[00:00] kirkland: as mentionned in http://www.debian.org/doc/debian-policy/ch-files.html#s-config-files [00:00] k [00:00] mathiaz: what's the logic to tell if this is on install, and not upgrade? dpkg --compare-versions .... something? [00:01] kirkland: [ -z "$2" ] [00:01] mathiaz: or test for $2 [00:01] ah [00:01] okay [00:01] kirkland: on install $2 is empty [00:01] mathiaz: cool [00:01] kirkland: $2 holds the latest *installed* version of a package [00:02] kirkland: http://women.debian.org/wiki/English/MaintainerScripts [00:02] kirkland: ^^ has a good explanation on how scripts are called with which argument [00:08] mathiaz: http://paste.ubuntu.com/353701/ [00:08] mathiaz: something like that [00:09] mathiaz: should create eucalyptus-local.conf if this is either an install, or an upgrade from a version before we started using that file [00:09] kirkland: hm right. [00:13] kirkland: I would put the seed_local_conf call to the pkg install section of the maintainer script (below - with euca_conf -d) [00:13] kirkland: and add some more specific code to handle upgrades [00:13] kirkland: what would should be done on upgrades? [00:13] kirkland: because after the upgrade, euca_conf is going to write to eucalyptus-conf.local [00:14] mathiaz: by "install section", you mean $1 = "install" ? [00:14] kirkland: shouldn't eucalyptus.conf be copied to eucalyptus-local.conf? [00:14] mathiaz: OOOOHHHH [00:14] mathiaz: there's already a -z "$2" [00:14] mathiaz: sorry [00:14] mathiaz: yeah, let me rework this [00:14] kirkland: yes - that's what I was refering to [00:16] mathiaz: sorry, sorry, sorry [00:16] mathiaz: completely missed that [00:16] mathiaz: okay, i'm rolling now :-) [00:20] Hi All. [00:20] This is royally doing my head in. [00:21] My switch 100% supports ad network bonding [00:21] I followed this tut [00:21] http://blog.brightbox.co.uk/posts/howto-do-ethernet-bonding-on-ubuntu-properly [00:21] how can I turn this off: "You have new mail in /var/mail/david" I really don't need to be notified. [00:21] My Router & all linux boxes can ping the box that has the bond setup [00:21] but my workstation and iphone cannot connect nor ping it :( [00:22] I know that the box is communicating with the router because its picking up DHCP leases, and if I change the IP address based on the MAC and then restart the bond, it picks the new IP up fine [00:25] mathiaz: okay ... http://paste.ubuntu.com/353712/ [00:26] mathiaz: but i think i still have one problem [00:26] mathiaz: i'm looking at the debian women's chart now [00:26] mathiaz: so the good news is that we can get rid of all of those euca_conf calls [00:27] kirkland: I don't think you want to seed_eucalyptus_conf [00:27] kirkland: that should be conffile [00:27] mathiaz: right, i wanted to ask you about that .... [00:27] mathiaz: so it is a conf file [00:27] mathiaz: but it's changing drastically on upgrades now [00:27] mathiaz: we need to copy that data over to -local [00:27] mathiaz: and *then* replace it with our new conffile [00:28] mathiaz: it *is* a conffile [00:28] mathiaz: i have that code elsewhere in my diff [00:29] mathiaz: which we'll handle for new installs correctly [00:29] mathiaz: but for upgrades, we need to gently move that data over [00:29] kirkland: hm... well - dpkg will automatically prompt for the user [00:30] kirkland: because /etc/eucalyptus/eucalyptus.conf has changed both *in* the package and on the local system [00:30] mathiaz: right, this is messy [00:30] kirkland: right - that will only happen once on upgrade [00:31] kirkland: there are ways to workaround that though [00:31] mathiaz: i think this is why i keep going back to making eucalyptus.conf *not* a conffile [00:31] mathiaz: let's imagine eucalyptus.conf is more like eucalyptus-local.conf in this new model [00:31] mathiaz: it is the thing read/written by euca_conf [00:32] mathiaz: we can easily take something that was a conffile, and drop it from dpkg's installed files list [00:32] mathiaz: if it's in /etc, it should just remain [00:33] kirkland: yes - but then you have to take care of it in the maintainer scripts [00:33] " kirkland: there are ways to workaround that though" [00:33] mathiaz: ^ ? [00:33] kirkland: we could handle the upgrade by moving the old eucalyptus.conf file in the preinst on upgrades [00:33] mathiaz: oh, okay [00:34] mathiaz: that's legit? [00:34] kirkland: it's *possible* :D [00:35] kirkland: that being said we should make sure it's done after eucalyptus is stopped [00:35] kirkland: on upgrade eucalyptus is stopped, then move eucalyptus.conf to eucalyptus-local.conf [00:36] mathiaz: okay; how to verify eucalyptus is stopped? [00:36] mathiaz: "stop eucalyptus" in the preinst? [00:36] kirkland: it's done by the upstart debhelper token [00:36] pting: m1.small instanced type on ec2 gets 1 "EC2 compute unit" which is generally less than half of the real CPU on which the VM is running. What you are seeing is normal and you are getting all that you're paying for. [00:36] kirkland: I don't know if the new conffile eucalyptus.conf would actually be installed though [00:37] kirkland: since dpkg would notice that the old eucalyptus.conf conffile is no longer there and it may not reinstall the new one [00:37] kirkland: I think you should ask in #ubuntu-devel - slangasek may be able to help in that [00:38] smoser: "ping external to ec2 never works." - This depends on your security group configuration. If you want to allow ping then allow icmp in the security group. [00:38] kirkland: moving conffile around is tricky - the samba pkg did that at some point IIRC [00:38] mathiaz: okay, pinging slangasek [01:01] New bug: #460316 in clamav (main) "clamav-base package contains very big clamav data files (main.cvd and daily.cvd) from clamav-data package" [Wishlist,Confirmed] https://launchpad.net/bugs/460316 [01:11] New bug: #428179 in clamav (main) "package clamav-base 0.94.dfsg.2-1ubuntu0.5 failed to install/upgrade: el subproceso post-installation script devolvi? el c?digo de salida de error 3" [Medium,Confirmed] https://launchpad.net/bugs/428179 [01:11] New bug: #430418 in clamav "clamav-milter socket permissions wrong" [Medium,Triaged] https://launchpad.net/bugs/430418 [01:15] mathiaz: cool! [01:15] mathiaz: euca2ools/boto happier? [01:16] mathiaz: please upload to lucid as soon as you're satisfied [01:16] kirkland: well - I'd like some feedback on the package first [01:16] kirkland: some of my testing is working [01:16] kirkland: but my main stress test script is still failing [01:17] kirkland: and I haven't figured out why yet [01:17] mathiaz: hmm, okay [01:17] mathiaz: http://paste.ubuntu.com/353732/ [01:17] kirkland: and I need to grab something to eat now [01:17] mathiaz: i'm testing the euca_conf/eucalyptus.conf changes now [01:17] kirkland: I haven't had lunch yet [01:17] mathiaz: heh [01:19] kirkland: line 98: it think you want install [01:19] kirkland: hm well - nervermind [01:19] mathiaz: hmm, i'm testing for upgrade [01:19] kirkland: oh yes - you actually want install [01:20] kirkland: postinst doesn't receive upgrade as $1 [01:20] kirkland: it receives install as $1 [01:20] mathiaz: that's preinst [01:20] kirkland: ahhhh - you're right [01:20] kirkland: I need to grab something to eat [01:20] mathiaz: kirkland - 1 ... mathias - 10000000000 [01:20] :-) [01:36] New bug: #502878 in samba (main) "Samba 3.4.0 won't let win98 clients to connect" [Medium,Triaged] https://launchpad.net/bugs/502878 [03:41] New bug: #487275 in eucalyptus (main) "eucalyptus.conf should not be a conffile" [High,In progress] https://launchpad.net/bugs/487275 [03:42] !ops [03:42] Help! Channel emergency! soren, lamont, mathiaz or tom [03:51] copied my smb.conf from a different machine, same username same pass same dirs, can't get it to open? === mcasadevall is now known as NCommander [05:14] Hi? [05:18] guess not [05:22] Pickley: Hi! [05:23] Do you have a problem that we can help you with? [05:23] FireCrotch: Hey [05:23] Server used to boot straight through Grub [05:23] now it just waits for me to have to hit eneter [05:24] and also it hangs after it fsck's the partitions [05:24] cant get sound on ubuntu-server, i have pulse audio installed, vlc shows it playing the file, alsamixer works. but no sound. any ideas [05:24] Pickley: well lets solve the grub problem first - what does it display when you have to press enter? [05:25] FireCrotch: It starts booting, but it used to automatically boot through into the server [05:25] FireCrotch: Do I just need to change it back to auto select? [05:26] Pickley: Yes, if you've changed it from the default [05:26] FireCrotch: I haven't just happened today, might be because it says there are two kernels now [05:27] Pickley: Yeah, the grub configuration probably just didnt update properly [05:27] FireCrotch: Ok, I can sort that out once I get it botting properly :P [05:27] Pickley: Are you able to boot into single user mode? [05:28] FireCrotch: How do you do that? Pass something in grub? [05:29] Pickley: Yes. the word "single" as a parameter [05:29] Ok, I'll try that now [05:29] can anyone help me ;\ [05:29] Pickley: fyi it goes at the end of the kernel line [05:30] FireCrotch: Doesn't work, also get a i915_handle_error, but it worked before with that. [05:31] Can only hit Ctrl+Alt+Del to reboot really [05:34] Pickley: so it just completely freezes after it does its fsck during boot? any kind of error output from that? [05:34] FireCrotch: It just shows the i915 thing and then nothing else loads [05:35] FireCrotch: except I had that error before [05:35] Thinking it might just be my hard drive failing lol [05:35] the i915 error is referring to the intel 915 chipset [05:35] Yep [05:35] It normally worked with that anyway [05:36] Hmm [05:36] Not sure [05:36] and really don't want to lose my data [05:36] lol [05:37] although I do have the most important stuff backed up luckily [05:38] probably is a bad hard drive. how old is the drive? [05:38] Probably 3-4 years old [05:38] probably older [05:38] wouldnt surprise me then [05:39] Got another 8gb drive coming so will swap it out [05:39] *80 [05:40] Thanks anyway :D [05:42] FireCrotch: think I should just get of it when I get the new one? [05:42] probably [05:43] :D [05:43] Not like I'll use more than 80gb [05:43] The other issue I was having was the server kept dropping out [05:43] But that might be due to the drive as well [05:43] yeah, hdd problems can show up in various ways [05:44] :D [05:44] Guess I get to reinstall everything [05:45] Pickley: well you could always try to get the data off the hard drive onto the new one [05:45] Yeah [05:45] I don't think it's worth it [05:45] At least /etc/ :) [05:46] lol maybe [05:46] Server wasn't set up long ago so doesn't matter too much [05:47] Had only really set up SVN, Rails, Ruby, LAMP and Samba === rberger_ is now known as rberger === bigbrovar_ is now known as bigbrovar === rberger_ is now known as rberger [09:32] hm. seems tzdata is updated all the time. it's not like timezones are changing, is it? === MTeck-Linux is now known as MTeck-ricer [09:42] RoyK: check the changelog for what changed I guess ;) [10:07] anybody using a SheevaPlug computer? http://en.wikipedia.org/wiki/SheevaPlug [10:23] * RoyK wants [10:35] cemc: I am. [10:39] twb: is it worth buying one? [10:40] vlan trunking works? openvpn? :-) [10:40] I don't know. Mine is just a build slave. [10:40] stock 2.6 kernel ? [10:41] what distro on it ? [10:41] It's running whatever is in Squeeze at the moment [10:41] aha [10:41] The onboard mtd ships with Ubuntu 9.10 [10:41] twb: downsides? [10:41] as far as you can tell :) [10:41] It's bulky, and it's a wallwart if you're using the US connector [10:43] I have no idea how to compare that 1.2ghz ARM processor to a normal one, you're building stuff on it ? [10:43] d-i doesn't support mtd yet, so you can't actually install Debian (or Ubuntu) onto the onboard storage, and using USB for the root filesystem means that it takes a few reboots before it can actually bootstrap the kernel [10:43] d-i ? [10:43] debian-installer [10:44] so how is it ship with ubuntu 9.10 then? [10:44] With dd or similar. [10:45] err, s/is/does/ [10:45] so you can 'install' ubuntu on it, just not the usual way? [10:45] I mean you have to hack it a bit :) [10:45] FSVO a bit [10:49] hm [10:51] You should also see #openplug (iirc) [10:54] twb: do you know about openwrt and routers that support it? how does the installation part compare to that? as in easier/harder? [10:54] I'll take a look there too, thx [10:54] cemc: I never installed onto the mtd, I left it alone [10:54] ah, ok [10:55] I did a normal d-i netboot onto a USB key. [10:55] I was also extremely annoyed that the bootloader's PXE support needed next-server hard-coded, instead of using DHCP. [11:58] twb: can you show me a cat /proc/cpuinfo on it phlease? [11:58] cemc: it's offline at the moment, sorry. [11:58] np [12:17] cemc: http://pastebin.com/f7e6c1ef8 [12:18] Curious lack of jazelle for an OMAP3 [14:50] anyone knows how to get a list or remove packages which have an apt source which was disabled on upgrade? [15:10] orogor, apt-get autoclean [15:11] orogor, then apt-get autoremove [15:13] orogor: do you mean packages were installed from (say) universe, and now that you've disabled universe, you want to remove those packages? [15:14] well more stuff like medibuntu which is automatically disabled on upgrade [15:15] then after that unless medibuntu is not re enabled, there s no update for theses packages [15:17] orogor: in aptitude's GUI, these packages will appear in the "Obsolete" drop-down [15:17] Er, s/drop-down/collapsible list/ [15:17] You should be able to just hit purge (_) on that list to purge them all [15:19] can i get that in cli ? [15:20] From the line interface, it's probably something like ~O or ?obsolete() [15:26] i try cleaning up old stuff, that computer has been throught a few major upgrades already [15:27] orogor: you may want to investigate deborphan and friends [15:27] debfoster? It's been a while [15:28] why spend all the time going through the list when apt will remove them [15:29] because the list will also include pckages downlaoded and installed manually and for which no repository was provided [15:29] and maybe i dont want to remove theses [15:29] bogeyd6: because d-i, at the very least, doesn't set markauto flags correctly [15:30] Granted, I tend to markauto EVERYTHING and have a single equivs stub, because I find that easier than debfoster/deborphan [15:30] But I don't feel like teaching people how to do that [15:33] ok so if someone manually installed something, how does it get disabled on upgrade? [15:34] It's "disabled", in the sense that it matches ?obsolete(), as soon as you install it. [15:37] so aptitude search ?obsolete shows someone a list of all packages manually installed and no longer has its reqs? [15:37] bogeyd6, as i said , let s say you install something from ppa or medibuntu, which is a common case [15:38] I'm talking about the set of packages that are installed and are not in apt. [15:38] ppa and medibuntu will propose update to karmic, but still on dist upgrade, ubuntu will disable these repositories [15:38] orogor: you must have a bloody strange dist-upgrade. [15:38] thus unless you manually re enable the repository, all the packes installed from theses are left with no upgrade path [15:39] really thats what it does here [15:39] it disable everything except the canninicall aproved/maintained packages, (can t remember the name) [15:39] I hate ubuntu [15:40] i specially hate the network manager, it s plain bogus and should be removed [15:41] orogor: FWIW, it's not installed by default unless you use a desktop CD [15:41] I especially liked how network manager + nis ==> 30 minutes to boot [15:42] i specially like how it destroy my network config at each reboot [15:42] i tried all the official settings to solve this , it seems there s no issue [15:42] if you define your network interface in /etc/network/interfaces NM won't touch it [15:42] it does [15:43] then that's a bug [15:43] thats why i say it s bogus [15:43] and it has always been [15:43] have you tracked down the cause? reported a bug? [15:43] Unfortunately, ubiquity doesn't define wired interfaces in /etc/network/interfaces by default, so NM gets its fingers into them [15:44] nope, still i bet it s a known one, all the peoples i know since ever have this issue [15:44] which bug? [15:44] well for my special case, sometime it goes differently [15:45] a nix with a static ip , plusgged on a network with a dhcp server, use the dhcp provided adress and not the static ip [15:45] why not do a dhcp reservation? [15:45] it becomes very problematic, when the computer with the nic witht he static ip is a net gateway [15:46] jcastro: maybe he bought a $20 router and its DHCP server can't do reservations [15:46] i bet because int he first place it should use the static ip and not the dhcp one [15:47] still now it s less an issue , before, at every dhcp lease renew it would get a new ip , now it s only at boot [15:47] orogor: just purge NM [15:48] did that [15:48] but still that shouldn t be [15:48] it has other issues with wireless, of the same kind [15:49] basically it keeps overidding manually set settings with some other ones it thinks is best [15:49] orogor: that's its job :-/ [15:50] no , there s auto, dhcp, and whetever parameters it should do that only if theses are set [15:51] btw, anyone with amd64 is having problem with amarok crashing on startup under karmic? [15:51] New bug: #502262 in autofs (main) "autofs4 doesn't timeout cifs mounted shares" [Medium,Incomplete] https://launchpad.net/bugs/502262 [15:51] That's a desktop issue [15:52] yup, but maybe you also have a desktop computer :) [15:52] Nope [15:52] Sorry [15:52] duh [16:02] New bug: #505178 in likewise-open (main) "error while launching in kde" [Undecided,New] https://launchpad.net/bugs/505178 [16:03] any way to get this notification turned off? "You have new mail in /var/mail/david" [16:05] Italian_Plumber: yes [16:07] Thanks for answering my question. :) I'll ask another. How do I get this notification turned off? "You have new mail in /var/mail/david" [16:07] I don't remember [16:10] hello! [16:11] i was wondering if someone can give some info on how the ubuntu cloud is making sure it doesn't loose any data when some bare metal dies [16:11] New bug: #502762 in samba (main) "[karmic] winbind installation causes apt to fail" [Low,Incomplete] https://launchpad.net/bugs/502762 [16:12] niekniekniek: presumably, the same way RAID5 does [16:13] i.e. redundancy [16:13] yeah i know it has walrus and stuff [16:13] but i would like to know how it works [16:14] what if 5 pc's all have 80 GB's harddisk space [16:14] how many can die before one looses data [16:14] and how does it perform in terms of live migration when such an event occurs [16:14] stuff like that :) [16:16] anyone? [16:16] did you try wikipedia? [16:17] Here's what you really need to know abaout RAID, which will put you above my boss: If you have a RAID array, keep spare HDs around, so that when one fails, you actually have one to hot-swap! [16:18] Oh... and that spare laying around? It can't be one that was pulled out of that same RAID array when it failed last year. [16:21] i know what RAID is... [16:22] when creating a cloud i wouldn't like to be dependant of one machine with RAID [16:22] Probably a netapp or something.. [16:24] but that netapp would also need a spare [16:24] i'm not sure if ubuntu cloud uses that way... maybe it uses all the storage in all the vm hosts as storage [16:24] it seems it uses ata over ethernet [16:58] niekniekniek, you need to check into cloud data striping and http://nicolas.barcet.com/drupal/files/SkillsMatter-WhatIsUbuntuCloud_0.pdf [17:04] is there support for LVM volume snapshot merging in 9.10? i saw a launchpad request for inclusion in 8.10 but can't seem to find anything else about it... [17:05] is there a website that shows what kinds of things you do with cloud? === dendrobates is now known as dendro-afk [17:06] bogeyd6: Most things? [17:06] well more detailed like uhm [17:06] is it similiar to vmwares vmotion? [17:08] niekniekniek: I /think/, that if the storage controller dies, then the cloud nodes will run off their local copy of it. === dendro-afk is now known as dendrobates [17:21] twb: I read a little bit about this sheevaplug thingy, and as I understand the last version to run on it is Ubuntu 9.04, and after that that's it ? no more (up-to-date) ubuntu on that? [17:23] I have been searching everywhere but I can't find out if there is a possible solution to replace AD in windows server with openldap and samba in ubuntu, and is it posible to apply GPO [17:23] From a ubuntu server for a windows client? [17:34] Anybody> [17:34] ? [17:34] bogeyd6, jpds, thanks! [17:35] jpds: i'll try to find out if it does [17:35] bogeyd6: i'll give it a read === dendrobates is now known as dendro-afk [17:38] jpds: come to think of it, i guess it is not like that [17:38] because live migration won't work if there's a local copy [17:38] so i guess no local copy [17:38] ? [18:22] ok i read some more === deegee_1 is now known as deegee [18:22] seems to me the ubuntu cloud has a single point of failure [18:22] http://www.ubuntu.com/system/files/UbuntuEnterpriseCloudWP-Architecture-20090820.pdf [18:23] the CLC/WS3 [18:23] anyone? [18:46] hi..are there any application whihc i can deploy window xp and linux client desktop image on a multiple client pc on the network..installation purpose and image a desktop os [18:49] ruben23: that would be clonezilla [18:49] niekniekniek: both for windows and linux OS..? [18:50] Features of Clonezilla: Therefore you can clone GNU/Linux, MS windows and Intel-based Mac OS, no matter it's 32-bit (x86) or 64-bit (x86-64) OS. For these file systems, only used blocks in partition are saved and restored. For unsupported file system, sector-to-sector copy is done by dd in Clonezilla. [18:51] sweet feature: Multicast is supported in Clonezilla SE, which is suitable for massively clone. You can also remotely use it to save or restore a bunch of computers if PXE and Wake-on-LAN are supported in your clients. [18:53] niekniekniek:ow ok very nice..your using it..? [18:54] im new to this, do you have somehow a kind of how to link.. [18:54] have you tried google? http://www.clonezilla.org/ [18:55] if you have a specific question i can maybe answer it [18:57] and yes, i've used it in the past.. back then it required some more handwork then i expected, but it worked ok [19:00] where should i go to learn more of the ubuntu cloud system? [19:16] New bug: #459403 in ubuntu-docs (main) "OpenLDAP server instructions out of date: slapd no longer creates initial directory (dup-of: 463684)" [Undecided,Fix released] https://launchpad.net/bugs/459403 === dendro-afk is now known as dendrobates [20:38] I just noticed karmic has python 2.6, and no 2.5. So ubuntu is fully upgraded to the latest python? There are no 2.5 dependencies?? Im surprised. I thought it would be more work to get things working on 2.6 since it doesn't maintain backwards compatibility. [20:41] err, that would be python3. nevermind. [20:41] Also Karmic has 2.5. It's Lucid that has only 2.6. === dendrobates is now known as dendro-afk [21:12] ScottK: lucid still has python2.5. [21:13] ScottK: and python2.4, apparantly. === dendro-afk is now known as dendrobates [21:21] New bug: #505278 in openssh (main) "ssh-add -D deleting all identities does not work. Also, why are all identities auto-added?" [Undecided,New] https://launchpad.net/bugs/505278 === mcasadevall is now known as NCommander === dendrobates is now known as dendro-afk [23:26] New bug: #505297 in clamav (main) "package clamav-base 0.95.3+dfsg-1ubuntu0.09.04 failed to install/upgrade: id: clamav: No such user" [Undecided,Incomplete] https://launchpad.net/bugs/505297 [23:29] I'm trying to use backuppc for the first time and I'm getting this error "tree connect failed: NT_STATUS_BAD_NETWORK_NAME" and I'm wondering what I should do [23:30] does it mean I have the wrong workgroup name? [23:44] is anyone else using backuppc? [23:54] metalf8801: hm. looks good - why not combine that with an opensolaris server with dedup+compression :) [23:55] I can't get it to work on my Ubuntu server which is what I want to do first [23:57] RoyK: what do you mean it looks good? its not working [23:58] haven't tried it, but it looked good [23:58] http://i.imgur.com/UEeKS.jpg lol