=== mjohnson15_2 is now known as mjohnson15 | ||
xibalba | know of a way to firewall a shell user on aa system? say i give you a shell on my box / 192.168.1.51. i dont want you accessing my home network at 192.168.0.0/24 but i , from my reza account, want ot be able to hit my home network | 00:23 |
---|---|---|
=== freeflying is now known as freeflying_away | ||
Patrickdk | I can't remember if that is still supported with iptables or not | 00:24 |
xibalba | i wonder if the ol' tcpwrappers could do it | 00:24 |
Patrickdk | only if every program they used was build with tcpwrappers support | 00:25 |
sarnold | xibalba: no, tcpwrappers is a voluntary deal, it wouldn't be a very good firewall | 00:25 |
xibalba | ok | 00:25 |
sarnold | xibalba: iptables --uid-owner is probably what you;re looking for, I've never tried it though. | 00:25 |
xibalba | thanks i'll take a gander | 00:25 |
=== thomi_ is now known as thomi | ||
=== freeflying_away is now known as freeflying | ||
=== peter is now known as Guest12214 | ||
=== marrusl is now known as marrusl_afk | ||
=== marrusl_afk is now known as marrusl | ||
=== gfrog_afk is now known as gfrog | ||
=== freeflying is now known as freeflying_away | ||
makara | i want to keep revisions of conf files | 09:15 |
makara | i've read of RCS | 09:15 |
makara | apparently the syntax is cumbersome | 09:15 |
makara | just looking for something simple and local | 09:15 |
jamespage | zul, urgh - the webob update that just trickled through busted nova | 09:28 |
=== gfrog is now known as gfrog_meeting | ||
makara | how can I measure the bandwidth savings that my squid proxy cache has provided? | 09:52 |
makara | and is it possible for to get squid hooked up to avahi (like squid-deb-proxy) so I don't have to reconfigure my browser settings every time I come to the office? | 09:54 |
makara | (i can't configure it as a transparent proxy) | 09:57 |
yolanda | jamespage, i'm looking at the heat failure, missing a build-dep | 09:57 |
jamespage | yolanda, yep - wanna propose a branch for that? | 09:58 |
yolanda | sure | 09:58 |
jamespage | yolanda, it might not build after that - webob just bumped version which is causing test failures as requirements < 1.3 in most projects | 10:01 |
jamespage | but please do try :-) | 10:01 |
yolanda | ok, trying now | 10:01 |
yolanda | jamespage, it built! | 10:04 |
jamespage | yolanda, if you wanted a bit of SRU practice: bug 1250654 looks like a good bit sized one | 10:05 |
uvirtbot | Launchpad bug 1250654 in python-keystoneclient "memcache key length error " [High,Fix released] https://launchpad.net/bugs/1250654 | 10:05 |
yolanda | ok | 10:07 |
yolanda | jamespage https://code.launchpad.net/~yolanda.robla/heat/icehouse_add_lockfile/+merge/198699 | 10:08 |
jamespage | yolanda, niggle but please leave the changelog entry as UNRELEASED - and add your entries using dch -t | 10:08 |
yolanda | jamespage, i added entry with dch -t, is there something wrong? | 10:09 |
jamespage | yolanda, did you do a dch -r as well? | 10:09 |
yolanda | no | 10:09 |
jamespage | how did it change from UNRELEASED to trusty ? | 10:10 |
yolanda | i manually edited that | 10:10 |
yolanda | jamespage, updated MP | 10:11 |
jamespage | yolanda, so we should not be updating from UNRELEASED until someone actually does an upload to the archive | 10:12 |
jamespage | then they do dch -r which updates the target and the timestamp on the changelog and do the upload | 10:13 |
yolanda | good to know | 10:13 |
yolanda | i normally tend to set the latest distribution name | 10:13 |
yolanda | jamespage, i have a question about SRU process explained here. https://wiki.ubuntu.com/StableReleaseUpdates | 10:43 |
yolanda | it says i need to fill an [Impact] , [TestCase] and [Regression Potential] comments. Where do i have to do it, in the same bug? | 10:44 |
jamespage | yolanda, yes - in the description | 11:17 |
yolanda | jamespage, and then i grab the bugfix from trusty and i update the package for saucy? | 11:18 |
jamespage | yolanda, no | 11:18 |
jamespage | you will need to cherry-pick the fix from upstream github and make it work for the version on saucy. | 11:18 |
jamespage | yolanda, the bugfix in trusty is a new upstream release - you can't sru that | 11:18 |
yolanda | jamespage, i meant that, sorry | 11:19 |
yolanda | just look at that commit | 11:19 |
jamespage | yolanda, the bug reporter is quite friendly - maybe you could ask him for a more specific test case configuration | 11:27 |
jamespage | I suspect he'd also verify a fix for you as well once its in the proposed pockets | 11:27 |
yolanda | jamespage, and SRU comments should be filed in same bug description, or create some other bug? i saw that in Launchpad there are specific SRU bugs | 11:28 |
jamespage | yolanda, same bug - why would you raise another one? | 11:29 |
jamespage | the SRU fixes that bug after all | 11:29 |
yolanda | jamespage, just asking because I saw other independent bugs in Launchpad, I want to be sure that i'm doing the right thing :) | 11:29 |
jamespage | yolanda, if we have minor release exceptions, then yes we would probably raise a specific SRU bug | 11:30 |
rbasak | smoser: are we treating bug 1250390 as resolved? I know I haven't verified it, but it should be fixed now, right? | 11:44 |
uvirtbot | Launchpad bug 1250390 in maas "No published Saucy armhf ephemeral images" [High,Triaged] https://launchpad.net/bugs/1250390 | 11:44 |
rbasak | jamespage: ^^ | 11:44 |
jamespage | rbasak, ta | 11:44 |
zul | jamespage: of course it did | 12:31 |
jamespage | zul, \o/ | 12:31 |
swaT30 | hey guys, sorry if you had already answered this, but any ETA on 2013.1.4 to be pushed into updates? I see that it has been in proposed for a while | 12:32 |
zul | jamespage: ill look when i get in | 12:32 |
=== PriceChild is now known as PriceChilde | ||
=== PriceChilde is now known as Pricey | ||
swaT30 | @jamespage any idea? | 12:40 |
jamespage | swaT30, sorry - about what (guess I must have missed something) | 12:41 |
swaT30 | any ETA on 2013.1.4 hitting updates in the UCA? | 12:41 |
swaT30 | :) | 12:41 |
jamespage | swaT30, hmm - yes that has been sat in proposed for a while | 12:49 |
jamespage | lemme kickoff the automated testing - if its all ok I'll promote today | 12:50 |
swaT30 | would be great to have it :) | 12:50 |
swaT30 | thanks ! | 12:50 |
swaT30 | jamespage: I'll keep an eye on http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/grizzly_versions.html | 12:51 |
swaT30 | I assume that's as good a spot as any? | 12:51 |
=== gary_poster|away is now known as gary_poster | ||
makara | i've logged in as user JIM and I need to run a command as GAX | 13:11 |
makara | what 2 do? | 13:11 |
makara | i don't know GAX's password | 13:12 |
jamespage | swaT30, yes - thats good | 13:12 |
jamespage | swaT30, there is also a notifications ML | 13:12 |
smoser | rbasak, its not fix-released for sure. | 13:30 |
smoser | as in no 'released' images. | 13:31 |
makara | any idea how to execute a script on system start as given user?' | 13:37 |
rbasak | makara: there are various ways. You could add something to /etc/rc.local that uses su or sudo. Or add an upstart job to /etc/init/ that uses setuid/setgid. Or even an @reboot job to cron. | 13:40 |
makara | rbasak, and if I put "su - teamcity runAll.sh" in /etc/rc.local then why wouldn't get hung up on asking me for a password? | 13:44 |
makara | like it would in a shell | 13:44 |
rbasak | makara: rc.local runs as root, which should be able to su to any user without a password. Check it works by using "sudo -i" to get a root shell, and then test the su command you propose there. | 13:45 |
jamespage | swaT30, tested OK - should be out on the archive in the next 2 hrs | 13:48 |
paws | hello, can anyone tell me whats a good Cloud storage software, that i can host on my own server ? | 14:04 |
zul | jamespage: whats patch-branches in nova icehouse packaging? | 14:05 |
jamespage | zul, uh? | 14:06 |
zul | jamespage: yeah... | 14:06 |
zul | jamespage: anyways its gone | 14:06 |
jamespage | zul, I don't see that | 14:06 |
zul | jamespage: its an empty file in the debian directory | 14:07 |
jamespage | not a clue | 14:07 |
Felipe_C | Hi all, Has anyone heard of or actually tried those Paralella Boards to create a proof of concept mini cloud cluster? | 14:07 |
jamespage | zul, its been there along time | 14:08 |
zul | jamespage: yeah well if anything breaks then ill blame myself | 14:08 |
zul | jamespage: https://code.launchpad.net/~zulcss/nova/webob/+merge/198751 | 14:10 |
zul | jamespage: https://review.openstack.org/#/c/61742/ as well | 14:13 |
zul | yolanda: i just merged your heat fixed | 14:22 |
yolanda | thx | 14:22 |
swaT30 | jamespage: thanks! on the ML, got the notifications | 14:24 |
zul | jamespage: trove is in openstack-ci now | 14:30 |
zul | jamespage: https://code.launchpad.net/~zulcss/cinder/webob/+merge/198752 | 14:41 |
bittin | Hired an ubuntu server from azure :) | 14:54 |
jamespage | zul, how do you always manage to remove dep headers from patches? | 14:56 |
zul | jamespage: i dunno its an art i think :) | 14:56 |
jamespage | zul, I'm putting in a conditional wait in nova-compute upstart that if libvirt-bin is installed, it will wait for it to reach running before starting | 14:56 |
jamespage | should close out a couple of race bugs I've seen | 14:57 |
jamespage | I *think* we need todo the same with neutron-ovs-cleanup | 14:57 |
zul | jamespage: sounds good | 14:57 |
zul | jamespage: yay trove needs a unpackaged dep | 15:09 |
=== freeflying_away is now known as freeflying | ||
jamespage | zul, feedback on your MP | 15:16 |
jamespage | sorry - git review | 15:16 |
makara | exit | 15:17 |
zul | jamespage: i saw ill fix it up | 15:17 |
jamespage | zul, I don't think we should unbound it - but < 1.4 is OK imho | 15:17 |
zul | jamespage: agreed | 15:18 |
jamespage | zul, adam_g: https://code.launchpad.net/~james-page/nova/compute-plus-others/+merge/198763 | 15:23 |
zul | jamespage: couple of things, while you are at it | 15:27 |
jamespage | zul, sure | 15:27 |
zul | jamespage: i think the openssh-client can go away in the build-depends-indep | 15:27 |
zul | jamespage: curl can go away as well | 15:28 |
jamespage | zul, what was that for? | 15:28 |
zul | jamespage: im not sure i think it was a diablo thing | 15:28 |
zul | jamespage: nova-uml should go away | 15:28 |
jamespage | zul, actually openssh-client - still a runtime I think | 15:29 |
jamespage | migrations | 15:29 |
zul | jamespage: ah yes | 15:29 |
jamespage | live migrations and cold migrations | 15:29 |
zul | ssh key-gen as well i guess | 15:29 |
zul | jamespage: but yeah git rid of uml as well :) | 15:30 |
=== freeflying is now known as freeflying_away | ||
jamespage | zul, we don't like uml do we? | 15:30 |
zul | jamespage: makes the package less complicated ;) | 15:31 |
yolanda | jamespage, i added a fix to rabbitmq charm https://code.launchpad.net/~yolanda.robla/charms/precise/rabbitmq-server/ha/+merge/198768 | 15:49 |
d1n0 | I am having problems with maas. I have imported the boot images, but when I try and pxe boot a node.. it doesn't see the pxe template | 15:49 |
zul | jamespage: https://code.launchpad.net/~zulcss/swift/1.11.0/+merge/198769 | 15:52 |
yolanda | jamespage, currently rabbitmq is relying on HA to create clusters. That is different on active-active. Should we set a config var in rabbit to specify if we want active-active or active-passive? | 15:55 |
jamespage | yolanda, can we not determine which mode using relations? | 15:56 |
jamespage | its only HA if it gets related to hacluster | 15:56 |
jamespage | zul, can't drop openssh-client | 15:57 |
zul | jamespage: ok cool | 15:57 |
yolanda | jamespage, i see, and act differently and set different vars depending on it. For example in active-active i need to send all the rabbitmq hosts | 15:58 |
jamespage | yolanda, yes - that's right | 15:59 |
yolanda | i could set a var in rabbitmq to tell glance for example, if rabbit uses active-active or active-passive, and glance reacts and create proper config | 15:59 |
yolanda | makes sense? | 15:59 |
=== matsubara_ is now known as matsubara | ||
jamespage | yolanda, I think it does - the current HA approach passed a VIP to signal its in HA mode right? | 16:01 |
yolanda | yes | 16:01 |
jamespage | yolanda, do we need rabbitmq >= 3.0.0 todo the active/active stuff? I think we might be not 100% sure | 16:02 |
yolanda | jamespage, i don't think so, openstack doc says it's tested with 2.7.1 | 16:04 |
jamespage | yolanda, OK - the setting of HA policy in your change is scoped to >= 3.0.x | 16:04 |
yolanda | jamespage, that's only needed for version >= 3.0.x, just followed openstack guide | 16:05 |
jamespage | yolanda, ah - OK _ good-oh | 16:05 |
yolanda | jamespage, i am starting with glance, to make glance react to rabbit_hosts needed to grab version from precise-havana | 16:06 |
yolanda | doesn't seem to work with older versions | 16:06 |
zul | Daviey: when you get a chance can you look at pytyhon-designateclient for me (its in source new, dependency needed for openstack-trove) | 16:20 |
=== malakhi_ is now known as malakhi | ||
=== Jikai is now known as Jikan | ||
hallyn_ | rbasak: trusty is not supported by uvt-simplesreams-sync? | 16:45 |
zul | jamespage/roaksoax: https://code.launchpad.net/~zulcss/swift/1.11.0/+merge/198769 | 16:48 |
rbasak | hallyn_: it comes from a different stream. uvt-simplestreams-libvirt sync --source http://cloud-images.ubuntu.com/daily release=trusty arch=amd64 or similar. | 16:50 |
hallyn_ | rbasak: any reason not to have uvt figure that out? | 16:50 |
hallyn_ | (lxc-create does :) | 16:50 |
hallyn_ | (you can see i'm big on not typing more than i need to :) | 16:51 |
rbasak | hallyn_: it would be nice, yes. It's a bit of a tricky thing to do with the simplestreams API, and I think it should probably be pushed back there in some way. lxc-create will have the same issue when it's simplestreamed. | 16:51 |
rbasak | hallyn_: I'd like to tell the simplestreams API "here are your streams, here are my filters, go". But I'm not sure about how multiple streams will interact that way. | 16:53 |
rbasak | smoser: ^^ | 16:53 |
rbasak | hallyn_: btw, I uploaded the latest PPA earlier | 16:55 |
hallyn_ | cool, thanks. | 16:55 |
jamespage | zul, OK - I think that nova mp is ready to go - ditto the neutron one | 16:59 |
jamespage | zul, looking at that now | 16:59 |
jamespage | why do we need to disable the tests? | 16:59 |
zul | jamespage: they are broken for now | 16:59 |
zul | and we need to do a mir for nose | 17:00 |
smoser | rbasak, i dont hink i follow | 17:03 |
jamespage | zul, blimey - swift needs some love | 17:04 |
jamespage | zul, which bit of nose? | 17:05 |
zul | jamespage: the functional test bits | 17:05 |
jamespage | zul, yeah - I mean't which dependency? | 17:06 |
zul | jamespage: hold on lemme double check | 17:06 |
smoser | rbasak, com.ubuntu.cloud.daily:server:trusty:amd64 is very explicitly different than com.ubuntu.cloud:server:trusty:amd64 | 17:06 |
smoser | that is by design. | 17:06 |
smoser | i woudln't suggest munging that all together. | 17:06 |
zul | jamespage: libjs-jquery-hotkey, jquery-goodies, libjs-jquery-isonscreen, and javascript-common | 17:07 |
jamespage | zul, is that for docs? | 17:07 |
zul | jamespage: i believe so | 17:07 |
rbasak | smoser: yes, they are. But uvtool should munge them together, as an exception, for the development release. | 17:09 |
rbasak | smoser: this makes them analogous to what people expect by setting their sources.list. | 17:09 |
smoser | i dont think so. | 17:10 |
smoser | there is no released trusty. | 17:10 |
smoser | if you say trusty released, it can't get you such a thing. | 17:10 |
smoser | if you say trusty and daily, it can | 17:10 |
rbasak | I want the latest available of everything. For stable releases, I want the stable image. For the development release, I want the development release. | 17:10 |
rbasak | Sorry. For the development release, I want the latest development image (ie. a daily). | 17:11 |
smoser | why? | 17:11 |
rbasak | uvt-just-give-me-it. | 17:11 |
smoser | that daily might not work at all. | 17:11 |
rbasak | That's fine. That's what using the development release means. | 17:11 |
rbasak | The CLI should provide sensible defaults. If I specify a release, then the sensible default is the stable release image for stable releases, and the daily image for the development release. | 17:12 |
jamespage | zul, sorry - don't get it | 17:12 |
rbasak | It is fine for this to be uvtool-specific and Ubuntu-specific. I'm just saying that I want some way to explain this to the simplestreams API. | 17:13 |
zul | jamespage: http://people.canonical.com/~ubuntu-archive/component-mismatches-proposed.svg | 17:13 |
rbasak | Or else we need to figure out how uvtool should handle the interaction. | 17:13 |
jamespage | zul, there is actually no point in running them - SKIPPING FUNCTIONAL TESTS DUE TO NO CONFIG | 17:13 |
jamespage | https://launchpadlibrarian.net/154000942/buildlog_ubuntu-saucy-i386.swift_1.10.0-0ubuntu1_UPLOADING.txt.gz | 17:13 |
smoser | i personally dont really think that "daily image for development release, but stable imgaes for all others" is likely whta a developer wants. | 17:13 |
smoser | more likely i'd want dailies of everything. | 17:14 |
zul | jamespage: agreed | 17:14 |
jamespage | zul, I don't see why that would require disabling tests | 17:14 |
smoser | cstack does handles this. sanely. | 17:14 |
jamespage | zul, also the swift-doc package is empty | 17:14 |
zul | jamespage: ok ill have a look at getting this fixed then | 17:14 |
rbasak | hallyn_: ^^ interested in this conversation? | 17:15 |
jamespage | zul, thats a regression - 1.10.0 has it | 17:15 |
jamespage | populated | 17:15 |
hallyn_ | no? :) | 17:15 |
* hallyn_ looks | 17:15 | |
rbasak | smoser: I think you assume that the CLI somehow knows that this is a developer speaking. It doesn't know that. | 17:15 |
rbasak | smoser: given that it doesn't know, I think it's reasonable to default to the released image for stable releases. | 17:15 |
rbasak | smoser: but given that the user just asked for "trusty", and that there is no released image for that, and it's known to be a development release, I think it's reasonable to provide a daily in that case. | 17:16 |
hallyn_ | 17:13 < smoser> i personally dont really think that "daily image for development release, but stable imgaes for all others" is likely whta a developer wants. | 17:16 |
hallyn_ | i disagree. that's exactly what i want | 17:16 |
smoser | its not what i want | 17:16 |
smoser | and i'm a developer | 17:16 |
hallyn_ | it's myworkaround for you cloud image folks being too lazy to give mea stable devel image | 17:17 |
smoser | hallyn_, its not cloud image folks. | 17:17 |
smoser | there is no stable devel image of any ubuntu | 17:17 |
smoser | when there is (next thursday) | 17:17 |
smoser | you'll have something in the stable stream of trusty | 17:17 |
hallyn_ | that's alpha1? | 17:17 |
smoser | it will be correctly labelled 'alpha-1' rather than release | 17:17 |
rbasak | I'm happy to add a --developer option that changes sync's default to use the daily stream. | 17:18 |
hallyn_ | in the end (ignoring smoser's being pedantic) i don't care who builts the smarts in - uvtool or simplestreams. | 17:18 |
rbasak | However, ifi the user doesn't do that, I still want trusty to be available by default, on request. And I don't think it's reasonable to default to daily images across the board. | 17:18 |
hallyn_ | but i wanted to test something in canonical-kernel-team/ppa just now for trusty, and wanted the 'best' trusty cloud image i could find | 17:19 |
hallyn_ | <shrug> i'm with rbasak. | 17:19 |
smoser | hallyn_, that would be the dailies ;) | 17:19 |
rbasak | So I think "best" is released for stable releases, and daily for development releases, despite any alpha-1. | 17:19 |
hallyn_ | smoser: i'm just a user of uvtool. i don't care where it comes from, i just want "the most stable" image that exists | 17:20 |
smoser | no you dont. | 17:20 |
rbasak | hallyn_: would you want alpha-1 or a daily for trusty, if both existed and the daily were newer? | 17:20 |
smoser | so its fine with me if uvtool has logic like: | 17:20 |
hallyn_ | rbasak: i guess i'd want daily | 17:20 |
smoser | if release_requestsed == ubuntu_development_release(): change_stream_to_daily() | 17:21 |
rbasak | smoser: trouble is that right now uvtool just passes through the release=... filter without understanding it. | 17:21 |
rbasak | I would prefer to not special case that understanding. | 17:21 |
smoser | you want a special case understanding | 17:22 |
smoser | but you dont want to special case it | 17:22 |
smoser | you want someone else to special case that | 17:22 |
rbasak | I want a general way of handling this case. | 17:22 |
smoser | but the someone else here thinks that when you say "released" you get "released". and when you say "daily" you get "daily" | 17:22 |
rbasak | I'm fine with telling the simplestreams API about this understanding. I want to do it in an API call. | 17:22 |
smoser | i dont think so. daily is explicitly and by design not released. | 17:23 |
rbasak | If the user specifies released or daily, then that should override default behaviour and do what you asked for. | 17:23 |
smoser | you can munge those two things together however you choose. | 17:23 |
smoser | as i said, cstack does this "right" for me. | 17:23 |
rbasak | If the user doesn't specify released or daily, released/dailiness should be based on the status of the release requested. | 17:23 |
rbasak | And of course the release requested should also have a default, but that wouldn't change this behaviour. | 17:24 |
smoser | rbasak, then you can implement that. | 17:24 |
rbasak | smoser: you want me to parse simplestreams filter rules myself? | 17:24 |
rbasak | smoser: I want the simplestreams API to do the parsing, but for me to give it the rule. | 17:24 |
smoser | you want 2 completely different things. available only from 2 completely different, completely unrelated sources. | 17:26 |
rbasak | Yet the release on one corresponds to the release on another. | 17:27 |
rbasak | They are not completely unrelated. | 17:27 |
smoser | happenstance | 17:27 |
rbasak | No. I have a specific reason for wanting that relation. Therefore, provided that my reason is reasonable, they are related. | 17:27 |
smoser | scott james remnant and scott moser are not related. | 17:28 |
rbasak | They relate at the CLI. 1) "Trusty!"..."ok, you want daily." 2) "Saucy!"..."ok, you want released." | 17:28 |
smoser | even if it would be convenient for robie basak if they were. | 17:29 |
rbasak | The names precise, quantal, raring, saucy, trusty are related. The daily images are related to that set. So are the released images. | 17:29 |
smoser | just like my example. | 17:29 |
smoser | same name | 17:29 |
smoser | different thing | 17:29 |
smoser | i can't fix your chromebook. | 17:29 |
=== yofel_ is now known as yofel | ||
rbasak | The daily ends up being a released. | 17:29 |
smoser | happenstance | 17:30 |
smoser | honeslyt. | 17:30 |
rbasak | Nothing like your example. | 17:30 |
smoser | you're welcome to buildin the knowledge that 'com.ubuntu.cloud.daily' is somehow related to 'com.ubuntu.cloud' | 17:30 |
rbasak | I think a problem here is that you do not accept the original use case that makes me want this. | 17:30 |
smoser | but you have to go to 2 completely different end points of data to find those things. | 17:30 |
smoser | i think this is easily solvable. | 17:31 |
smoser | and i've told you I'm ok for you to solve it. | 17:31 |
rbasak | Well, I can parse the simplestreams filters. So we can duplicate the code, and lock in the simplestreams filter query language. Great. | 17:31 |
smoser | you can create filters. and mirror 2 streams to 2 different locations. | 17:32 |
smoser | i dont see how that is duplicating anything. | 17:32 |
rbasak | I need to do that automatically. Without expecting the user to set it up for me. And I need to handle the transition of a release, where I want to switch from use of a daily image to a released image. | 17:33 |
rbasak | Oh look - they're related! | 17:33 |
smoser | they're not related. | 17:34 |
smoser | was that unclear ? | 17:34 |
rbasak | Where I was using a trusty daily image, I will after release want to use a trusty released image. Thus they're related. | 17:34 |
smoser | dont handle it. | 17:34 |
smoser | after trusty is released, get the released stream | 17:35 |
rbasak | The point of uvtool is to hide this complication, and just do what is reasonable. | 17:35 |
rbasak | Automatically. | 17:35 |
rbasak | The only way I can see of doing this is to either hardcode the assumption that the release key maps between daily and released, or to publish the mapping somehow. | 17:36 |
rbasak | (oh look - they're related) | 17:36 |
rbasak | I guess we're at an impasse, which basically just means that the pain will continue. Sorry hallyn_. | 17:36 |
smoser | cstack does this right. | 17:37 |
rbasak | Except that you've defined what you think is right, and we disagree. | 17:37 |
rbasak | Somebody who says "precise", without any further information, can be inferred to mean "stable, please". And someone who says "trusty" presumably means "development, please". | 17:38 |
smoser | and then based on some even that occurs at some point in the future, your interpretation of what they mean changes. | 17:39 |
smoser | which is odd. | 17:39 |
smoser | and unexpected. | 17:39 |
smoser | just because the calendar says 2014-04-17. | 17:39 |
hallyn_ | that event being a release? | 17:39 |
rbasak | It matches how Debian releases work, so is not at all unexpected. | 17:39 |
hallyn_ | that's not expected | 17:39 |
hallyn_ | uh, not unexpected | 17:39 |
smoser | it is unexpected to the user. | 17:39 |
rbasak | It is not, since that's what the release names are defined to mean. | 17:39 |
rbasak | That's how we all use the term. | 17:39 |
hallyn_ | the user to whomthat is unexpected would nto specify a release at all, and run latest LTS | 17:39 |
rbasak | You've invented your own system of daily and released for cloud images (with good reason), but that doesn't mean that these concepts apply to release names. | 17:40 |
smoser | it is not good design for behavior to change | 17:40 |
smoser | based on events non-controllable by the user | 17:41 |
smoser | "I did nothing, yet it broke" | 17:41 |
smoser | where "it broke" is "it behaved differently" | 17:41 |
rbasak | "I had a bug in Trusty, which I fixed, but why isn't showing up in my uvtool test? Wait, why is it using alpha-1?" | 17:41 |
rbasak | The development release is *defined* to behave differently. | 17:42 |
rbasak | It is not good design to surprise the user. | 17:42 |
hallyn_ | would it help if 'uvt-s-s query' showed trusty entries in red for "danger"? :) | 17:42 |
smoser | rbasak, i dont know how to solve your problem. with the interface you have. | 18:04 |
zul | jamespage: https://code.launchpad.net/~zulcss/trove/trove-cleanups/+merge/198803 | 18:05 |
smoser | since you'd have to look at 'release=' which could be 'release~([tT][rR].*)' | 18:05 |
smoser | or something else that would match that. | 18:05 |
smoser | perhaps you should, being an 'ubuntu' tool, explicitly know certain things about releases, and treat them as releases. rather than just as arbitrary keys. | 18:07 |
smoser | ie: uvt-kvm trusty | 18:07 |
smoser | rather than | 18:07 |
smoser | uvt-kvm release=trusty | 18:07 |
smoser | that is the bit that cstack knows. | 18:08 |
d1n0 | where does a maas node get its sources.list from? | 19:27 |
d1n0 | its a pxe boot node, and im getting a 403 error when trying the packages | 19:27 |
smoser | d1n0, it can be specified by maas. | 19:28 |
smoser | but it should default to archive.ubuntu.com | 19:28 |
d1n0 | smoser; why would i be getting a 403 though, thats what confuses me | 19:29 |
smoser | agreed. | 19:29 |
smoser | that is confusing. | 19:30 |
smoser | are you able to pastebin it ? | 19:30 |
smoser | it == the console log | 19:30 |
d1n0 | Unfortunately not, because after it gets the 403.. the pxe boot fails and it shuts down. | 19:31 |
d1n0 | logfile: /var/log/squid-deb-proxy/access.log shows the 403 error | 19:32 |
sarnold | d1n0: note that s3 systems return '403' when they mean '404'. | 19:34 |
sarnold | it's confusing. | 19:34 |
d1n0 | hmm, now you confuse me even more.. lol | 19:35 |
d1n0 | smoser; how can I change what the maas sends over to it? I want it to use a local mirror | 19:36 |
smoser | d1n0, it depends on what version of maas. | 19:36 |
d1n0 | whatever is in 13.10 | 19:37 |
smoser | but i think in cloud-tools version or saucy, it is in the web ui. | 19:37 |
smoser | i think | 19:37 |
d1n0 | i thought so too, but it looks like they took it out between 12.04 and 13.10 | 19:37 |
smoser | sarnold, well, its not likely s3 is involved here. | 19:37 |
smoser | and to be fair, thats general good practice for security. | 19:38 |
smoser | ie, you dont tell the attacker "username not found" you tell them "username/password invalid". | 19:39 |
d1n0 | editing the node only allows editing the hostname and mac addresses (along with changing the default release) | 19:39 |
sarnold | smoser: yeah, s3 seemed unlikely, but an s3 workalike crossed my mind. I figured I'd throw it out ;) hehe | 19:39 |
smoser | i do agree it sucks for s3 | 19:39 |
smoser | mirrors explicitly. | 19:39 |
smoser | d1n0, it wouldnt be on the node | 19:39 |
sarnold | yeah; it's a reasonable enough position in general. hehe. :) | 19:39 |
smoser | it would be somwhere else. | 19:39 |
smoser | in d'setting' or something | 19:39 |
smoser | settings. | 19:39 |
d1n0 | changed, now for another attempt at a pxe boot | 19:43 |
d1n0 | now back to my vague invalid mac address error | 19:56 |
d1n0 | nothing like getting one or more of your mac addresses is invalid when trying to commission a node :s | 20:14 |
zul | adam_g: ping can you +1 https://code.launchpad.net/~zulcss/swift/1.11.0/+merge/198769 please | 20:20 |
adam_g | zul, done | 20:20 |
zul | adam_g: thanks | 20:21 |
hxm | hi | 20:38 |
hxm | i have a server with many domains and I could want to let the users receive and send emails from their domains, as in example facebook does | 20:39 |
hxm | forwarding emails, is that possible without using postfix and all those tools that are hard to set up properly for not get as spam | 20:39 |
Demosthenex | so ubuntu LTS 10.04. after booting for FIVE MINUTES no one, not root, no accounts can login. timeout nonstop. ssh with key is ok, it seems to be a password authentication chain issue. ideas on debugging? | 20:42 |
Demosthenex | ah, i had to remove smb crap from pam common-auth. | 20:48 |
mgw | What is the purpose of the packages at http://archive.ubuntu.com/ubuntu/pool/? | 21:29 |
mgw | To answer my question… it seems to host the actual packages pointed to from the various Packages.gz | 21:50 |
mgw | There's a package in pool that works fine in precise (python-gnupg), but it's not in Precise's Packages.gz…. other than downloading the .deb, how could that be installed via apt? | 21:51 |
smoser | hallyn_, its not perfect, but it bevaves the way you happen to think it should | 22:15 |
smoser | https://code.launchpad.net/~smoser/uvtool/sm-features00/+merge/198838 | 22:15 |
smoser | rbasak, ^ | 22:15 |
* smoser gone | 22:15 | |
hallyn_ | smoser: don't be angry | 22:16 |
hallyn_ | :) | 22:16 |
=== jerrad is now known as Underbyte | ||
=== gary_poster is now known as gary_poster|away | ||
hallyn_ | smoser: looks perfect | 22:49 |
=== mibofra- is now known as mibofra |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!