=== zz_natorious is now known as natorious === natorious is now known as zz_natorious [09:28] Merged stackforge/cloud-init: Change the API of wait_any_url to return a tuple of url and response https://review.openstack.org/191100 [09:56] claudiupopa: Review of https://review.openstack.org/#/c/191137/ is blocking the other two things you reviewed from landing. :) [09:56] (Apologies if you were already getting to it) [09:56] ah, missed it. [09:56] thanks [09:57] This means that the documentation for windows specific modules will not be created by autodoc? [09:59] claudiupopa: It does; but that's unavoidable unless we make the things we want to document importable on the doc-building host. [09:59] can it be done manually at least? [10:00] claudiupopa: I don't know, let me have a play. [10:01] claudiupopa: (If you could approve that, I'll submit any progress I make as a separate change) [10:02] ok, no need to block this. We'll write them by hand. [10:02] smoser: I've updated https://code.launchpad.net/~daniel-thewatkins/cloud-init/lp1464253/+merge/261849 with full parameter names and no use of shell=True. [10:04] Merged stackforge/cloud-init: Don't try to generate autodoc for Windows modules. https://review.openstack.org/191137 [10:05] Merged stackforge/cloud-init: Fail the doc build if we have any warnings. https://review.openstack.org/191149 [10:06] Merged stackforge/cloud-init: Add doc8 checks to docs tox target. https://review.openstack.org/191162 [10:07] claudiupopa: Thanks for the reviews. :) [10:07] No problem. ;-) [10:25] Daniel Watkins proposed stackforge/cloud-init: Clean up stale auto-generated autodoc files. https://review.openstack.org/191139 [11:04] Daniel Watkins proposed stackforge/cloud-init: Clean up stale auto-generated autodoc files. https://review.openstack.org/191139 [11:36] claudiupopa: http://paste.ubuntu.com/11718962/ lets us autodoc more of the Windows code, but does make the Windows code uglier as a result. [11:37] that's quite ugly. [11:38] Well, you chose an ugly platform. ;) [11:38] If we bike-shed for long enough, Windows will probably be UNIX-based, right? :p [11:39] claudiupopa: Yeah, it's totally ugly; but I don't think we'll do any better with autodoc. [11:39] I'll have a look and see if there are any tools that can do autodoc-style stuff, but with static analysis. [11:39] Those should work OK. [11:39] That would be cool. [11:39] We could write one if not. [11:40] claudiupopa: Yeah, though Sphinx plugins are fairly painful to write. [11:41] can't we hook autodoc instead? [11:42] claudiupopa: Having a quick look through autodoc's source, it's pretty tied up with having a Python object that it can examine. [11:43] Preprocessing the files to remove everything except for docstrings _might_ work... [11:44] But I suspect that having some scripting which does this will be a lot less painful than trying to integrate it in to the Sphinx run itself. [11:45] Yeah, I mean it doesn't need to be an autodoc extension, as long as it builds .rst files from .py, all should be good. [11:47] Yeah. [11:48] Let's see if Joshua doesn't already have something like this. ;-) [11:48] :D [11:49] harlowja: ^^ ^_^ [12:20] claudiupopa: harlowja: Thinking about it, I'm not sure that our use case (for reporting) really needs a library. [12:20] claudiupopa: harlowja: If there was a library that would give us _handlers_ for free, that would be more interesting. [12:21] But the pub-sub sort of a model is overkill; we'll only have one publisher, and a constrained number of subscribers (which will all be configured in the same place at the the same time, from configuration). [12:21] If we want to use pub-sub more generally within cloud-init, then it would obviously make sense for the reporting framework to use that. [12:21] But I don't think it justifies it in and of itself. [12:22] yeah, makes sense, it's a little overkill for our use case. [14:05] Odd_Bloke: harlowja: do we actually need something heavy weight as this https://github.com/stackforge/cloud-init/blob/master/cloudinit/version.py#L9? [14:05] in certain conditions, it looks for a git repo and tries to run a git command. [14:09] claudiupopa: Are you hitting problems with it? [14:10] yeah. [14:10] http://paste.openstack.org/show/294140/ [14:10] Not related per se (I think), but I can't stop noticing the git calls. [14:11] It's executed from a local editable installation. [14:12] Yeah, it doesn't look directly related. [14:12] But that isn't ideal. [14:14] On the other hand, we shouldn't hit that in a proper installation. [14:14] harlowja: Thoughts? === zz_natorious is now known as natorious === natorious is now known as zz_natorious === zz_natorious is now known as natorious [17:30] sup [17:31] just got in [17:31] claudiupopa i there another claudiu in your work place? [17:31] * harlowja gets confused when i see another claudiu [17:31] lol [17:32] Odd_Bloke u no want my pub/sub thingy, lol [17:33] as far as https://github.com/stackforge/cloud-init/blob/master/cloudinit/version.py#L9 up to u guys [17:34] i'm fine with the second way, or just a static string (that someone remembers to update) [17:35] harlowja, there are 2 claudiu in cloudbase [17:35] i knew its! [17:35] and it is very confusing [17:35] :-P [17:35] i had a conversation at ODS with someone who said 'hi, i'm claudio from cloudbase' [17:35] ya, i had to double-check on https://review.openstack.org/#/c/191168/ [17:35] and i said... um i don't think you are [17:35] lol [17:36] did u ask for ID then? [17:36] lol [17:37] i sent him a gpg encrypted email with a password that was required before i would continue the conversation [17:37] my social skills are l33t [17:37] lol [17:37] ha [17:37] def === natorious is now known as zz_natorious [18:18] l [19:04] hi! [19:05] I'm playing around with cloud-init. I don't really need a "cloud", I'm using the nocloud data source with QEMU, which works great as I can attach files as disk devices on the VM [19:05] is there a way to do something similar with conainers? [19:07] jaksi, yeah. [19:08] http://ubuntu-smoser.blogspot.com/2013/08/lxc-with-fast-cloning-via-overlayfs-and.html [19:08] essentially you just write the same files that go on that disk into /var/lib/cloud/seed/nocloud-net/d [19:09] smoser : quick question, I cannot seem to get linux vms to get the random generated password by nova injected by cloud-init [19:09] ah, great, thanks a lot! [19:09] smoser : curl 169.254.169.254/openstack/latest/password is just plain empty [19:11] smoser : I've been trying alot different configs but I guess that until the nova metadata is empty it will never succeed, any idea what is wrong or a direction to verify look for ? [19:11] ByPasS, cloud-init doesn't pay attention to that. [19:11] use ssh keys. [19:15] ByPasS, but that said, i really haven't ever looked at it. [19:15] i can verify that on my cloud [19:15] 'password' is present in curl http://169.254.169.254/openstack/latest/ [19:15] but empty at curl http://169.254.169.254/openstack/latest/password [19:21] smoser : oh let me check [19:24] yea thats basicly my issue, I do agree to use ssh keys but it seems the project manager wants password injection for some obsucre reasons [19:28] echo "root:$(curl http://169.254.169.254/openstack/latest/password)" | sudo chpasswd [19:28] ^ ByPasS something to that affect might work. [19:28] ByPasS, i wonder, is that a one time read thing ? [19:30] smoser : runcmd on first boot you mean ? [19:31] smoser : could be I will test it but as far as I know, with cloudbase-init for windows the metadata is still present upon reboots etc [19:31] yeah. [19:31] then, yeah, its porbably not read-once [19:31] (which you might want it to be :) [19:32] alexpilotti, probably knows how it works. i'd have to look at code. [19:47] smoser : thatd be great I will still try it just in case :) who knows === zz_natorious is now known as natorious