=== harlowja is now known as harlowja_away === praneshp_ is now known as praneshp [11:12] smoser: I'd hardly call that a fix :p [11:13] smoser: think you can merge those last bits regarding freebsd? [15:16] hiren_: ping [15:16] harlowja_away, http://paste.ubuntu.com/8260489/ [15:16] digital ocean has user-data now, and the ec2_utils.MetadataMaterializer is able to crawl it. [15:17] and utlemming they add vendor-data [15:17] :) [15:26] smoser: you worked out something about cirros? [15:26] harmw, either you or i broke htat script [15:26] and possiby i [15:27] but i thin kwe might be good now. [15:27] and i'm thinking about putting that into trunk. [15:27] ok, cool [15:27] well my branch certainly compiles :p [15:29] some openstack-qa people were asking for mtu [15:29] so i was going to try to get that in as 0.3.3 [15:29] nice [15:29] I think ill be able to testdrive something tonight [16:42] harmw: hey! === harlowja_away is now known as harlowja [17:06] smoser thats good right? (the digitial ocean stuff) [17:07] harlowja, yeah. really good. [17:07] woot [17:07] really nice that it worked. [17:07] who wrote that awesome metadata stuff [17:08] oh i know [17:08] hahaha [17:08] the one thing that was weird [17:08] _decode_leaf_blob [17:08] it returns a list of blob.splitlines() [17:08] which is kind of odd. [17:08] as in this case, using it got vendor-data and user-data as an array [17:09] ya, i wonder why i did that, ha [17:09] my guess is boto does that [17:09] so i do that [17:09] well, it was really for metadata only [17:09] which, in amazon, any leafs might be lists.. i'm not even sure if there are any leafs with newlines in them. [17:10] https://github.com/boto/boto/blob/2.15.0/boto/utils.py#L278 [17:10] sorry 281 there [17:10] right, but no ton leafs. [17:10] oh. hm.. [17:10] maybe it makes sense there. [17:10] i dont know. [17:10] want to change that? [17:11] but since 'user-data' and 'vendor-data' are relaly just blobs in that MD [17:11] we dont want them interpreted in this case [17:11] have to see i guess. [17:11] kk [17:11] maye just make it a method you can pass in [17:11] to render your own leafs [17:11] or you could subclass it too [17:12] who would be subclassing it? [17:12] well, in this case i would :) [17:12] for the digital-ocean reader [17:13] harmw, http://bazaar.launchpad.net/~cirros-dev/cirros/trunk/revision/317 [17:13] that gets your cirros-udchpc stuff functional in trunk. [17:13] smoser do u want to try the boto reader, and see what it returns just to double check? [17:13] functional in that it had some bugs that i' not sure were your fault or mine (quite possibly my changes). and also actually enables patching of buildroot's busybox so we use it. [17:14] sure [17:14] otherwise sure, i can make it more pluggble [17:15] wow. [17:15] i didn't realize the boto one was like it was [17:15] ? [17:15] doing the json.load [17:15] i thought you had added that. [17:15] never [17:15] ha [17:15] line 278 htere is kind of wierd [17:15] if it starts with { , then its *obviously* json :) [17:16] lol [17:17] of course! [17:17] thats why i put maybe_json_object [17:17] cause maybe it is, haha === shardy is now known as shardy_afk [17:25] smoser https://code.launchpad.net/~harlowja/cloud-init/customized-leaf-decoder/+merge/233565 [17:25] have fun :- [17:25] :-P [17:25] mixed leaf decoder [17:25] i like the name [17:27] :) [17:57] another thing we can do with the doc cleanup/make betterness is start to use https://pypi.python.org/pypi/doc8 (which i helped create/am the main person making); its getting sucked into most openstack tox.ini files for sanity checks on docs [18:06] smoser: I hope my buildservice picks up that cirros change ;) [18:07] you planning a new release ? [18:08] harmw, yeah, building something as 0.3.3~pre1 right now. [18:09] and then if that looks sane, which i think it does from my test builds, then we just call 0.3.3 [18:09] cool [18:10] dont forget th fbsd stuff ;) [18:11] yeah. i know. [18:13] harmw, looking [18:14] harmw, what about the config drive stuff that Rajun.. was doing [18:14] is that important ? [18:14] i think probably, for config drive at least [18:14] yea, but I'm afraid I haven't looked at it [18:15] so I don't realy have an opinion on it [18:15] let me build an update pkg for fbsd first, and then I'll think something of it :) [18:19] harmw, done. [18:23] ah nice [19:11] smoser: that stuff from Ragin looks fairly ok [19:12] old functionality is still in place, and I'm asuming he tested if it works on fbsd === zz_gondoi is now known as gondoi [20:19] smoser: morning! [20:24] alexpilotti, its not morning [20:24] but ok [20:24] :O) [20:25] smoser: what do you think about having -z in nc, on cirros? :-) [20:26] smoser: In tempest we need to test security groups, checking if a given port is open, cross-tenant [20:26] smoser: aka cirros -> cirros [20:26] is it something you can just turn on ? [20:27] smoser: we need to update nc in the cirros image [20:28] smoser: I mean, we can go with workarounds of course [20:28] sbut if you can easily update it, it would be nice [20:29] no. sorry. [20:29] was a question as to if it *can* be turned on in cirros [20:30] not really [20:35] do you know if turning on NC_110_COMPAT will do it ? [20:41] alexpilotti, http://git.busybox.net/busybox/tree/networking/nc.c [20:41] it seems like that might do it. [20:42] smoser: as in uploading the file and compiling it in place during the test? [20:42] well, no. i have to enable it. [20:42] but i'm wondering if that would be sufficient for you [20:43] ah ok perfect [20:43] yep, it looks very ok [20:44] in what cirros version do you think you could get it in? [21:01] well, hopefully i could just turn that on for 0.3.3 [21:01] which i was hoping would be "real soon now" [21:13] harmw, 0.3.3~pre1 is there now [21:13] i'll plan on trying a build with alexpilloti's request [21:13] and if that seems to not completely crash 'nc', then i guess we'll just call that 0.3.3 [21:18] smoser: tx! :-) === gondoi is now known as zz_gondoi === harlowja is now known as harlowja_away === harlowja_away is now known as harlowja