[00:10] kirkland: there? [00:38] cloakable: That's almost certainly not true. [00:40] ScottK: Well, for the purposes of the ioncube loader. [00:40] It wouldn't suprise me if it was all Free software, but all GPL is highly unlikely. [00:41] possibly [00:42] Tine 2.0, WordPress, Dokuwiki, RoundCube, Status.net :) [00:51] Why aren't server tasks (such as mail server, lamp server, VM host, etc.) documented more thoroughly? I assume those are the "easier" aka more preferred methods of service installation, yet the Server Guide shows traditional apt-getting packages. [00:51] Am I missing something? === sheenobu is now known as criminy [01:01] lukehasnoname: Maybe you want tasksel? [01:02] * jpds is fine with traditional apt-get. [01:10] let me clarify my question to make sure it's valid: When using tasksel vs grabbing all the packages manually, are the configurations of the services different? Would tasksel provide options or configurations not offered by std installation? [01:23] lukehasnoname: no [01:36] pmatulis, I guess it's not as much of an issue, then. Still, I see little mention of tasksel in the server guide, which surprises me, given how useful it can be (esp to intro admins) [01:38] lukehasnoname: if tasksel can be applied to topics in the server guide then, yes, i agree with you. [01:43] lukehasnoname: feel free to file a bug [1] or get involved [2] [01:43] [1] https://bugs.launchpad.net/ubuntu/+source/ubuntu-docs/+filebug [01:43] [2] https://wiki.ubuntu.com/DocumentationTeam/SystemDocumentation [01:44] Sounds good. [02:17] what would be the best soulition to build a WiFi hotsopt access server? === unreal_ is now known as unreal [03:11] anyone here have any experience with ioncube loaders? [04:59] I just watched Toy Story for the first time [04:59] Another reason to like Debian [05:06] lukehasnoname, how so? [05:13] scar, It was a fun movie and I think it's cool the Debian releases are named after characters in a movie [05:59] i want to run nginx on boot (lucid). i am reading here https://help.ubuntu.com/community/UbuntuBootupHowto that 'Ubuntu is transitioning away from the sysv init system'. what do u guys think? [06:01] and if the answer is: use init script i have another question - will this script run on lucid? (it was tested on karmic) -http://code.google.com/p/nginx-init-ubuntu/ [06:11] oren: The last Ubuntu release to actually use sysv was Dapper (4 years ago). Since then it's all been emulated by upstart. [06:12] oren, I believe you can still use your sysv init script. But if you want to be really cool you'll learn upstart scripts [06:13] except upstart scripts maight look completely different in 10.10 or later, but yeah ;) [06:14] JanC, augh [06:17] Now that Mako and Kyle Rankin are rich from the Ubuntu Server Book, having it open sourced and quasi-open for edit online would be incredibly useful [06:18] That book is so very awesome [06:21] ScottK: lukehasnoname` i never used those scripts before. can i simply take the nginx script (my link) , make it executable and locate it inside init.d, and run: sudo update-rc.d myscript defaults 98 02 ? [06:22] (that's the instruction i see on ubuntu docs) [06:22] and what is this book you guys are talking about? [06:23] oren, The Official Ubuntu Server Book [06:23] written by two Ubuntu/Linux guys [06:24] It's not terribly useful for writing init or upstart scripts (mostly because Upstart's API changes daily, it seems) but it has good instructions for a lot of common tools and services in Ubuntu [06:26] lukehasnoname`: so my steps above are accurate? and where can i get this book? [06:28] effing IRC boots me all the time and I get these backup names... anywho oren it looks right, judging from the man page for update-rc.d [06:28] never used it myself [06:30] ok. i'll try it. thanks === mushroomtwo is now known as mushroomblue [06:38] oren: docs for writing upstart scripts for karmic & lucid can be found in init(5) [06:38] Daviey, testing the kernels now, i think it broke between 2.6.32 and 2.6.33 [06:39] JanC: do u mean /etc/rc5.d [06:39] i appologize for my ignorance..not sure what init(5) means [06:40] i know there are run levels where each folder has scripts that are executed. [06:40] rc0.d, rc1.d etc [06:41] init(5) means the man page for init in section 5 of the manual pages. So do man 5 init to read it. [06:41] hi [06:41] can some one please help me about this post https://lists.ubuntu.com/archives/ubuntu-server/2010-June/004345.html ? [06:44] oren: what jmarsden says, or http://manpages.ubuntu.com/manpages/lucid/en/man5/init.5.html if you want to read it online (that site has all Ubuntu manpages) [06:47] JanC: jmarsden awesome [06:47] JanC: You're welcome [06:53] is this the ubuntu server book you guys praised? http://www.amazon.com/Official-Ubuntu-Server-Book/dp/0137021186 if not, what is the best server book i should buy? [06:53] only 7 reviews but they all liked it [06:53] Daviey, headed to bed now, it appears to be 2.6.33 that broke the image registration [06:54] Daviey, so my original idea was right but i am not sure how i came up with it originally, after more testing it seems reliable that it is 2.6.33 though [06:55] Daviey, there aren't any 2.6.33 rc builds on the ppa so that is as narrow as i can make it [06:55] * ccheney bbl [08:13] what's the best solution for automatic mounting/unmounting of usb drives for ubuntu server? [08:14] daishadar: mount/umount ? [08:15] jpds: um, how do you make those automatic? like when i plug in a usb device i want it to be mounted [08:28] daishadar: this is what you are looking for: http://packages.ubuntu.com/intrepid/usbmount [08:28] oops - s/intrepid/lucid/ [08:29] funnily enough i'm still running an intrepid server... thanks, that looks like what i need [08:30] daishadar: just install via apt-get. what it does is add some rules to udev, to trigger the mount when the usb devices are plugged in [08:45] alonswartz: nice! === oubiwann is now known as away [09:16] New bug: #596780 in tomcat6 (main) "tomcat6 still missing JSTL ("glassfish-javaee")" [Undecided,New] https://launchpad.net/bugs/596780 [09:21] New bug: #596781 in tomcat6 (main) "JSP examples refer to Tomcat 5 (not 6)" [Undecided,New] https://launchpad.net/bugs/596781 [09:52] <_ruben> joy .. seems i made an error in my /etc/network/interfaces file .. now it wont come up after reboot, not even in recovery mode [09:54] <_ruben> kinda wierd it complains about the start/stop of lo and eth0 though, didnt alter those [10:03] <_ruben> ah .. bug #512253 [10:03] Launchpad bug 512253 in ifupdown "Mistake in /etc/network/interfaces keeps the system from booting" [Undecided,Confirmed] https://launchpad.net/bugs/512253 [10:29] apache+mod-rewrites fun. Where a simple rewrite "test" should be RewriteRule ^/$ redirected.php. Only that gives "You don't have permission to access / on this server." Removing the caret seems to fix this, but everything I can find online says nothing about excluding carets. Nice way to start :X [11:22] Eucalyptus is an open source cloud platform which is compatible in frontend with Amazon EC2 API, but implements a totally different design structure in backend. [11:23] What is the diffrence on backend and frontend api? [11:23] http://cloudy.liangz.info/2009/11/provisioning-on-existing-cloud.html <-- written here. [11:26] <_ruben> frontend: the looks of a car .. backend: the engine (for example) inside [11:26] <_ruben> you can change the engine and still be able to drive the car [11:26] <_ruben> when done properly that is [11:26] Can you give me some examples of the backend api? [11:28] Is there any backend api at all? [11:29] Or is the only the backend design structure diffrent? [11:29] -the [11:40] <_ruben> sanderj_: in this case the backend entails the software+hardware used .. euc is an opensource suite, whereas ec2 is most likely (mostly) closed source (never used it, so dont know the details) [11:42] _ruben, so the api for tracking how much resources each customer is using in euc is done trough a frontend ec2 compatible api? [12:58] any experience using confluence + webdav? [13:28] UEC QUESTION: What should the VNET_BRIDGE value be on a NC in MANAGED mode? [13:28] Does it still need to be br0 in MANAGED mode? [13:31] New bug: #596859 in ntp (main) "ntp profile denies write access to serial devices" [Undecided,New] https://launchpad.net/bugs/596859 [13:47] morning :-) [13:48] Sommer: Morning! [14:15] kirkland: ping [14:15] hggdh: morning! [14:16] kirkland: good morning, sir [14:17] kirkland: we have a new (tentative) patch for the public IP issue on UEC from Greg [14:17] kirkland: I am not sure how to work on the bzr to set it up [14:20] hggdh: awesome [14:20] hggdh: would you like me to show you how? [14:20] hggdh: i'm in a meeting at the moment [14:21] kirkland: I would love to know how to do it, it would free you... [14:21] kirkland: and I can wait if needed, right now in Lexington [14:22] hggdh: good point, yeah [14:22] hggdh: okay, i'll put it into a PPA [14:23] hggdh: and send an email explaining how i did that [14:23] hggdh: and just poke me if you need more [14:24] kirkland: ack, and thanks [14:25] lamont: what do you think of bug #596363 [14:25] Launchpad bug 596363 in bind9 "db.root needs update for i.root-servers.net AAAA record" [Undecided,New] https://launchpad.net/bugs/596363 [14:27] zul: I think meh. [14:27] but yeah, it prolly does [14:27] not worthy of an SRU, might be worthy of throwing into a point release for lucid, definitely worth doing for maverick [14:32] hggdh, which public ip issue is that? (/me wonders if it what he saw before) [14:35] kirkland, ^ ? [14:42] * ccheney brb [14:42] guys, problem: I just installed PAE kernel and my X server is broken. Please advise === kobrien_ is now known as kobrien [14:43] I assume the pae kernel doesn't support my graphics card [14:46] ok, embarrassing question [14:46] 'not worthy of an SRU, might be worthy of throwing into a point release for lucid, definitely worth doing for maverick' === away is now known as oubiwann [14:47] i didn't realize there was any distinction between point release and SRU. i thought that point releases just picked up state of the archive, and to get into that you had to follow SRU [14:47] I'm doing a pxe installation using kickstart. I'm placing a custom sudoers file for the installing machines. WHY THE HELL DOES UBUNTU ALTER MY SUDOERS FILE ?! lol it adds "$admin ALL=(ALL) ALL" which I don't want .. [14:47] lamont, zul is that incorrect ? === shade_ is now known as \shade\ [14:48] smoser: yes thats correct [14:49] tgalal, well, it does that because thats how you become root in ubuntu. i wonder when are you placing the custom sudoers file (ie, in post?) [14:49] smoser, yes in post [14:49] smoser, any advice? [14:49] smoser: inclusion in a point release might well require an SRU [14:49] smoser: I've pretty much avoided figuring such things out [14:50] tgalal, no, i've no advice. i guess that some things just happen after 'post' . [14:52] smoser, I'm trying to have sudo on my machine passwordless out of the box. That's why I place my custom sudoers file in post. The problem with that line is that makes all users of group admin can gain root priviledges but WITH a password. Is their a workaround that you're aware of ? [14:52] well, definitely you can modify that to your liking [14:52] kirkland: http://google-opensource.blogspot.com/2010/06/introducing-google-command-line-tool.html [14:52] but i don't know about through kickstart [14:53] hmm ok thanks [14:57] * ccheney back === mathiaz_ is now known as mathiaz === oubiwann is now known as away [15:18] so kickstart has " rootpw --disabled" for passwordless sudo.. but it's not working, advices? === oubiwann is now known as oubiwann-away === oubiwann-away is now known as oubiwann [16:11] With Ubuntu Server, how would I put files onto the server? [16:11] Would I mostly control everything with a remote computer? [16:17] Hi everyone ! === dendro-afk is now known as dendrobates [16:23] Mrcheesenips: it is 100% up to you! [16:24] Mrcheesenips: I myself use a remote computer to connect to the server and do things using FTP/SSH [16:24] oh' [16:24] I see :D [16:24] Mrcheesenips: What are you thinking of using the server for? [16:24] A Garry's Mod Server [16:25] Are you familiar with that game? [16:25] Mrcheesenips: You will want to install an SSH client and do everything through that. [16:25] Mrcheesenips: I believe you are following a guide of some sort to do that? [16:25] Yeah [16:26] Mrcheesenips: Would it be possible to see the guide you are going to follow? [16:27] Its actually included in the readme with the tool [16:27] I don't think its on the web [16:27] Ok, what tool are you using? [16:27] New bug: #596923 in lm-sensors (universe) "Asus F8Sg high CPU temperature" [Undecided,New] https://launchpad.net/bugs/596923 [16:27] Its the Half Life Update Tool [16:28] To download and install the Garry's Mod content from Valve's servers [16:28] to run the server [16:28] Then I have to run the server program itself [16:28] And there is a Linux Client supported [16:29] Mrcheesenips: Ok I understand, that's great! [16:29] yus :D [16:29] I understand the basics of Ubuntu Server now [16:29] thank ya [16:29] Mrcheesenips: I was just thinking instead of trying to move the update tool from machine to machine it would be quicker to download off the internet and execute. [16:29] ya [16:29] I was thinking the same thing [16:30] ScottK: ping fyi im going to update dovecot to the new release if debian hasnt done it yet [16:31] zul: OK. I think I have the changes for renaming postfix-dovecot here. [16:31] ScottK: can you push them to a bzr branch or give me a debdiff? [16:34] zul: Yes. I just found them .... [16:35] zul: Untested, but I think complete: http://pastebin.com/tdvTjR6P [16:37] ScottK: thanks [16:37] zul: No problem. Thanks for asking and taking it into your merge. [16:41] Anyone that is used to work with webmin please ? i would like to add the first domain name to my server, bind9 is installed, webmin too, but i don't know where to start [16:41] !webmin | Mateo_ [16:41] Mateo_: 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. [16:42] !ebox [16:42] ebox is a web-based GUI interface for administering a server. It is designed to work with Ubuntu/Debian style configuration management. See https://help.ubuntu.com/community/eBox [16:43] Mateo_: I'm not happy with ebox - but be warned about webmin [16:43] guntbert: why not happy ? [16:45] Mateo_: it doesn't use the system config files (as webmin does) but runs it's own config -- which I don't like - but I don't use webmin either except for very limited tasks [16:45] ok guntbert [16:46] guntbert: so i'm going to install it and then i will try to setup my domain name [16:48] Mateo_: your call :-) [16:49] Mateo_, webmin is gonna break your junk [16:49] !webmin [16:49] 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. [16:52] isntall is in progress [16:53] what tool should i use for managing partitions from the command line? [16:54] irvy: fdisk ? [16:55] thx [16:55] is there anything that would give a slightly more guided user experience [16:55] irvy: there is cfdisk too - a little nicer [16:55] like the editor in the installer [16:55] sweet, thanks [16:57] now, i wonder what this random 2155mb fat32 partition that came pre-loaded on this server is.. [16:57] irvy, a restore partition? [16:57] guntbert: did you installed the 1.5 release of ebox ? [16:57] i feel as though i should wipe it [16:58] there was no os loaded [16:58] there's an 83mb utility partition [16:58] followed by 2155mb bootable fat32 [16:58] hm [16:58] it's dell if that makes a difference heh [16:58] Mateo_: I really don't remember, I installed ebox once and removed it almost immediately again - just not my type I guess :-) [16:59] jjohansen, hi! [16:59] jjohansen, i need some more kernels from you :) [16:59] ccheney: oh, what do you need [16:59] irvy: dell servers come with a "service partition" (or so) it is useful for support issues [17:00] ahh perhaps i'll refrain from wiping it just yet then [17:00] what would you guys recommend for a filesystem on a raid5 array of 3x 1tb drives [17:00] jjohansen, bug 588861, i need kernels between 2.6.32 and 2.6.33 if possible to help narrow down which one broke eucalyptus image registration [17:00] Launchpad bug 588861 in eucalyptus ""pad block corrupted" error when trying to register an image with 2.6.34 kernel" [High,Confirmed] https://launchpad.net/bugs/588861 [17:00] i want a single partition for data [17:00] jjohansen, 2.6.32 and 2.6.32-15.5 seem to work fine, but 2.6.33 does not, and there didn't seem to be any 2.6.33-rc's on the ppa [17:01] Daviey: just curious, do you have any concept of what woudl be a *good* time for a 'virsh save'? I'm getting 50-60 seconds to produce what becomes a 100M image [17:02] ccheney: give me a sec [17:03] jjohansen, ok [17:05] ccheney: okay, I if you give me a bit I will build you some kernels [17:05] jjohansen, thanks! :) [17:09] ccheney, Daviey: have either of you tested a .35 kernel for this bug? [17:09] i didn't test 35 stock but i did try 2.6.35-4-server [17:09] i can test out 35 stock if you would like should only take about 15m to run my script [17:10] ccheney: or a maverick .35 [17:10] blah how to mount an external driver remotly ? [17:11] normally i am logged in on the machine and it automounts but i had to reboot. now my external drives are not accessable [17:11] jjohansen, the v2.6.35-rc3-maverick? the one in the archive '2.6.35-4-server' had the issue [17:12] ccheney: ah right, I was just looking at the set listed at the end [17:12] ccheney: don't bother then [17:12] jjohansen, ok [17:18] Is preseed like kickstart? how would I do automated/handsfree ubuntu server installs over the network? [17:20] hey guys i have a vps, and i want to know if i setup a primary dns server could i then use open dns as a 2ndary and tertiary dns servers if the primary fails [17:24] Ng: hey, do you have an automated testcase to verify bug 588293 ? [17:24] Launchpad bug 588293 in qemu-kvm "Memory leak" [Medium,Fix committed] https://launchpad.net/bugs/588293 [17:24] suppose maybe the ltp aio testcases in a guest might push it faster [17:26] just leaving fedora in a guest does not appear to suffice... [17:27] hallyn: I just fire up 4 VMs (on a quadcore 8GB box) and run spew (exactly, I leave "spew -g --raw --statistics -t -v 5G foo" running in screen) [17:28] hallyn: the 4VMs appear to be sufficient to saturate the IO of the host, and 4*2GB VM will leak sufficiently large to be OOM killed within 8 hours [17:28] Ng: i see, i missed that part - so you have to push the host in order to do this [17:29] how do I install ubuntu jeos ? just hit f4 to pick minimal ubuntu virtual machine and then select install ubuntu server ? [17:29] hallyn: I don't think the level of IO I'm doing is triggering the leak, given the comments in the bug about people having to deal with leaks every few days/weeks, I just happened to have found a way to provoke it very quickly ;) [17:29] Ng: thanks! I'll give that a shot [17:29] (will leave my laptop unusable of course :) [17:29] or just get the vmware ubuntu 10.04 jeos appliance ? [17:30] hallyn: I suspect it would still be obvious with one VM doing as much IO as it can. The difference is quite dramatic - I either come back after 8 hours to find 4 almost exactly 2GB VIRT kvm processes, or horror and destruction ;) [17:31] blah how to mount an external driver remotly ? [17:31] normally i am logged in on the machine and it automounts but i had to reboot. now my external drives are not accessable [17:45] hey guys about name servers. if i have a primary name server (bind9) can i use opendns as 2ndary and tertiary fall over name servers just incase [18:03] I just finished installing server edition; I selected the 3 options dns, openssh, and virtual machine host. it rebooted and booted but turned the screen off [18:03] evga 680i, radeon x1600 [18:03] RAID 5 [18:04] can someone help me just install a standard vga driver, BLIND [18:04] is there any documentation regarding lucid lynx server and request-tracker3.8 ? [18:06] what is the /etc/init.d/ filename of the openssh server? [18:06] If I get it started, I see where it was leased the ip .233 [18:08] after I rsync a file onto one of my nfs shares, trying to access it with apache gives me "Stale NFS file handle: access to /download/Sparkle.xml failed" [18:08] so I cd into the directory, and do an "ls" [18:08] and that fixes it [18:08] seanjohn: /etc/init.d/ssh === oubiwann is now known as oubiwann-away [18:18] i'm blind here :) I just installed ubuntu server, detected everything on my nvidia 680i and core 2 quad, but, after install was complete and rebooted, it booted up BUT the screen turned off. Can someone give me instructions on what exactly to type to get openssh server going? I selected openssh, dns, and virtualization host as the packages. [18:18] If i try to connected to the ip leased to it, it says "connection refused" so, it IS booted up [18:19] the screen won't turn on whatsoever [18:19] tried every key on the keyboard === oubiwann-away is now known as oubiwann [18:34] anyone good with samba shares for windows? [18:34] Hi, I installed UEC and as far as I can tell neither the CC nor the NC is starting. When I try to start them and it only seems to start the apache components. It seems that there should be java components but I can't figure out what is supposed to start them [18:35] I'm interested in giving Ubuntu Enterprise Cloud a spin.. and was wondering if it's possible to run Windows Server as a virtual image? [18:37] Egonis: it is, but I think it would be really complicated to have them automagically join an AD [18:38] mathiaz: ping [18:39] goldins: Essentially, I need the Windows Server only to run Blackberry Enterprise Server, and Pervasive.SQL for a weird windows application. So is 'AD' (which I presume you mean Active Directory) necessary in that configuration? [18:43] it sucks, especially that the entire distro is probably on the cd of 700 mb, that ubuntu doesn't ask what services should start at runtime, static ip config, and etc. so that YOU DONT NEED A SCREEN AFTER THE INSTALL. [18:44] this is a server edition; I feel sorry for rack-mount usage [18:45] seanjohn: Install the openssh-server package to get an SSH server... [18:45] that was part of the selection [18:45] of packages [18:45] seanjohn: Setup one host, and image it for multiple usages. Real simple./ [18:45] Then it should be there and be started automatically at boot. [18:45] "connection refused" on attempt to connect [18:46] it doesn't use vga for command-line only? it seems to try to use opengl for a COMMAND LINE [18:46] only explaination of why it would't support all gpu's [18:46] Umm. [18:46] You mean the console? [18:46] yeah [18:46] It uses fbcons. [18:46] Not opengl. [18:46] hey psi-jack [18:47] well, the install went great BUT on reboot, after booting, it turns off the monitor [18:47] Anyway. I had an actual question and not a not-pit to do. ;) [18:47] I want to disable network-manager from starting at boot, but update-rc.d doesn't seem to be allowing me to disable it. [18:47] seanjohn: If you really need highly customized installations, you can do preseeding, see the Server Guide https://help.ubuntu.com/10.04/installation-guide/i386/appendix-preseed.html [18:48] "customized installations" it setup raid 5 easier than it did just a standard graphics adapter [18:48] the raid 5 works cause its booting [18:48] I can blindly login and type shutdown -r now and it will restart [18:49] i tried doing sudo /etc/init.d/ssh start [18:49] i tried doing /etc/init.d/iptables stop [18:50] grrr, i feel like going for gentoo [18:50] you should be able to boot without the framebuffer console AFAIK [18:51] JanC: is that for me? [18:51] the post of the MB goes so fast, it doesn't present a flash screen for ubuntu before booting [18:52] keep left shift pressed down during boot to get into grub... [18:53] and if openssh is installed, it should run and work [18:53] don't forget you can't login as root though [18:54] Egonis: sorry, I think that explicitly is not something you want to use the cloud for: the BES and these other applications would be much better to run in a straight KVM without all the cloudy stuff [18:55] goldins: Is there a moron's guide to Cloud Computing? I obviously don't have my head wrapped around it [18:55] Anyway. ;) [18:55] As for upstart jobs, is there like any tool that truely works to manage those? [18:56] Egonis: see the ubuntu server guide for a start [18:56] RoyK: Yes.. :P I suppose I need it dumbed down further [18:56] https://help.ubuntu.com/10.04/serverguide/C/uec.html [18:56] Egonis: I'm not sure, the idea is for clustering I'm pretty sure; if you for example want to be able to launch 20 nodes that are running the same job with different parameters, that's the kind of thing you want to have a cloud around for [18:57] so for instance, a web application is what a cloud environment would be best for? [18:57] no... [18:57] if so, does the language matter? I write Mono.ASP apps [18:57] a math application [18:57] wow... I'm so far off then. :P [18:57] Egonis: cloud computing is just virtualisation gone large [18:57] so matlab types of scenarios? [18:57] Egonis: anything, really [18:58] Egonis: personally, I wouldn't use virtualisation for number crunching [18:58] How about documentation that details out how to work with scripts in /etc/init.d? I'm basically just trying to disable network-manager from starting at boot-time, along with a couple others that update-rc.d isn't working on. [18:58] Egonis: we have a few bunch-of-cores machines doing that sort of thing [18:58] RoyK: Okay... I'm looking to aggregate my existing set of servers... I was hoping this would be an easy solution for expansion of VM's. My mistake. Thank you for clarifying [18:59] Psi-Jack: /etc/init.d/skeleton is a good start [18:59] RoyK: Not making them. Disabling or enabling them at boot. [18:59] oh [18:59] As I said, update-rc.d isn't working on network-manager, for starters. [18:59] with 10.04 a few things changed there - I don't remember the details on how to disable the ones not in /etc/rc2.d [19:00] Egonis: I think it could simplify things a bit - what sort of virtualisation are you using? [19:00] Yeah. And little to no documentation to be found, either, which is 1 of my 2 biggest gripes about ubuntu present-date. ;) [19:00] upstart == incomplete, and grub2 ;) [19:01] RoyK: At the moment, VirtualBox for client application testing.. but we have separated physical servers at the moment. A mix and match of Windows Server 2003, SBS 2003, Server 2008, and Linux [19:01] Psi-Jack: What problem are you having with grub2? [19:01] Psi-Jack: heh - we still stay at 8.04 for now on most of our servers :) [19:02] RoyK: Looking to do some house cleaning [19:02] Egonis: kvm is quite good [19:02] Egonis: vbox isn't really a server tool [19:02] RoyK: I will check it out, thank you. Yes, vBox has been used only for at-hand testing.. not for server instances [19:03] Jordan_U: It's alpha. :p [19:03] I used vbox for server instances on opensolaris for some time - it hung the host after some time - switched to ubuntu/kvm and problem solved [19:04] there are some bugs with ubuntu+kvm too, especially with high network load between host and guest, but it hasn't hit me too hard [19:04] Jordan_U: And the way Ubuntu sets it up is painfully minimal. Like, for example. how do you get to the command prompt of grub when you absolutely need to? ;) [19:04] Psi-Jack: Hold shift to get the menu, press 'c' to get to the grub shell. [19:04] RoyK: I had another bug on that myself, where it wasn't adding vnet# to the bridge device. [19:04] Jordan_U: press escape [19:05] RoyK: So, my firewall was rejecting everything from my kvm guests. ;) [19:05] Psi-Jack: ufw? [19:05] Not to mention it wasn't in the bridge, so no network capability. [19:05] RoyK: Heck no. shorewall. [19:06] Psi-Jack: there are ways to configure ufw to allow the guests - don't know about shorewall [19:06] New bug: #596739 in mysql-dfsg-5.0 (universe) "package mysql-server-5.0 5.1.30really5.0.75-0ubuntu10.3 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/596739 [19:06] RoyK: I just manually brctl addif vnet0 br0, and it started working perfectly thereafter. [19:06] heh [19:07] * ccheney bbl lunch [19:07] So, still nobody knows how to stop network-manager from starting at boot-time when it doesn't use /etc/rc2.d/ etc? [19:08] I'm guessing it has something to do with /etc/init/*.conf, for example /etc/init/network-manager.conf [19:14] Ahh.. I'm seeing now. === alonswartz_ is now known as alonswartz [19:15] Basically I have to edit the "start on" part of /etc/init/network-manager, but .. To what? [19:34] * ccheney back [19:39] smoser: around? [19:39] here [19:41] smoser: is ec2-describe-images with an AMI ID going to be super slow? [19:43] define super [19:44] http://paste.ubuntu.com/453040/ [19:44] SpamapS, its mostly a cpu bound operation [19:47] smoser: its a once-per-run operation, and only if people choose "auto" instance type, so I won't worry about running it. [19:49] SpamapS, yeah, it does suck. when using euca-describe-images its much faster (~ .5 seconds) [19:49] kirkland: can you talk? I want to discuss one thing with you [19:49] the time still stinks though. [19:49] i really only see 2 good ways to avoid it [19:50] a.) user specifies the type [19:50] b.) we cache the describe-images information for a given ami-id. [19:51] the issue with the caching is that right now, we are completely "region un-aware" [19:52] while possibly very unlikely it is possible that if we cached "ami-12345678 is amd64" and didn't attach the qualifier "on us-east-1 EC2 region" then we could have a collision and have wrong data [19:52] SpamapS, ^ [19:52] i guess there is also [19:52] Hey folks, I've got some UEC questions... [19:53] I recently set up a simple cloud with a CC and one NC [19:53] but the defaults seem to prevent me from keeping anything across reboots [19:53] smoser: yeah unless we know about the env vars we can't do much caching really. [19:53] c.) only do 'auto' if specified as 'auto'. otherwise, pass nothing (which will result in i386 on ec2). this makes the default 'run' case faster... sort of. but it makes it more difficult to run. [19:54] if I shut down an instance, it is impossible to restart it, and if I reboot the NC, everything is wiped clean [19:54] SpamapS, i would consider peeking at EC2_URL on this. [19:54] shoot, but euca2ools doesn't necissarily have it set. [19:54] yeah, we'd have to deal with knowing the ocnfiguration [19:54] anyone know of any solutions? [19:54] and keeping it in line with those tools. [19:55] smoser: you have to look at the args and the env var.. this is where I start thinking API calls instead of exec... ;) [19:55] SpamapS, args and environment aren't enough [19:55] you'd also have to search ~/.eucarc [19:55] and parse that [19:55] so its one more thing [19:56] SpamapS, the reason i didn't want to do API calls is that the closest thing we have to "standard configuration" is having one of those sets of tools configured to work [19:57] i wou'dnt be opposed to creating "yet another ocnfiguratoin format" [19:57] but it seems like overkill for this [19:58] anyone know of a slick bandwidth monitor that will tie into webmin nicely? [19:59] !webmin | Krazyderek [19:59] Krazyderek: 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. [19:59] It just seems really crazy to me to have a cloud with no ability for persistence... [19:59] oh dear.... that explains a couple of things [20:00] !webmin [20:00] 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. [20:01] Krazyderek, it already has a bandwidth module [20:01] If there is a more appropriate channel for me to go to for UEC issues, please let me know [20:01] @bogeyd6 which one? [20:02] coder7: You can try in #ubuntu-virt perhaps? [20:02] @bogeyd6 or rather, is there any pre installed? or should i just pick from one of the lists i find on a ubuntu page? [20:03] Krazyderek, already preinstalled tbh [20:03] @bogeyd6 specifically i want to see real time usage of local users on a intranet on a fileserver [20:03] Krazyderek, http://www.webmin.com/standard.html [20:04] coder7: I guess you have seen https://help.ubuntu.com/10.04/serverguide/C/uec.html, see also #ubuntu-virt, #eucalyptus [20:04] !ntop | Krazyderek [20:04] hmm [20:04] Krazyderek, ntop [20:05] @bogeyd6 that's a good one? [20:05] the best one [20:05] Krazyderek, https://help.ubuntu.com/community/Ntop [20:05] @bogeyd6 ya the "bandwidth monitoring" in webmin gives me an error "Neither of the System Logs modules are installed on this system and supported by your OS. The Bandwidth Monitoring module cannot operate without one of them." [20:06] Krazyderek, thats because webmin doesnt fully work on an ubuntu sytem [20:06] i guess "all except windows" isn't entirely correct on the webmin page then lol [20:07] jmarsden & guntbert: thanks... and yes I've looked through the docs [20:08] coder7: I assumed :-) I hope the other channels are more helpful today - Good luck :-) [20:08] @bogeyd6 hmmm ntop isn't playing well so far [20:09] Krazyderek, #ntop [20:10] ntop is in the 10.04 repos [20:10] @bogeyd6 **ERROR** ....open of /var/lib/ntop/prefsCache.db fail ed: File open error [20:10] Mon Jun 21 16:08:06 2010 Possible solution: please use '-P ' [20:10] Mon Jun 21 16:08:06 2010 **FATAL_ERROR** GDBM open failed, ntop shutting down.. . [20:10] Mon Jun 21 16:08:06 2010 CLEANUP[t140709457868544]: ntop caught signal 2 [20:10] Mon Jun 21 16:08:06 2010 THREADMGMT[t140709457868544]: ntop RUNSTATE: SHUTDOWN( [20:11] !paste | Krazyderek [20:11] Krazyderek: For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://tinyurl.com/imagebin | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [20:12] sorry [20:13] Krazyderek, did you use the init script or just run ntop from the command line? [20:13] cmd [20:14] well derp [20:14] Krazyderek, if you didnt read the link i gave you for ubuntu community ntop setup you deserve to fail :( [20:14] bogeyd6: please stay helpful [20:15] @bogeyd6 [20:15] @bogeyd6 lol [20:15] fair enough [20:15] kk [20:15] go read that stuff bro [20:15] its step by step [20:16] Krazyderek, its supposed to run as a service and you visit a website for more specific setup [20:16] @bogeyd6 awww my bad [20:17] @bogeyd6 should i / how do i uninstall ntop and start over? [20:17] no [20:18] just start the step after apt-get install ntop [20:18] ok [20:18] Krazyderek, https://help.ubuntu.com/community/Ntop#Configuration === unreal_ is now known as unreal [20:27] @bogeyd6 oh dear... lol [20:27] it doesn't want to monitor my bonded ethernet [20:32] Krazyderek, dunno [20:35] Krazyderek, just googling around shows me some kernel type things they are talking about but I generally have no idea on whats up [20:35] @bogeyd6 i also forgot to restart the service [20:36] Krazyderek, good, because no other articles mention a problem with bonding [20:36] I have a belkin Wifi Adaptor, that I would like to use to setup a home web, print, music, ftp, and ssh server. How can I do this? [20:36] and i thought to myself that to dump an entire 48 port switch to one machine you would need dual gigabits [20:39] Is this even possible? [20:42] nevermind, this came up right on the install. [20:49] Actually it didn't work can I receive some assistance [20:50] blsmith, i just stumbled in. can you give a synopsis of your problem? no gurantees that i can help though. :-) [20:50] *guarantees [20:52] CpplsWeird: I am attempting to set up a home server using a Belkin Wifi USB Adapter [20:53] hmm [20:55] I was under Ubuntu 10.04 Live CD and it worked just right.During the installation it was attempting to automatcally set up the network and found wlan0 and I attempted to set that up without sucess. [20:55] @bogeyd6 even if i stop ntop i can't get it to start on bond0 using sudo ntop -i eth0 [20:55] *installation of the server [20:56] init script Krazyderek [20:56] /etc/init.d/ntop restart [20:56] blsmith, a home server to serve what specifically? [20:56] print, web, ftp, music, and ssh [20:56] lol, this is going to take some time. :P [20:57] @bogeyd6 that restarts it on eth0 [20:57] Krazyderek, then check the init script [20:57] CpplsWeird: possible samba for backup of local computer. If I can just setup the network to work, then all should fall into place [20:57] ok so right now you're stuck on getting your usb wireless adapter working on ubuntu live cd? [20:58] CPPlsWeird: Sorry the adapter works on the live cd but I can't set it up on the server side., I continually get the Network autoconfiguation failed msg. [20:59] CpplsWeird: Should I not set up the network right now? [21:00] blsmith, try this http://ubuntuforums.org/showthread.php?t=12045 [21:01] however be sure to use your adapters interface, wlan0. [21:02] CpplsWeird: The network's password is under WPA/WPA2 TKIP + AES, not WEP [21:02] **ERROR** RRD: Disabled - unable to create directory (err 13, /var/lib/ntop/rrd/graphics) [21:03] @bogeyd6 i keep getting that error when i try to monitor bond0 [21:08] blsmith, looks like you have some fun ahead of you. http://ubuntuforums.org/showthread.php?t=202834 [21:08] CpplsWeird: Espically when i get debootstrapping errors when installing [21:10] "debootstrapping error" [21:10] ive never heard of such a thing [21:10] !google debootstrapping error [21:10] Error: I am only a bot, please don't think I'm intelligent :) [21:10] CpplsWeird: I think thats what it said. [21:11] hmm i guess i'll just take it as it is for now [21:11] CpplsWeird: It alright I have another cd, that i'm going to tryp [21:12] yes, because if you stick enough cd's in the cdrom drive its bount to magically start working sooner or later. [21:12] *bound [21:13] CppIsWeird, thats the theory i'm going with. [21:13] and i thought my strategy of cussing at people until things work was inefficient. [21:14] CppIsWerid: Its working now, and I know why, I starting making a copy of Ubuntu 10.04 Desktop, but without realizing it I started the server version, and never through it away. So This cd I have now is working fine. [21:15] CppIsWeird: I'm installing without any network settings and will change them using the guide you gave me. === tschundeee_ is now known as tschundeee [21:21] exit [21:23] i have an ubuntu server providing services on a completely private network, disconnected from the internet. what's the simplest solution for providing email to hosts on the network for integration with other services like wiki, request tracker, nagios, etc? [21:26] panfist: did you see https://help.ubuntu.com/10.04/serverguide/C/email-services.html ?, then there is sudo tasksel to install all mail software,... [21:27] i was hoping there was some sort of condensed version that let me skip over all the internet related material [21:36] for example, if all my mail isn't going from my domain to any outside domain, it's all on the LAN only, that would simplify a lot of work done by the MTA [21:46] New bug: #597021 in apache2 (main) "apache2 can not load fcgid_module" [Undecided,New] https://launchpad.net/bugs/597021 [21:48] Hi. I'm looking at installing a Zimbra server. Among underlying OS options are RH/Centos55 & Ubuntu 8 LTS. Iiuc, both are on the road to being replaced soon-ish. OS upgrades *under* an app make me ... nervous. I come from an RPM-based distro bkg, but have rad that Debian/Ubuntu have the "install it once, never worry about upgrades" package mgmt. [21:48] Tbh, I'm not sure what that means :-/ [21:48] Can anyone comment on how straighforward an Ubuntu 8 LTS -> 10 LTS should be? If "painfuL', I lean to sticking with Centos ... [21:53] DannyT: Ubuntu 8.04 LTS server is supported for 3 more years, but upgrading from 8.04 LTS to 10.04.1 LTS should work out fine [22:07] DannyT: what that means is that you can upgrade the distribution painlessly. If you've ever used Fedora's preupgrade feature, this is how that feature would work if it ever worked correctly [22:07] JanC: Sure, I know it "works out" ... lots of folks do it. I've been burned by promises of "long term support" (not @ Ubuntu ...) where the support--rats abandon ship the moment I try to do _anything_ other than what's shipped with the original release. I guess a better way to say what I'm looking for -- I'd *like* to install today-ZImbra on toady's supported Ubuntu 8 LTS, then when Ubuntu 10 LTS is supported by Zimbra (soon-is [22:07] Yeah, I know its a "fuzzy" question -- best I can manage, atm [22:08] basically upgrading the distribution never requires running an installer. It just upgrades all the packages to their new versions and if you have a new kernel you need to reboot [22:08] goldins: Ok, that's sort of what I'm hoping to hear. Add to that that the community "in here" hasn't called me an idiot for wanting to "do anything" to the LTS distro in the first five minutes .... we're ahead of the game! ;-) [22:09] DannyT: everything after "(soon-i" was cut off (IRC doesn't support lines that long) [22:09] unlike centos, which doesn't actually have preupgrade yet so upgrading the distribution (from 4 to 5 or 5 to 6, say) which requires you to burn two DVDs and run an installer, possibly necessitating several hours of downtime. [22:09] JanC: Sorry, didn't know that: (soon-ish ...) *not * end up with a hellish nightmare trying to updated to 10LTS ... [22:11] goldins: Didn't realize that re: Centos, and therfore I assume RHEL? ... good to know [22:12] Does Ubuntu Server 10 LTS run "happily" in/as a Xen DomU? THat'd be my fastest way to get started .... [22:28] hallyn: [serge-hallyn] Forward port libvirt (to 0.8.1?): TODO [22:28] hallyn: that looks to be DONE [22:31] hggdh: can you ensure that you update https://blueprints.edge.launchpad.net/ubuntu/+spec/server-maverick-uec-testing with your progress this week? [22:31] hggdh: i think some of those items are to be solve by you guys up there this week [22:33] kirkland: will do [22:33] kirkland: uh, yeah. though i am trying to get a version of that which actually does 'virsh save' in reasonable amount of time [22:33] (but when i add the patch to fix that, i break 'virsh restore', still dont' know why) [22:33] hallyn: okay, i updated that spec a bit [22:33] * hallyn looks [22:34] kirkland: cool, thanks. [22:34] i'd forgotten that was on that bp [22:35] kirkland: should 'make virsh save work' be a separate item on there, or just keep it off since there's a bug (owned by Daviey) ? [22:35] hallyn: hmm, why does daviey own that one? sounds like it should be yours [22:36] kirkland: cause he's had it since february :) [22:37] hm, maybe he doesn't [22:43] ok now i'm getting cranky. even when i manually rebuild jdstrand's package with no other changes, it still won't restore vm's any more [22:46] hallyn: libvirt uses debian source version 3 (quilt) [22:47] hallyn: it is possible that even though you removed the patch from series or debian/patches that it is still applied in a 'debian' patch in debian/patches [22:47] hallyn: check down in debian/patches for a patch that starts with 'debian' [22:49] hallyn: the some thing happened to me during the merge. I'm sure this new source package quilt version works for someone, but I've yet had it not get in my way [22:49] jdstrand: yeah, i just found that. had to fight to keep from loudly exclaiming wtf [22:49] s/some/same/ [22:49] hallyn: yeah, it is a bit weird to get used to [22:49] so this means what, i left the patches applied in working tree and it automatically did a quilt refresh? [22:49] hallyn: exactly [22:50] so now i have no idea whether my previous tests were also fubared [22:50] hallyn: so you need to 'export QUILT_PATCHES=debian/patches" then 'quilt pop -a' before building [22:50] then use quilt normally [22:50] all right thanks - i'll try again later... [22:50] sure, np [22:51] * jdstrand kicks debian source format 3.0 (quilt) on hallyn's behalf [22:51] thanks, i feel better :) [22:51] :) [23:10] jjohansen, any updates about kernels? [23:35] ccheney: sorry not yet [23:36] jjohansen, ok np [23:38] ccheney: I had a question about the 2.6.34-lucid kernel you tried [23:38] jjohansen, ok [23:38] was that one of tim's backport kernels? [23:39] its the one from http://kernel.ubuntu.com/~kernel-ppa/mainline/ [23:39] ah, so just from before renaming to maverick [23:39] okay thanks [23:39] happy to help :) [23:40] send me an email whenever you have the new ones ready and i can run my script on them and see if i can't find the spot it broke :) [23:40] ccheney: will do [23:48] Curious. Do upstart scripts generally read from /etc/default/$NAME like LSB scripts do? [23:51] Psi-Jack: hmm.. great question [23:51] /etc/init/ssh says [23:51] # if you used to set SSHD_OPTS in /etc/default/ssh, you can change the [23:51] # 'exec' line here instead [23:51] exec /usr/sbin/sshd [23:51] which is, rather lame IMO [23:52] Yeah. I looked at others and found out it doesn't, you have to enclose it into a script section in the conf. [23:52] irqbalance.conf was a good one to follow. [23:52] Psi-Jack: but some of them use a "script" clause instead of an 'exec' line [23:52] I'm converting some LSB scripts to upstart to see if I can get keepalived and linux-igd to play nice together on two servers.