17:13] Nope [17:13] It has just the serial out [17:13] ok.. you might start by removing tty?.conf from /etc/init [17:13] but first, copy one to ttyS0.conf [17:13] edit it, change ttyx to ttyS0 [17:13] n boot [17:14] (that's assuming you are using ubuntu 9.10) [17:14] 9.04 [17:14] .. before 9.10 the usual /etc/inittab file is still used.. but 9.10 switched to upstart and hence /etc/init/.conf files instead [17:14] hmm lemme look at my 904 machine [17:15] # Example how to put a getty on a serial line (for a terminal) [17:15] # [17:15] T0:23:respawn:/sbin/getty -L ttyS0 115200 vt220 [17:15] Is the end of inittab (rest is commented) [17:16] makes sense.. so standard inittab syntax [17:16] hmm, my 904 machine does seem to have /etc/inittab *confused* [17:16] (would be nice if ubuntu stuck with something for more than 6 months...) [17:17] ok, people.. so where the bleep is the inittab in 904? [17:17] Mine is as you said /etc/inittab [17:18] hmm, that doesn't exist on mine [17:18] maybe its different on the amd64 build [17:18] looks like /etc/event.d for my install.. strange. [17:18] Mine is one generated using debconf [17:19] anyways, you should be able to remove the tty1-tty6 lines from inittab so you stop erroring on them [17:19] and I'd guess that above ttyS0 line should be what you need to spawn a getty on the serial line [17:19] assuming ttyS0 is the valid device, AND it supports 115.2kbit [17:19] But it never spawns it [17:19] soren, can you make me the assignee of https://blueprints.launchpad.net/ubuntu/+spec/server-karmic-ec2-release-process ? [17:19] Yeah it does [17:20] what if you run it manually via your SSH sesssion? [17:20] might spit you a useful error :) [17:21] The rest (tty1-tty6) are commented in inittab [17:21] running it manually just gave me access [17:21] haha [17:21] i can log in via serial now, wtf? [17:22] Is the inittab bit never ran? [17:22] ahh. so tty1-6 shouldn't be spawning in the first place if it's using /etc/inittab ?? [17:22] thats my guess. [17:22] do you have /etc/event.d ? [17:22] yep [17:22] my 904 install.. has no /etc/inittab [17:22] but.. looks like my getty's are spawned from there [17:22] maybe try editing /etc/event.d/tty1 [17:22] and change tty1 so ttyS0 on the getty line [17:23] err.. exec /sbin/getty with the same params that worked from the command line [17:23] lol, that is a real hack [17:23] looks like the last line of the file is just an exec [17:23] well i suppose you could.. nuke tty1-6 files.. create ttyS0 file, etc.. that'd be "proper".. lol. [17:23] I wonder if its related to this arm board running jffs2 on a 512mb nand [17:23] It takes *FOREVER* to boot [17:23] Im having trouble with ubifs for time being [17:24] hmm. well, dmesg output should be timestamped [17:24] so you can see where it's hanging up [17:24] I get a giving up message on bootup saying mount took longer than 120s [17:24] wow, insane. [17:25] yeah its jffs2 [17:25] will the device cooperate with a most standard filesystem? [17:25] err mORE [17:25] It reads the entire disk before mounting [17:25] * NRVate giggles [17:25] It needs to be a nand filesystem [17:25] It needs ubifs tbh, but im struggling to get it working [17:26] It was jaffs2 before but I also had issues with no one seemingly using it [17:26] huhm.. ive yet to play much with all that [17:27] Basically theres no controller chip that does the wear leveling for you that you find in sd cards or ssd drives, so the file system has to handle it for you [17:27] makes sense [17:27] ubifs looks real new [17:27] along with logfs.. interesting.. haven't seen this before [17:28] so.. ubifs came out in 2.6.27 [17:28] 9.04.. uses 2.6.28, right? [17:28] wonder if there are significant fixes in the later kernels you may need? [17:28] Yeah too new.. the tools for it dont exist on the debian sarge install that comes with this board, so im struggling to set it up to use ubifs for the newer filesystem and kernel i run on it [17:28] Im running 2.6.31 on it [17:28] ah, gotcha. [17:29] It has some patches in it that some bloke wrote to make it read the nand faster [17:30] right on. i assume the root fs is on the nand device? [17:30] yep [17:30] I cant make it boot from the sd card because of some hard coded rom issues [17:30] hehe that complicates things a bit [17:31] Im trying to cross compile mtd-utils but ive had issues [17:32] and mtd-utils only has ubifs support in 9.10 (while im 9.04) [17:32] right im off to see my GP cheers for the help NRVate [17:32] whew.. sounds fun! [17:33] good luck chimp, sounds like a fun project [17:33] i've been wanting to play with some ARM stuff [17:33] keep seeing their office on the ride home haha [17:33] cheers :) [17:34] Where, london? [17:34] ScottK: will take a look [17:34] Austin, Tx [17:34] nxvl: Thanks. === dendro-a` is now known as dendro-afk [18:12] Hello, I chose to configure networking later during 9.10 ubuntu server install. After install and bootup networking init script simply doesn't work. Are there any further steps besides configuring /etc/network/interfaces? Ifup eth0 works... === orudie_ is now known as Pavel_ [18:17] New bug: #479365 in squid (main) "assertion failure" [Undecided,New] https://launchpad.net/bugs/479365 [18:25] anyone? [18:25] :S [18:27] mephx: if you have some log output of how it doesn't work, that would be very helpful [18:30] kane_: i can't find any relative output regarding what's happening [18:31] ifup works, /etc/init.d/networking restart clears the current configuration and does not re-set it [18:31] mephx: odd -- did you edit that file at all? [18:32] the init script? [18:32] yeah [18:32] no [18:32] i installed 9.10 server on 2 servers [18:32] simultaneous [18:32] mephx: and they both have this problem? [18:33] one of them got network configuration, the other one i chose to configure it later [18:33] the one configured during install worked, the other doesn't [18:36] I could always add ifup eth0 to local, but that looks $#itty [18:36] mephx: /etc/init.d/networking basically does an ifup/ifdown. try capturing the output and see what it does [18:39] mephx: another thing to look at would be the runlevels this is being executed at [18:39] also, it does not show on rcconf [18:40] networking [18:41] can't neither get it to output anything [18:41] mephx: you can't get /etc/init.d/networking to output anything? === Authority_ is now known as Authority [18:42] kane_: i'm now debugging it by hand [18:42] ifup -a differs from ifup eth0 [18:42] mmm [18:43] lol [18:43] lacking auto eth0 line [18:44] ifup -a sucks [18:44] thanks for the headsup [18:45] don't know why, but i always associated auto line in ifaces to dhcp [18:45] mephx: you did the hard work :) glad we got it fixed [18:46] thanks again kane_ [18:46] my pleasure === WALoeIII_ is now known as WALoeIII [19:14] hi :) [19:14] guys.. does anyone of you has a luck with Xen and Karmic ? [19:16] .. [19:16] :) [19:32] I'm looking at the Model API reference... is there something like an afterLoad and afterUnload callback? [19:32] Oops... wrong channel.. lol [19:36] New bug: #479428 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5 failed to install/upgrade: subprocess new pre-installation script returned error exit status 128" [Undecided,New] https://launchpad.net/bugs/479428 [19:43] New bug: #284416 in whois (main) "whois package contains unrelated binary 'mkpasswd'" [Low,Fix released] https://launchpad.net/bugs/284416 [20:02] hi there [20:02] does anybody know, if ubuntu is certified for oracle (db)? [20:20] spy6, it's not [20:23] are there any scripts /window 3 [20:24] sammy: What? [20:24] are there any ways to view the status of multiple init.d services at the same time? [20:24] gentoo has things like rc-status that show you the started/stopped status of all services or all services in a particular run level [20:25] sammy: You could try server --status-all , but I'm not sure if thats the recommended way to do it. [20:29] if you mean `service --status-all` :) that seems to give me what I'm looking for. I'm okay with a complete list of all init scripts- its not terribly important that I be able to filter it by current run level. thank you! === dendro-afk is now known as dendrobates [20:37] New bug: #259426 in autofs (main) "autofs shoud be started after slapd" [Low,Confirmed] https://launchpad.net/bugs/259426 [21:07] Is http://www.symfony-project.org/book/1_2/08-Inside-the-Model-Layer available for doctrine? [21:07] Wrong chat >.< [22:43] "A friend" needs some advice on letting Windows users have a single logon no matter which computer on the network they are using. Is there an Ubuntu service which can be used to do this LDAP-like thing for Windows? Is there a better place to ask? [22:46] vmbuilder jeos runs successfully, but when I start the VM the guest hardware cannot find the harddrive.... [22:46] erichammond: windows calls this a domain (active directory) [22:48] hi where i can download netboot file for ubuntu desktop...? [22:48] Greetings, channel. [22:48] fro pxe server [22:51] Might anyone have any ideas or tips for removing extra services and reducing the amount of code on an Ubuntu server running Apache, MySQL, PHP & friends? [22:52] I'm looking to lighten the footprint and increase security. [22:53] mooselix, I would look into kvm+JEOS [22:54] Just Enough Operating System [22:54] Its a version of ubuntu that is tiny and works well with VM. [22:55] fxhp: Thanks for the "active directory" tip. I'm finding howto's but they mostly seem to describe how an Ubuntu client can connect to a Windows server. I'll keep looking. [22:56] What are you trying to do? [22:56] Build a ubuntu server that allows for windows computers to authenticate with it? [22:56] That would be LDAP [22:59] Why does my jeos VM no have a harddrive? [23:02] fxhp: Yes, authentication, and if possible, loading browser cookie/bookmark settings. Looks like we're going to have to bring in somebody who knows Windows a bit more than we do. [23:04] LDAP is used for authentication, for browser settings you would need windows "Group Policy" which is part of active directory... I don't think that functionality is part of ubuntu [23:06] New bug: #479621 in eucalyptus (main) "Re bundled karmic image fails to boot" [Undecided,New] https://launchpad.net/bugs/479621 [23:08] Does anyone here have experience with vmbuilder and jeos? I need help. [23:14] Heres an interesting question can I, and what happens if I overwrite the kernel while running the os (my guess is maybe the kernel is entirely in ram and it wont notice) [23:14] This is for an arm board [23:14] Going to overwrite the nand and update the kernel while running [23:14] :) [23:15] Kernel is in its own parition [23:16] is there an issue with software raid and 2.6.31? === kane___ is now known as kane_ === erichammond1 is now known as erichammond [23:58] Still having trouble with my VM's seeing their harddrives [23:58] If anyone has a chance to help me, that would be great.