[00:06] zul: jamespage: any recommendations for getting wider testing of the qemu-kvm proposed packages in ppa? I may just put up a blog post, wait 1 week, then push to raring... [00:22] Anyone got experience with roundcube on ubuntu 12.04 [00:23] Ran into some problems configuring roundcube with apache... [00:32] hello masters! [00:33] i search to know if it exist a version of openldap for 12.04, more recent than the version openldap (2.4.28) package [00:33] thanks very much! [00:35] What is a good way to trouble shoot why certain people cannot connect to my server some can connect others cannot [00:35] paco11: 2.4.28-1.1ubuntu4.2 looks like the most recent for 12.04 LTS; a newer version is in 12.10 [00:35] Danawar: define "connect" [00:36] Teamspeak clients and minecraft clients some cannot resolve lagcraft.co.uk [00:36] Im gussing [00:37] Im not sure if it is a propergation problem or a setup problem [00:37] do your clients tell you the error message is "cannot resolve"? Or is their error message vague enough that you don't know what the real problem is? [00:37] sarnold: can i install the 12.10 version (2.4.31-1ubuntu2: amd64 i386) on my 12.04 server? [00:38] paco11: check to see if -backports has a newer version compiled for the LTS release available, first [00:38] sarnold: minecraft doesnt give good error messages sadly =[ [00:38] paco11: maybe it'd work no trouble, maybe it would be no end of trouble. :( [00:38] Danawar: ouch :( [00:39] Im wondering if .cloudflare. which i signed up to and added to my name servers today [00:39] Might be intercepting and droping the legit traffic [00:39] Danawar: how long was your dns cache time on your old dns hosting? [00:39] As it has had a good part of the day to propergate and people were using the teamspeak 2 hours ago [00:40] How would i find that out sarnold i use .cloudflare. freedns and brought the domain on 123-reg# [00:40] sarnold: precise-updates: 2.4.28-1.1ubuntu4.2: amd64 i386 | precise-backports: Sorry, your search gave no results :( [00:40] Danawar: do you have your old zonefiles? [00:42] sarnold: Sorry i wouldnt even know how to get these i am new to hosting websites and all these name servers =p [00:42] paco11: if you've got a problem with the 12.04 LTS version, you may wish to file a bugreport with your reason for wanting the newer one... [00:42] Danawar: ah. :) well, allowed cache times are quite often less than a day... [00:43] sarnold: no, i don't have any problem, also, i want to use 'mdb' instead of 'hdb' backend [00:43] sarnold: i will wait a day so it has more time to propigate with cloadflare luckly the site is not mission critical just play about =D but i hope it doesnt take this long in the future because we have a dynamic ip :*( [00:44] paco11: that sounds like a reasonable enough problem :) the amount of effort to do an SRU may be more than you're willing to do, but still, filing a wishlist bug report won't take too long [00:45] Danawar: oh -- then make sure your current dns cache timeout is something short, like ten minutes or less.. it'll increase the load on your dns servers, but may save your users some real hassle in the future [00:47] sarnold: a bug report to ask for the creation of the package 2.4.33 of openldap for 12.04 to use mdb? [00:47] sarnold: when you say my dns cache where is this located local machine or on my hosting company? [00:48] paco11: sure. it might be rejected, but if you provide a reasonable reason, it might yet start the SRU process rolling.. [00:49] Danawar: they're located all over the world, at ISPs and in every end user's phone, laptop, desktop, etc. :) === daniel_-_ is now known as daniel_- [01:01] * Patrickdk wonders why people think dns *propigates* [01:02] I think dns is the only thing google hasn't started indexing [01:07] Patrickdk: I've been content to let that slight misconception sit, since it seems to do no real harm, and it's not too far from the truth anyhow [01:08] truth/how it's observed [01:09] but propigates also implies when the *master* is down, changes won't go away, but if ns servers go down, it will go away :) [01:19] New bug: #1014872 in mysql-5.5 (main) "no debug symbols for mysql-server" [High,Fix released] https://launchpad.net/bugs/1014872 [01:19] New bug: #1062716 in mysql-5.5 (main) "Regression in privileges of mysql debian-sys-maint user" [High,Fix released] https://launchpad.net/bugs/1062716 === chilicuil_away is now known as chilicuil [01:38] sarnold: i'm testing install openldap-2.4.31 from quantal, because i anylized that the depends packages are the same..... [01:40] sarnold: libldap-2.4-2 is installed.....but to complete openldap-2.4.31 package, it's saying me libldap-2.4-2 is required.....it's crazy..... [01:40] did you insteall the -dev package? [01:41] patdk-lap: a package would depend upon the -dev? [01:41] of libldap? [01:41] hmm [01:42] Hi all. How can I get apt-get to non-interactively deal with this? http://bpaste.net/show/xQUhXWNFR019nMn4qYoO/ [01:42] (this is output from "apt-get upgrade -y" [01:46] Zal: DEBIAN_FRONTEND=noninteractive aptitude safe-upgrade? [01:46] ok, I've tried that environment variable, but not with aptitude, just with apt-get. [01:47] Funny thing is, the script worked fine until today (I guess there was not an upgrade available until today?) [01:47] .. an upgrade that changed a configuration file that you had also changed [01:48] there's a simple dpkg --force-confnew or --force-confold but I don't know how to send thoes command line options through from apt-get .. [01:48] aha, thank you sarnold. I tried force-confold too, no dice. I'll have to revisit those, maybe I did something wrong. [01:52] Hey everyone [01:52] Well, now that I understand the exact nature of the change in behavior, I can at least work around it if I don't find a direct solution. [01:53] (basically by modifying that file *after* upgrade) [01:54] thanks again sarnold, big relief. And thanks paco11. [01:55] Got soem trouble with apache2 here, trying to set up 2 virtual hosts to host two sites on one machine and one IP, however the guides that all seem to say basicallyt the same thing for setup, do not result in working sites or configs for some reason [01:57] Do you have a stanza for each in sites-eneabled? [01:57] Im wondering if there is something really basic im missing that i should do before these configurations are supposed to work [01:58] I have a file in sites available, and used the a2ensite function to add them both, i imagine they should be there [01:58] could be but you did not answer my question [01:58] because i hadnt read it yet lol [01:59] you should have two files one for each site [01:59] corrrect [02:00] whats the quick way to check files in the present directory [02:00] and they include a [02:00] ls [02:00] check sites enabled for sym links back to sites-available [02:01] ok in sites eneabled i have 3 pages one is 000-default [02:01] the other 2 are the ones i set up [02:02] i better check on the virtual host thing, last time i tried supllying it with the WAN ip it errored out [02:02] And you have a NameVirtualHost 173.14.101.201:80 line in default [02:02] now is that IP supposed to the servers specific lan IP or the WAN ip] [02:03] ooooooooooo now you are saying this machine is a router too [02:03] on my machine that is a public address [02:03] not really a WAN address [02:03] now, its a regular server with a network card, the sites are to the WAN Ip and forwarded to the server [02:04] via a router [02:05] Ah. And you want these visible on the outside (public) [02:06] of course [02:06] then you have a router problem first [02:06] im just wondering if server sidei t should be the network card's IP or the WAN ip [02:07] idk about that, but you could be right, i navigate to the default site fine [02:08] netstat -an and see what the deault server is listening on [02:08] lets just try this part out to see if it works, ive changed the virtualhost files to include the IP instead of a *:80 [02:08] *:80 should have worked [02:09] I'd expect *:80 to work better... [02:09] simpler at least [02:10] you do jhave to use the fqdn of the sites to get apache to go to them [02:10] that is what name virtual host means [02:10] you may to rig some entries in hosts [02:10] i have registered domain names, where would i add that to? [02:11] sarnold: with regards to my issue earlier i asked a guy to connect to he server and he was unable to also he tryed the website and that worked! O.o [02:11] where does dns point too? [02:11] to [02:12] the WAN ip, my router basically, and that forwards traffic to the server [02:12] for the two added domain names? [02:12] yes [02:13] Danawar: it's a real shame that minecraft mangles the error messages. it'd almost be worth writing a little tool that tries to connect to the same port that minecraft uses to try to get a real error message back. [02:13] and you can ping those fqdns [02:13] i can double check but i set those up hours ago, i had to change the addreess as i moved physical locations, but both addreess or domain names should point o he current IP [02:13] WAN ip that is [02:13] Indeed, port scanning the server returns the open port and other people are conncting some not which is relly strange [02:14] dig the domain names [02:15] when i enter the doman name for eiher, it brings me to a default apache page, but not the respectful index.html files i set up for the specific virtual host directories [02:15] you could have the old IPs in cache, they could be stale on your dns server [02:15] otherwise it would give me an error, as there is no server running apache at the old wan IP [02:16] http://superuser.com/questions/330586/apache-virtualhost-isnt-serving-a-page [02:17] I think you know what your are doing [02:18] just one detail off [02:18] yeah, I mean its very strange that this is not working, im very confused abotu why it is not [02:18] whats that? [02:19] wish I knew [02:21] I have 50 or so virtual sites running on Ubuntu 12.04 [02:21] I don't mind sharing my config files [02:22] that would be awesome [02:22] ive been frustrated with what to do about this lol [02:23] send me an email - tarvid@ls.net, I'll reply with my config [02:23] TheGuy: .. have you reloaded the apache config? [02:23] good point service apache2 restart [02:25] reloadied it twice, once after each change, i can try reloading it again though, wont hurt if it doesnt work [02:25] hehe, okay.. [02:25] might loook in the log files to see if there were errors on the restart [02:26] ahh, forgot to include the namevirtualhost directive, im not sure where that goes though [02:26] said the two virtual hosts overlap and one takes precedence [02:27] i set that in the 000-default conf right? [02:27] we are getting somewhere [02:27] 000-default is in sites-enabled, you should put it in sites-available [02:28] sites-available/default [02:28] and what should i set for it, is there some special conf for have the two sites? [02:29] only one per IP in default [02:29] then there is an entry in each stanza [02:30] [02:30] ok, so under the virtuahost *:80 i set it, or do i replace that with namevirtualserver ? [02:31] or do i set up two copies of this for each site in the default ? like have two complete virtual host entires in the file [02:31] and then of course which file are we in - NameVirtualHost goes in default i am currently in the default of sites available [02:32] and i see the default page i had copied and edited for the new siet configurations [02:33] I am downloading and sending you some files [02:33] alrighty [02:34] been quite a while trying to get this to work, i have a 3rd site for a non-profit to put up but i dont want to mess with it untill i can reliably run it lol [02:34] and pardon me for a minute, smoke, you guys want some coffee? [02:36] check your mail [02:37] go it [02:37] can I just delete the old kernels in /boot [02:39] ok, isee how hats set up, list he ' NameVirutalHos direcives at he op first [02:39] In default, I use a few for obscure and not entirely rational reasons [02:40] Then there are two key lines in the domain stanza are these if mods going o be something necessary? [02:41] How would I be able to add a mainline kernel to my build? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1077547/comments/43 [02:41] Launchpad bug 1077547 in linux "Kubuntu 12.10 Kernel issues" [Medium,Incomplete] [02:42] Is there a specific repo or ppa that needs to be added? Or do I have to download the files manually and do everything that way? [02:42] Just one line in default and two in the domain stanza (three if you want to alias www [02:44] i should specifiy the document roo though as well correct? [02:45] Yes [02:45] and the Directory [02:45] Asking this in here due to: Linux server 3.5.0-18-generic Otherwise, I would as in the main ubuntu channel [02:45] as well as the directory, changed to the directory that the files for each site are contained in respectfully to the files they are configured by [02:46] Apache will look for index.* in the document root [02:46] the options are overrides to those in default [02:49] hmmm, it seems to be doing the same thing as before... strange, although you use a different directory for the files, mine are in var both e virtual site directories [02:50] I have no idea what that means but they need to be in user space [02:51] readable from apache and where you say they are [02:52] all rather real [02:52] like in /var/www? [02:52] alright, i changed the default again, to have the namevirtualhost ip:80 and then i changed in the virutahost part just under it to the IP:80 and it gave me an error when reloading [02:53] yes. they are in /var/www/example.com sort of format [02:53] as the directories [02:54] I think the virtual part is right [02:54] /var/www/example.com is part of the directory for default [02:54] I'd put them in your home directory [02:55] its warning that the name or service is not known, failed to resolve the servername for ..IP.. (check dns) also, namevirutalhost *:80 has no virtual hosts (though i should have this part changed already [02:55] hmmm, i have my raid set up to be accessed on the var direcory lol [02:55] /var/www is the document root for default [02:55] * is fine for virtuals [02:56] I think /var/www is good only for an info page for the server [02:56] you know what, ill give that a go, i may have to reload the OS and mount the raid on the home directory, but im not sure why that would cause an issue in /var/ww/ [02:57] /var/www is the default document root [02:57] i had a single site fully functional on the direcory last week, but since ive been trying to figure this out it wont work wih a second virtual host for some reason [02:58] i hate clearing the disks on my raid, takes hours lol [02:58] I keep copying a sites-available stanza that works [02:58] I hate RAID [02:59] Had a mirror array, one drive died and RAID would not switch automatically [02:59] then the other drive died [02:59] yeah, i dont really need it for back up, since i can on my personal computer, but its an older server so i am testing a striped raid for faster reads/writes [02:59] bad idea [02:59] new hardware is cheap [03:00] oh wow, i did have one disk go bad on me once, but i rebooted the server, got into the raid config and copied it over no problems [03:00] bought this server for $100 [03:00] $500 with an SSD [03:00] $105 too much [03:00] where did you get yours? [03:00] I was given the RAID server [03:00] It over heated. [03:00] all the ones im looking at would cost me over a $1000 [03:01] Bought an expensive fan [03:01] Thanked the donor profusely. She smiled and said "free as in free cats". [03:02] In the bargain basement I pick up an old desktop from scottsbargains.com [03:02] Ahh i see [03:02] for new kit newegg, cheap barebones, lots of RAM but I am about to switch to SSD [03:03] yeah, webhosting isnt very demanding on hardware from what i can tell, most likely a bandwidth issue rather then hardware speeds [03:03] A new hudson in a book sized case with RAM is well under $200 [03:03] You are correct [03:03] yeah thats where i was shopping, a server set up was looking to be about $1200, more with a barebone, and $600 more to fully fit out to my desire lol [03:04] give up your desires, all you need to do is shovel bytes from hard disk to Ethernet port [03:04] but im looking into game hosting down the road, or similar functions, planning on running VMs, but one step at a time [03:05] You can stuff 8GB on a hudson for about $40 [03:05] SSD will cure your disk transfer desires [03:06] yes but will not cure my wallets butt hurt [03:06] those are getting cheaper, but its something i have to consider heavily, as i do need disk space, im storing videos for one of mysite [03:07] stick the videos on Amazon, let the customer pay [03:08] http://washingtondc.craigslist.org/nva/syd/3376788154.html [03:08] But it is a mistake [03:08] deffinately an option, got a flash player and videos all ready to go though, not hard to set that up on my site, but a youtube or amazon type of deal might be a good option [03:09] And YouTube is free [03:09] tarvid: heh, that's probably less than 1/10th the original price... [03:09] load average on my hudson is about 0.2 [03:09] sarnold, [03:10] huh, that is exceptionally priced [03:10] sarnold, and he has tons of big iron, all a mistake [03:10] tarvid: bought amd instead of intel? [03:10] tarvid: or .. how is it a mistake? :) [03:11] hallyn: emailing ubuntu-server/ubuntu-devel? [03:11] I bought AMD for the memory capacity [03:12] It is not a religion for me [03:12] I'd actually like to try some ARM arrays [03:13] massive arrays of CPUs excite me more than massive CPUs [03:13] and redundance beats RAID [03:13] I had thought about buying a mostly consumer grade computer that supported ECC memory, Im not sure if thats really a viable option though [03:14] Consumer grade has generally fewer SPOFs and better reliability [03:14] Besides two are cheap [03:14] ECC never got me anything [03:15] yeah, i've heard its mostly applicable to large databases and such [03:15] but not worth it for smaller scale [03:16] do the server cpus offer anythign extra in the areas of virtualization? (VMs? ) [03:16] zul: ? [03:16] It takes a lot of bullshit and gullible customers to make money in the business [03:16] IBM is the historic example [03:16] i emailed ubuntu-server last week [03:16] but HP whores too [03:17] VMs are a processor feature [03:17] hyper threading and hyper-V sort of thing, i think consumer grade procs can do this [03:18] I ran VMware, VirtualBox etc. on a consumer comp [03:18] run LXC [03:18] though i can't say my comp liked it much [03:18] still rough but more rational [03:20] Im thinking using the VM method i may be better able to set up the sites lol, though it still baffles me why it isnt curently working [03:21] And you will have too, most of your VMs will consume resources and sit idle [03:21] i suppose ic ould plug in the second ethernet cord for it, but im not sure that would be any easier [03:23] Centerton looks promising [03:27] it's been fun guys but time for me to get my feet up [03:27] zul: did you get the email i sent last week about qemu-kvm test pkgs? [03:49] sarnold: can dig lagcraft.co.uk +trace and let me know if there is any thing wrong with the name resolution? [03:50] or any one if sarnold is not arround i think i am having problems with my dns but dont know where :/ === gurps_ is now known as gurps [04:29] stgraber: so if we have an lxc-instance upstart job for each container, i wonder if it should respawn. [04:29] the problem with not doing so is that if it gets shut down and the admin does 'lxc-start' to restart it, then upstart won't now shut it down at reboot === rmk` is now known as rmk [04:33] hallyn: well, respawn would be a problem if you do a halt/poweroff from within the container [04:35] stgraber: lol [04:35] good point [04:35] then i think i'm ready to push this [04:36] no wait [04:39] hallyn: lxc-start will still be sent a TERM then KILL at shutdown [04:39] sigh if we go this route i could see a bug coming in a few months saying lxc-start should check if the container is autostart, and use upstart if so [04:40] SpamapS: seems to be getting the sigPWR i asked for it to be sent [04:40] well, I really see it as a distro thing, we'll now manage our auto started containers with upstart [04:41] stgraber: it's jsut that hand-starting the lxc-instance job is a lot more annoying typing [04:41] if someone wants to have an upstart managed container post-boot, they'll need to do: initctl start lxc-container NAME=container-name [04:41] SpamapS: but you're saying at the last moment. right [04:42] stgraber: right (or just 'start lxc-container NAME=name' :) [04:42] stgraber: perhaps have a 'lxc-start-auto' job that starts all the instances marked for auto [04:42] yeah i think we'll need something like that [04:42] well, don't we have that anyway (our main lxc job)? [04:42] * SpamapS prefers assertions to instructions [04:42] stgraber: but we dont' want to shut down the still-runnign ones [04:43] hmm, right [04:43] right, you want to say "make sure all the ones that should be started, are. [04:43] but, this sounds solvable later? may as well push the lxc-instance job now? [04:44] yeah, that sounds like minor implementation details for a limited set of our users, so we can deal with it later [04:44] thanks guys. good night. [04:45] ideally we'd have some way to allow for someone to "start" the lxc job again without having to stop it first, in which case it'd just start whatever containers aren't started yet. But I can't think of a way to do that with upstart. [04:45] and last I checked we don't have a reload/restart stanza we could use for that (restart does stop+start and reload does SIGHUP on the tracked pid) [04:46] status lxc-instance NAME=container-name | grep -q '^start' || start lxc-instance NAME=container-name [04:49] SpamapS: that's not the question. the question is where do we put such a command [04:49] such that " lxc" would have upstart trigger them [04:50] where someaction==restart and someaction==reload wont' work [04:52] hallyn: that seems like the main script portion of a task that is specifically meant to start auto containers [04:53] SpamapS: and how do we then trigger it? [04:53] oh [04:53] right, another separate task to do that. that's what i was saying above... not ideal, but... [04:54] I don't see why its not ideal [04:54] such is the way of upstart.. to every task, its own conf [04:54] very zen [04:55] oh hey! while hyou're here, [04:55] IIUC we should avoid using /etc/default/foo, [04:55] and instead put envs in /etc/init/foo.conf right? [04:55] how does dpkg handle conflicts then on pkg upgrades? [04:56] its a conffile [04:56] if you want it to be generated (and thus not subject to conffile handling) you can do that by generating it in postinst [04:58] no, ideally what i probably want is customizatoins to be done in .overrides, but ok, i just wanted to make sure i wasn't goin gto make a faux-pas in qemu [04:59] thanks, good night again :) === NomadJim_ is now known as NomadJim [05:18] Anyone got a recommendation for a JOBODs card? (just a bunch of disks). So I can sw raid 10 sata hdds on a 6port motherboard. I'm lost in the choices. & I want to make sure it's ubuntu compatible. [05:23] lvmer: LONG ago, I had great luck with 3ware for the cheap side [05:24] lvmer: its a "real" RAID and you can get them in the highest and lowest bus speeds. [05:29] ok I'll check them out. I'm going to be doing sw = software raid, so I'd like to avoid hardware pci-e raid costs if possible. or limit them. [05:30] I know I can 90% of the time - turn them off and run in JBOD mode but why get it in the first place xD. But idk if I can do jbod on a port multiplier card. It's all just so complicated & this tiny part is the only thing holding me back from putting up my 3rd (much larger) file server [05:35] spamaps: does something like this make sense?: http://www.amazon.com/SAS9211-8I-8PORT-Int-Sata-Pcie/dp/B002RL8I7M/ref=pd_cp_pc_0 [05:48] well I'm sorry I'm so lost. But google / forums aren't getting me anywhere. So I've got to ask - what would be a problem with something like this: http://www.newegg.com/Product/Product.aspx?Item=N82E16816124028 it's like $200 cheaper? Would sw raid still be ok? [05:49] I have no plans to boot off of it. I keep the OS separate from the RAID file server. If that helps. [05:49] how to change an ip address using terminal? [05:51] need halp [05:55] noliverh: you can use $ sudo ifconfig ethX IP === n0ts_off is now known as n0ts [05:56] trying to setup apache2 with ssl on amazon EC2 [05:57] chilicuil: how about their dns and gateways? [05:57] how can i put them in? [05:58] noliverh: for the gw, you can use: $ sudo route add default gw IP # and for the dns, you can edit $ sudo vim /etc/resolv.conf [05:59] (98)Address already in use: make_sock: could not bind to address 0.0.0.0:443 [05:59] noliverh: that's just for testing purposes, you may want to edit $ sudo vim /etc/network/interfaces # to set permanent values [06:01] chilicuil: how can see my gateway using ifconfig? [06:01] i mean... i cant see my gateway [06:01] im installing a ubuntu server now [06:02] i want it to be my PDC [06:03] noliverh: you can see it by running $ route [06:03] anyone? [06:03] wow! thanks [06:04] do i have to put a dns in my server since this already its ip address? [06:05] noliverh: I'd suggest to do it [06:07] what is vim? [06:07] is this an editor? [06:07] noliverh: yep, it's an editor, use $ nano instead if you had never heard about vim [06:08] chilicuil: im already in vim... how to get out of it? [06:08] :D [06:08] ok... how can i save it? [06:08] noliverh: type ":wq" [06:09] noliverh: if you're already editing u'll need to press before trying to enter ":wq" [06:09] chilicuil: how to save in nano? [06:09] :D [06:10] Ctrl-s , but I'm not sure, you should be able to see the shortcuts at the bottom of your screen [06:10] there's no save here... [06:10] ill try [06:11] why is it error writing: permission denied [06:12] noliverh, are you trying to write to a super-user-protected area? [06:12] say, /etc/ ? [06:13] yup [06:13] aws... i forgot the sudo [06:13] did you open it with sudo? [06:13] :P [06:13] okies [06:13] :D [06:14] then, how to show my dns? [06:15] chilicuil: what you mean by permanent values? [06:15] i want this to be my pdc [06:15] so i set up a dns here? [06:16] what should be done next? [06:19] noliverh: the values u set up with $ ifconfig, route and editing /etc/resolv wont survive a reboot, if you want permanent values, so your server get the same ip/gw/dns on every restart, you can edit /etc/network/interfaces https://help.ubuntu.com/12.10/serverguide/network-configuration.html [06:20] noliverh: once you confirm the values u've just tested works for you, I'd suggest write them in /etc/network/interfaces [06:22] chilicuil: okies... matsalams... :D [06:23] noliverh: (=, gl, I'm leaving === chilicuil is now known as chilicuil_away [06:40] Danawar: http://paste.ubuntu.com/1393796/ === jMCg_ is now known as jMCg === smb` is now known as smb === Akendo_ is now known as Akendo === n0ts is now known as n0ts_off === yofel_ is now known as yofel === deegee__ is now known as drussell [09:30] jamespage: morning [09:31] morning caribou [09:31] jamespage: I got a noob question for you regarding walinuxagent SRU [09:31] jamespage: how should I go about backporting the fix ? Here is my theory : [09:32] jamespage: get the diff from raring-proposed, try to apply the diff to both P & Q [09:32] jamespage: or manually modify if needed [09:32] jamespage: does that make sense ? [09:33] caribou, personally I would branch the version in raring; and then backport it in full to precise and quantal [09:33] using ~ubuntu12.04.1 and ~ubuntu12.10.1 respectively to deal with upgrades [09:34] caribou, we need the new version of walinuxagent in those two releases anyway for HE [09:34] (Hardware Enablement) [09:34] jamespage: interesting, I was instructed otherwise on another SRU I did with smoser & utlemming [09:34] jamespage: but this was for grub2 which can't simply be retrofitted to earlier versions [09:34] caribou, let me check what I did for the last SRU for this package [09:36] jamespage: I'm still familiarizing myself to doing the actual SRU backport effort [09:38] jamespage: Precise already has a -proposed branch with one of your merge : [09:39] jamespage: Backport for enablement of Windows Azure IaaS platform in [09:39] Ubuntu 12.04 (LP: #1014864) === Akendo is now known as Akendo_ [09:39] caribou: yeah - I was just looking for that [09:39] we did a straight backport to 12.04 of the version from 12.10 [09:40] using the ~12.04.1 suffix in the changelog [09:41] zul: cinder branch fixed up in distro now [09:43] jamespage: so doing a straight backport from raring to 12.04 & 12.10 would sync up all releases [09:43] caribou, yep [09:45] jamespage: ok, precise-updates & precise-proposed branches are identical [09:47] caribou, good [09:52] jamespage: bear with me as I'm not sure if I get the bzr cuisine right : [09:53] jamespage: Am I right to consider that it is only a matter of doing a "bzr branch" of the raring-proposed tree, fix the changelog, test and if ok, do a merge proposal ? [09:53] * jamespage thinks [09:54] caribou, I'd branch the precise branch and then merge in the raring branch so that you keep the changelog history for precise [09:54] and then add a new backport entry for the SRU [09:55] jamespage: ok, I'll try that === TuxLof is now known as TuxLof`work === ibiris|afk is now known as ibiris === mcclurmc_away is now known as mcclurmc === zeppo_ is now known as zeppo === chilicuil is now known as chilicuil_away === n0ts_off is now known as n0ts === alaing is now known as alaingMBCS [12:31] adam_g, roaksoax: so I spent some time this morning reviewing the haproxy/openstack API server approach [12:31] it feels really clunky with an external haproxy instance and very complex/fragile [12:31] I think I prefer adam's original suggestion that we should embed haproxy into each of the charms [12:32] using keepalived + haproxy in that configuration will provide a nice lightweight solution to this problem with closer configuration to the parent service [12:32]