[00:11] got pygrub trouble- trying to boot karmic 9.10 VM (pv) & get "no bootloader data" error. Can I add a psuedo grub1 setup - menu.lst - and get pygrub to work? [00:13] the dom0 is suse11.2 and my /boot is ext3. I want pygrub to use the /boot in domU [00:14] it seems silly to uninstall grub2, install grub1 when all pygrub needs is the menu.lst [00:19] if an nfs mount is 755 owned by www-data:www-data, what permission would allow any user to touch any file owned by www-data:www-data? [00:25] so i have a "special" version of xml-rpc-c installed (the advanced tree) - how can i uninstall it? 'make uninstall' did not work, and i've never really uninstalled anything i've compiled... === erichammond1 is now known as erichammond === erichammond2 is now known as erichammond === rberger_ is now known as rberger === FFForever is now known as FFForever-Away [02:06] nam-myoho-renge-kyo [02:45] Hi Ubuntees, I need help resolving dependency issue, I'm trying to install testdrive but apt-get complains : testdrive: Depends: cpu-checker but it is not installable. Running Ubuntu 9.10 64bit [02:47] livingroom: And why i it not installable? [02:47] KB1JWQ, exactly [02:48] Package cpu-checker is not available, but is referred to by another package. [02:48] This may mean that the package is missing, has been obsoleted, or [02:48] is only available from another source [02:48] E: Package cpu-checker has no installation candidate [02:48] Use a pastebin. [02:48] And what sources are you using? [02:49] ppa.launchpad.net/testdrive/ppa [02:50] for testdrive package [02:50] do you want me to pastebin sources.list? [02:51] and thank you for you interest [02:51] The only references I see to cpu-checker are in Lucid build repos. [02:52] So "WTF are you doing running pre-release packages on a server and expecting it to work halfway decently?" :-) [02:53] that is not what I am trying to do [02:55] as I've heard on ubuntu podcast and read on https://launchpad.net/testdrive testdrive will allow me to test development releases in virt enviroment [02:55] in an easy manner [02:56] thus allowing me to give back my 5 cents to development team and community [02:56] may i ask same question lvs/tun here? [02:57] so was I mistaken my understanding? [03:56] New bug: #537099 in postfix (main) "package postfix 2.6.5-3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/537099 [04:44] I have a Squid question. [04:45] We have some systems connecting to Amazon.com affiliate APIs via Curl in PHP. And it's hammering our server, we determined. [04:45] However, the data it brings back really only needs to be done like once a day. [04:46] Is this something we can use Squid for? [04:46] For instance, I could put amazon.com in /etc/hosts, pointing to the Squid server. And then I'm thinking we could do some kind of Squid configuration to cache those connections to Amazon.com by the day. [04:49] vinq1: why wouldn't you, you know, fix your PHP scripts? [04:50] Just run curl in a cron job, writing to e.g. /var/lib/local-copy.txt, and then have your PHP scripts access that. Or similar. [04:50] Ah, it's spaghetti code, unfortunately [04:51] I'm needing a quick-hit solution where we don't need to change the code for now. [04:51] If you were to use squid to do it transparently, you'd have to care about caching heuristics and whether the amazon API claimed the data was cachable (it probably doesn't) and all that. [04:51] Because I'm tied up coding another project. [04:51] Well, it doesn't hurt to try in an off-hours timeframe so that I could see if it would work. [04:52] Of course, burning the PHP to the ground and using something like Python or Erlang would also be a Good Move. [04:52] Well, we can have that religious conversation another time. So, I'm trying to find a simple (as can be) tutorial on Squid for this type of config. [04:55] I'm afraid I can't really help with that. [04:55] What's a guide on the web for just caching one domain with Squid? [04:56] And do I need to set this up as a Transparent Proxy (not certain yet what that means)? [04:56] If PHP is running curl(1) instead of libcurl, you could just set the proxy in ~www-data/.curlrc, make squid bind only to the loopback interface, and then force squid to cache harder as necessary. [04:56] Interesting! :) [04:57] The whole technique is a bodge, IMO. [04:57] The PHP is using the curl() API inside PHP, not shelling out to command-line curl. Would that matter? [04:59] twb: we're just looking for a temp solution to bandaid this just until I can leave the current project I'm on to put this fire out -- the right way -- by rewriting the code entirely. [05:07] grub2 is no obstacle for ub 9.10 PV guest... got around it by adding menu.lst and testing with pygrub [05:48] Is there a way to not have the lamp start when my computer is booted up? I want to start it manually sudo service apache2 start [05:49] any docs on how to do this? [05:51] edit /etc/default/apache2 or run update-rc.d as appropriate. [05:52] Note that these are potentially temporary solutions : at some point the answer will change to "edit /etc/init/apache2", but that doesn't appear to have happened yet. [05:52] Obviously, you may or may not want to do the same thing for your database, etc. [06:05] how much of a system load does it use if I leave it alone? [06:07] Kutakizukari: Hrm? A system doing nothing has no load. Some things that you install will start services, some of which may have recurring tasks. Some of what you install may install cronjobs that may cause actions at specific times. [06:09] persia, thank you for the help. [06:11] Just got the LAMP installed with Netbeans IDE and a copy of O'Reilly Learning php, mysql, & javascript. [06:12] guys, is there any snmpagent i can install on my ubuntu-server which tracks all network info so that I can use it in mrtg? [06:14] swift: `apt-cache search snmp` shows a bundle of stuff (I haven't tried any of it) [06:18] persia, the issue is that one of my routers doesn't have SNMP features.. and i want to monitor it from my ubuntu-server via mrtg [06:19] so you want something to ping it every once in a while, and to get that info into mrtg? [06:20] I suspect you'll find some example that's close to what you seek in the mrtg-contrib package. [06:20] persia, it's a router cum modem.... so, details including transfers, etc [06:22] So, if it doesn't due SNMP, how do you expect to extract that data? [06:50] persia, so i won't be able to get data from that router? [06:51] swift: I didn't say that. I only suggest that if you want data from the router, you need to be able to collect it from the router *somehow*. Getting that into MRTG is a secondary problem. [06:52] ok, persia, any insight on how i can get the data? [06:53] Check the documentation for the device from which you want data. [06:53] Many devices can enable SNMP is coaxed. Many that can't have some other admin interface that allows data extraction. [06:53] s/is/if/ [06:54] ?? [06:54] that was confusing.. please elaborate [06:56] So, read the docs on the device, and look for other sources of info on the device. [06:56] Some devices don't do SNMP by default, but can be made to do SNMP if certain actions are taken. [06:56] Other devices can collect data in some other admnistrative interface, and it's potentially possible to script extraction of this data. [06:57] I can't usefully elaborate more unless you happen to have a device with which I'm familiar (and none currently exist, as it's been that long since I played with SNMP) [07:58] thanks a lot persia!.. il look into the device docs.. [08:00] swift: Good luck. [08:10] ttx, around ? [08:10] smoser: yo [08:10] bad news on https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/531494 [08:10] Launchpad bug 531494 in upstart "cloud-init job not running in eucalyptus without ramdisk" [Critical,Confirmed] [08:11] it is almost definitely race, and seems like just about anything can kick it (although nothing i've found yet is reproducibly does so) [08:12] smoser: want me to reproduce it on my own set ? [08:13] i'd be interested to see if you can. [08:13] ok [08:13] i'm guessing your hardware is more like dustins [08:13] yes, quite identical [08:13] What hardware are you using out of interest? [08:14] ttx and kirkland have "cloud in a bag" hardware [08:14] Omahn: a set of laptops [08:15] wow. the link from eu-west-1 to ubuntu data center is quick [08:16] 53.9M/s over 160M [08:16] and prior to that 46M or something [08:18] anyone know how to use iptables to filter traffic to my bridge's KVM based VMs? [08:18] smoser: The disk is probably the bottle neck rather than the network link looking at those numbers. [08:18] it seems like no matter how I configure iptables, I can get to any of the VM's ports. [08:22] <_ruben> you might need to use ebtables instead for filtering on bridges [08:24] _ruben, what I want to do is only have iptables running on the host, not on the VMs. And be able to forward port 80 of a VM to port 8080 on a VM [08:25] I'm running Jetty as a regular user in a VM, and I want that service to be available as a regular HTTP server to the world. [08:26] _ruben, I don't think ebtables will help with that, would it? [08:28] <_ruben> tauren: now you got me confused? do you want to restrict the allowed traffic to the vms or grant outside access to it (and thus liberating the allowed traffic to the vms) [08:28] <_ruben> ? [08:29] _ruben, sorry. for my VM that is running Jetty, I want to only allow ssh, http, and 8080 traffic. But I want to forward 80 to 8080. [08:29] does that make sense? [08:29] for my VM that is running mysql, I want to only allow ssh and 3306. [08:30] if anyone wants to play http://paste.ubuntu.com/393120/ [08:30] that has ebs root daily builds for lucid from today [08:31] smoser: should I use the latest uec iamge, or something else ? [08:31] _ruben, I want to block it at the host, because these VMs will be under control of customers, and I want to limit what they can do. [08:31] the latest will fail. [08:31] i was just using 20100310 for consistency in an inconsistent world [08:31] 20100311? [08:31] sure [08:32] why would it fail ? [08:32] it will fail because of upstart and timing and your hardware being like kirklands [08:32] it will run fine on data center hardware [08:32] smoser: that's what I'm trying to confirm, right [08:33] right [08:33] let's go for 20100310 for consistency [08:33] yeah [08:36] <_ruben> tauren: ah ok, so you actually want both things i mentioned ;-) ... then again, i have near-zero experience with (filtering on) bridges.. [08:37] _ruben: ok, too bad, i appreciate your input! [08:37] ttx, i have to sleep. any thoughts you can provide or help any where on that would be great. i don't have high hopes for it for freeze today. [08:37] smoser: it's a beta1 nominated bug anyway, so we can still work on it [08:38] smoser: good "night" [08:38] yeah. thanks. [09:00] any ubuntu documentation available for mysql db replication ? [09:06] New bug: #537191 in apache2 (main) "package apache2 2.2.12-1ubuntu2.1 failed to install/upgrade: η υποδιεργασία dpkg-deb --control επέστρεψε κατάσταση λάθους 2" [Undecided,Incomplete] https://launchpad.net/bugs/537191 [09:18] smoser: it's working alright here. [09:18] and I'm in the worst case scenario / running CLC+Walrus+CC+SC on the same machine [09:24] hm, or not. [09:24] smoser: I can reproduce it by starting multiple instances at the same time. Some of them will fail. [09:25] ttx: thats not worst case.. you don't have nc there :) [09:25] NC on a separate machine :) [09:25] is the release date of ubuntu 10.04 already announced? [09:25] Dr_Jekyll: yes, we also know the release date of 10.10 [09:26] 10.04 should be released on April 29th. [09:26] alright, thanks! [09:51] what's the recommended kernel to load under vmware in ubuntu server 9.10? [09:51] as a guest os [09:51] 2.6.31 [09:55] anyone including vmtools === uvirtbot` is now known as uvirtbot [09:57] I am already running 2.6.31-19-generic-pae [09:57] I don't know how to install vmtools though [09:57] I tried to get the host to mount the image to the guest [09:58] not working for some reason [09:58] [x] connected ? [09:58] ^^ [09:58] th0mz: Error: "^" is not a valid command. [09:59] checking [10:00] vmtools do something with the network card right? [10:00] you can use vmnet3 [10:00] I'm running server, not desktop edition [10:00] I just need more performance from the network [10:00] but it increase performance [10:00] (try a dd before and after) [10:00] great. that's what I need. [10:00] server shout not run without [10:01] and if it is in a cluster it is really needed, (ballooning etc..) [10:01] ouch. yes… believe me I feel it ;-) [10:01] great pain. [10:01] I need to find a simple way to mount the image with the tools [10:01] right click, install vmware tool shoud do the trick. [10:02] when I do that, the vm gets stuck for 70+ seconds [10:02] [ 1113.022505] BUG: soft lockup - CPU#0 stuck for 77s! [swapper:0] [10:02] [ 1113.022505] Pid: 0, comm: swapper Not tainted (2.6.31-19-generic-pae #56-Ubuntu) VMware Virtual Platform [10:02] and a bunch of kernel-panic looking stuff [10:03] so that method is scary and not working :-) [10:06] any suggestions how to obtain the vmtools without using the standard method? [10:06] download, install [10:06] :] [10:10] how about open-vm-tools is that as good as? [10:11] I ran an apt-chache search vmware and found this [10:11] open-vm-tools - tools and components for VMware guest systems (CLI tools) [10:11] whoah, it wants to install a gazillion things [10:11] haha [10:42] New bug: #341979 in kvm (main) "Windows guest can't reboot when using smp." [Medium,Fix released] https://launchpad.net/bugs/341979 [10:52] cjwatson, I noticed you've been uploading to the libvirt package a bit lately. I was just wondering if you could have a look at my bug at LP #528934 and consider applying it? I've attached a debdiff, which I believe makes it easier for you. [10:52] Launchpad bug 528934 in libvirt "IPv6 shouldn’t be disabled by default in libvirt" [Wishlist,Triaged] https://launchpad.net/bugs/528934 [10:53] or soren. Whoever is interested. :) [11:33] jayvee: the only reason I've been uploading libvirt is for the libparted ABI transition - I'm not going to be a good reviewer for changes to it in general [11:33] sorry [11:33] np [11:36] Launchpad nicks usually == irc nicks, but I wish Launchpad would generate nicks from the e-mails of people in the changelogs to make people easier to track down. :) [11:40] jayvee: I, on the other hand, know libvirt reasonably well, but I know next to nothing about ipv6. :( [11:41] heh [11:41] soren: but you can see that libvirt disabling IPv6 on an interface is plain inconsiderate, right? [11:42] sure, there are arguments for and against the behaviour red hat put in it [11:42] It sounds inconsiderate, yes. [11:42] but I believe that the only scenario where you would disable IPv6 is if you're scared of it — it's an irrational fear [11:42] Yup. That's what I have. [11:42] :) [11:43] ..so I'mm kjkknik,jjjjjjjjjjjjjjnmmmmmmmmmmmmmmmmmmmmm mm [11:43] Sorry, had to get a thing out from under my m-key :) [11:44] ..so I'm kind of hesitant to change anything that as anything to do with ipv6. I have no idea of the sideeffects. [11:45] the funny thing is, is that *now* it's doing something to do with IPv6. my patch actually makes it have less to do with IPv6 by leaving it more alone. [11:45] put it this way: [11:45] That I understand. [11:45] all ubuntu installs since 4.10 have had IPv6 enabled on every network interface by default [11:45] at least, I think so :) [11:45] as far back as my memory goes anyway [11:46] it's the linux default [11:46] My hesitance kicks in because someone made a decision to make this change. If I go and revert it, /I [11:46] am changing something to do with ipv6. [11:46] * soren notes that typing is hard [11:47] hmm [11:47] what can I do to help, in this case? [11:49] soren: I suppose it would be easier to convince you to upload this patch if upstream had applied it too [11:49] but I've had little response from them [11:49] jayvee: Oh, yeah, sure. Then I'd do it in a heartbeat. [11:49] soren: upstream were nuts [11:49] jayvee: They are on IRC as well. [11:49] soren: if its the bug I think it was, I read the bug in detail. [11:50] https://bugzilla.redhat.com/show_bug.cgi?id=501934 [11:50] yeah [11:50] bugzilla.redhat.com bug 501934 in libvirt "libvirt bridge should have IPv6 disabled" [High,Closed: rawhide] [11:50] soren: it boiled down to 'zomg something wrong turn it all off' [11:50] note that Peter Bieringer "100% agrees" with my comment, and he wrote the IPv6 tutorial for Linux on TLDP [11:50] so there's some credibility there :) [11:50] jayvee: Now, that is useful info :) [11:50] anyone know how to display the ipnumbers used by kvm guests on the host? doing "virsh list --all" shows the hosts, but how can I find their IP numbers? [11:51] tauren: You can't. [11:51] soren, no? [11:51] I do fear I used too harsh language in my comments on the RH bugtracker though. [11:51] tauren: Just like you can't tell what the IP's are of your physical machines by looking at the box. [11:51] as you can see, IPv6 is something I'm passionate about. :) [11:51] hmm, so I have to log into each host and do an ifconfig to see them? [11:51] tauren: Are they using dhcp? [11:51] nope [11:51] tauren: How did you configure their network? [11:52] soren, bridge eth0 with static ips in the guests [11:52] tauren: Yes, but how? [11:52] tauren: Where did you enter the static ips? [11:52] Inside the virtual machine, right? [11:53] soren, in /etc/network/interfaces on each guest [11:53] yes [11:53] Right. [11:53] So that's where you go to find them. [11:53] well, i used vmbuilder to create them and specified the IPs in the command line. [11:53] I assume it put them into the interfaces file [11:53] It did. [11:54] tauren: may be offtopic, but from what I recall, OpenVZ makes it easy to find the guest IP addresses like what you describe [11:54] jayvee: openvz does not provide virtual machines. [11:54] not having used OpenVZ, take that with a grain of salt [11:54] * soren is on the phone [11:54] jayvee, funny you say that, i was using openvz before, which is why i wanted the same feature with kvm [11:56] I supposed I could set up DHCP and dole out static IPs based on mac addresses or something [11:56] then i could just look at the dhcp server to see the list of ips. [11:57] yeah [11:57] there is also ping -b, but unfortunately linux doesn't respond to broadcast pings by default [11:57] dang, was really hoping i was just missing a virsh command that would list ips. this is a bummer. [11:57] or use UEC :P [11:58] if I were to retrofit a system to figure out IPs on top of what you already have, I'd enable broadcast pings, and get each machine to register itself in reverse DNS [11:58] that way, ping -b would show all the hostnames with the IPs [11:58] but I think that's over the top for something small ;) [11:58] jayvee: dhcp3-server will do rdns rego for you [11:59] that's what I had in mind [11:59] interesting. ok, those are some ideas then. [11:59] and forward [11:59] but you have to tweak the config file [11:59] yah [11:59] off by default for compat with 3 [11:59] s/3/2 [11:59] and it's really confusing at first [11:59] Is UEC the eucalyptus stuff? [11:59] yeppers [11:59] anyone here looked into openqrm? === linux_guru_mania is now known as linuxgurumaniac [11:59] lifeless: I tried UEC on my desktop...I couldn't get it going. I think it's way too complex for me. [11:59] hello [11:59] hello [12:00] or it's just not in my use case [12:00] i tried doing a plain UEC install from CD a while back and it didn't work. didn't put any more effort into it. [12:01] * jayvee just installed MySQL in CentOS, and it looks like it can install i386 and x86_64 libs on the same system cleanly [12:01] * jayvee can't wait for debian multi arch [12:02] it's a shame, because UEC seems to be really cluster oriented, which I was really excited about [12:02] I tried to get some cluster dns script going with nsupdate, but I gave up because it was too difficult to get it working with upstart. [12:02] upstart is designed for daemons, not scripts. [12:03] in a cluster scenario, is it best to get machines to use dhcp addresses? [12:05] jayvee, soren, lifeless: the reason I want to know the IP numbers on the host is so that I configure iptables on the host to protect the guests. ie block everything except 22, 80, and 8080. Plus port forward from 80 to 8080 on the guests. [12:05] jayvee: single machine is a little tricky due to bugs [12:05] jayvee: did you follow my docs on the same (and do you ave a separate machine doing DNS/DHCP ? [12:06] lifeless: what is the recommended config for doing UEC? is 2 machines enough? [12:06] i was trying to set it up on a single machine too, maybe that's why it didn't work. [12:07] tauren: recommended config is three machines: 1) network infrastructure. 2) cluster controller 3) 1 or more node controllers [12:08] I have successfully brought that down to do two , by combining 2 and 3, though I ran into some bugs [12:08] * persia wonders if the IPv6 discussion got lost [12:08] indeed it did [12:08] lifeless, can cluster controller run in a VM? [12:08] tauren: I'm working on bringing it down to one, but that needs some more care. [12:08] tauren: I think it could yes. But not a UEC one :P [12:08] lifeless, and what exactly is network infrastructure? [12:08] tauren: DHCP, DNS [12:08] ahh [12:09] can UEC guests do PXE boots? [12:09] no, they boot 'machine images' [12:09] I'm really skeptical that UEC is at all production-ready if it's not even been tried and tested under a single server condition [12:09] its EC2 for home. [12:09] jayvee: its not a use case the devs aim at. [12:10] jayvee: they are aiming at 10s or hundreds of machine deployments [12:10] I've heard it works cleanly with 3 machines, and is tested at 2. One is a special case. [12:10] When I compared UEC with OpenQRM, it seemed OpenQRM was the better solution. [12:11] it too recommends 3 servers: storage server, openqrm management server, node server [12:11] but they have a way to run it all on a single system. [12:11] anyway, in case you want to check it out. [12:12] it's discussions like these that make me want to power up the DL320 in the garage [12:12] but we worked out the power cost us about $45 a month so it's been off since 2007 [12:14] jayvee: yeah, I had some of those but had to give them away due excessive noise and power consumption :( [12:14] yeah, that's how I got all mine :) [12:15] jayvee, how much power does it consume? [12:15] don't remember [12:15] though it'd be around a kilowatt, I'd imagine [12:16] i just installed a supermicro twin 1U system. It contains two servers in 1U, each has dual quad CPUs, 32GB ram. I'm sure it sucks up the power. [12:22] just checked my notes. supermicro says they tested that server in some configuration and it used 2783 BTU/hour. I think that works out to about 6.8A. (1W = 3.413 BTU/hr, 815W/120V =~ 6.8A). Sorry, getting OT now. [12:27] is there a easy power management package for Ubuntu servers to shutdown and start automatically? i have tried APCI and WOL === manic is now known as manicman [12:37] Hi - I'm trying to vnc into my headless ubuntu-server hardy box [12:37] I was following the directions from here: http://stevenharman.net/blog/archive/2008/12/13/vnc-to-a-headless-ubuntu-box.aspx [12:37] But all I get is a brown screen [12:37] Does anybody have any idea what might be wrong? [12:37] I'm thinking I'm probably looking at a blank desktop and need to install some packages for a minimal gui [12:37] but I'm loathe to install the ubuntu-desktop packages [12:38] since that would be such overkill for what I need [12:40] gzur: i have no ideas for your specific problem. but why don't you just ssh in to the box? [12:52] morning === jan247_ is now known as jan247 [13:00] zul: hey [13:01] Samba bugzapping: created https://code.launchpad.net/~bug-zappers/ubuntu/lucid/samba/bugzapping [13:03] ttx: sweet [13:04] ttx: still trying to wake up ;) [13:12] so yeah, soren, what do you think? :) [13:12] jayvee: It makes sense to me. [13:12] that's a good start. :) [13:12] ttx: Have you followed the ipv6/libvirt discussion at all? [13:12] ttx: A few hours ago. [13:12] or less, perhaps. [13:13] My sense of time is poor right now. [13:13] 90 minutes or so, by my logs [13:13] nearly 1.5 hours ago [13:13] tauren: I mostly use SSH - but I've got a spatial database (postgis) running on there that I sometimes need to interact with visually using a desktop GIS (quantum gis)- which is hard to to in vim :) [13:13] soren: no [13:14] soren: looking [13:14] ttx: Ok. [13:14] ttx: Basically, jayvee has a tested patch to make libvirt work with IPv6, but the upstream bug got messy, and there's a debdiff awaiting upload. [13:14] persia: Would you say it requires an FFe? [13:14] * persia isn't a release person [13:14] persia: I know. [13:14] persia: Regardless. [13:14] persia: I can't really tell if it's as much a new feature as a bug fix. [13:14] But I'd probably upload it without bothering the release team, because I consider not working with IPv6 a bug. [13:15] It's borderline. Depends on whether you believe IPv6 should work by default. [13:15] persia: I'm inclined to agree. I just wanted someone else to say so as well. [13:15] If you say it should, it's a bugfix. If you consider IPv6 a feature, it's a feature. [13:15] * persia knows that there exist IPv6-only production deployments of Ubuntu server [13:15] Wow. [13:16] Check the ipv6 bugtag : seems lots of folks don't even have IPv4 [13:16] That's crazy. That's like not having... [13:16] err... [13:16] something else that almost everyone has. [13:17] for my own personal use, I go without IPv4 all the time [13:17] it's like eating my own dogfood, I spose [13:17] jayvee: /whois reveals you're doing IRC over ipv4 at least :) [13:17] There's a limit to IPv4 address space. In some places, addresses are expensive. [13:17] soren, yes, I bricked my router yesterday [13:18] jayvee: Coulnd't handle the ipv6 address space? I know the feeling. [13:18] That thing is huge! [13:18] jayvee: Do you have the bug number handy [13:18] ? [13:18] soren: I am scared of both ipv6 and libvirt, if that helps :) [13:18] LP #528934 [13:18] Launchpad bug 528934 in libvirt "IPv6 shouldn’t be disabled by default in libvirt" [Wishlist,Triaged] https://launchpad.net/bugs/528934 [13:19] ttx: No, that's surprisingly unhelpful :) [13:19] jayvee: Ta [13:19] heh, actually I tftp'ed a 3.3 MB image when it was only supposed to be 3 MB. had to hotwire my parallel port to the router motherboard to fix it. I'll be telling this story for the next few weeks, I'm sure. [13:20] and because the router now has the 2.6 kernel, the bridging quirk that I used to deliver a second pppoe session to my ipv6-enabled ubuntu server no longer works. [13:20] that's my guess, anyway. been more focused on getting this apache solr installation working since I got it going again. [13:21] which doesn't support ipv6 out-of-the-box, btw. [13:21] ipv6++ [13:21] jayvee: I'm fixing these things before I upload, but just so you know: [13:21] The patch will not be applied if not listed in debian/patches/series. [13:21] ...so I've added it there. [13:22] I've also renumbered it to 9017 rather than 0011. [13:22] ooh [13:22] Debian tends to number them from 0000, and for patches that are local to Ubuntu, we tend to use 9000+. [13:22] I see [13:22] I didn't know that. [13:22] ..that way, if we merge with Debian, and they've added patches, our are still applied "on top", so to speak. [13:23] in my testing, the patch worked locally though. Maybe debuild and pbuilder behave differnetly? [13:23] The patch numbering bit tends to be team- or even package- specific. [13:23] * soren verifies that the package uses quilt.. [13:24] jayvee: No. [13:24] Well, yes, but not usually about patches :) [13:24] jayvee: Neither of those deal directly with the patches. [13:25] jayvee: How did you construct the patch? [13:25] (that may explain why you've seen it work) [13:25] well I first constructed it with a git checkout of libvirt, and tested from there [13:26] and then downloaded the libvirt source [13:26] ran cdbs-edit-patch or whatever it's called [13:26] made my change [13:26] and added the changelog entry [13:26] and tested the resulting .debs [13:26] Heh :) [13:26] cdbs-edit-patch would do it. [13:26] which definitely worked [13:26] Funny that. [13:27] I know it worked, because my ipv6 broke again when I installed the karmic version of libvirt again this morning [13:27] oh...when I rerolled the debdiff, I simply copied over the patch into debian/patches [13:27] and updated the changelog [13:27] so that was done manually [13:28] and that's why it didn't end up in series [13:28] aha [13:28]