[01:48] smoser https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/+merge/293957 should be up and running again [01:48] seems to pass unit tests (allthough a few new unittests fail) [01:48] aka one that runs 'Command: ['dpkg', '--print-architecture']' [01:48] or ''lsb_release', '-cs']' [01:48] stuff that's not being mocked afaik [01:49] https://gist.github.com/harlowja/d0a8d56ca460d0a7adbc99f1db3c8fe6 [09:17] I am creating an OpenStack instance with an ephemeral disk, and cloud-init creates a filesystem and mounts it as /mnt - is there a way to stop cloud-init from doing this? I want to use the ephemeral disk to create a drbd device. [12:23] ccard_, I just had to do something similar, check this out https://castro.io/2015/01/24/preparing-ec2-instance-store-with-cloud-init.html [12:23] I had to override the mounts and also remove an existing entry in fstab for the mount [12:25] GivenToCode: thanks, that's helpful. Ideally I'd like there to be no filesystem at all created on the ephemeral disk, since I want to create a drbd partition on it. [12:31] ccard_, can't help you there, amazon puts a filesystem on most of its ephemeral disks automatically. Though you can replace it with whatever you like via fs_setup or bootcmd, etc [13:06] harlowja, yeah, cpaelzer is working on those. [13:14] smoser: ? [13:15] smoser: out of my log since I sometimes disconnect over night (like tonight) due to vpn [13:15] smoser: was that a reference to the bug you opened with the unittests failing in some environments? [13:16] harlowja said: [13:16] smoser https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/+merge/293957 should be up and running again. seems to pass unit tests (allthough a few new unittests fail). aka one that runs 'Command: ['dpkg', '--print-architecture']' or ''lsb_release', '-cs']'. stuff that's not being mocked afaik https://gist.github.com/harlowja/d0a8d56ca460d0a7adbc99f1db3c8fe6 [13:16] thanks smoser [13:16] yeah the dpdk print architecture fix is in my last MP [13:16] dpkg [13:17] I can't type it anymore :-/ [13:21] :) [13:21] link ? [13:21] I thought it falls into your inbox, gimme a sec ... [13:22] smoser: https://code.launchpad.net/~paelzer/cloud-init/bug-1589174-fix-tests-in-adt-env/+merge/296643 [13:22] smoser: you might argue about the (unrelated) bddeb change, since ther emight have been a way I just haven't seen [13:22] smoser: but other than that most changes should be straight forward [13:23] smoser: tested in Ubuntu, ppa, sbuild and a CentOS container [13:44] cpaelzer, it does fall into my inbox with the slew of 300 other things :) [13:45] smoser: hehe [13:45] smoser: it takes most of my first 30 minutes each day to sort that out [13:46] smoser: don't tell me you have not all read in your inbox :-P [13:58] has anyone run in to issues in upstart with a service with start on stopped cloud-init running before cloud-init is finished? [14:41] GivenToCode, not that i'm aware of, and that surely should not happen,. [14:42] harlowja, what would you do if you wanted an Enum but did not want to depend on python2 package to provide python 3's builtin Enum ? [14:43] cpaelzer, you have a conflict in merge at url above [14:46] smoser: ok, I'm gonna refresh later [14:49] I should always before MP'ing next time ... [14:50] smoser, I found this: http://www.madorn.com/cloud-init-stages.html#.V1bVB7orJMM which helped me see I need start on stopped cloud-final [14:52] right. :) [14:52] GivenToCode, i thought maybe you didn't mean cloud-init but cloud-inti-final [16:05] 1227 wasn't out yet when I branched this morning [16:06] I merged and uploaded again [16:06] thou I'm not sure all in the merge worked correctly [16:06] the conflict was minimal [16:06] smoser: please let me know if there still is a conflict [16:07] did a pull into a trunk dir, found the conflciting rev 1227, did a merge into my dir, found the conflict, resolved and committed the merge [16:07] yet it seemded to have an empty diff [16:07] * cpaelzer lacks one more piece of bzr magic [17:04] ok backs [17:04] sooo smoser Emum that u don't want to use the pkg from [17:04] i can make a crappyEnum that will work [17:04] lol [17:05] in nearly the same manner [17:05] smoser do u want that ? [17:10] harlowja, my query was for stuff for curtin.. [17:10] there is http://stackoverflow.com/questions/36932/how-can-i-represent-an-enum-in-python [17:11] which kind of has a crappy enum for py2, but i just wanted to know what you woudl ot there. [17:11] prob make enum (maybe less crappy) that works for u [17:12] and try to make it so that it works with py3 (using native stuff) and py2 (using your stuff) [17:33] smoser i'll make some [17:33] thing [17:33] the enum seems pretty nice. do people use that ? [17:33] i do [17:33] lol [17:33] i am all the people [17:33] lol [17:33] my expectation was to use for basic constant like things [17:37] smoser ya, one sec, whipping up something for u [17:37] lol [17:37] looks like we used this in curtin (and cloud-init/reporting) before [17:37] pastebinithttp://paste.ubuntu.com/17095446/ [17:38] http://paste.ubuntu.com/17095446/ [17:38] hmmm [17:38] ya, that's pretty ghetto :-P [17:38] mine will be a little better, ha [17:38] i suspected as much. [17:51] https://gist.github.com/harlowja/f60d91128a1fcd99d30d27715d5a9a30 smoser [17:51] have fun [17:51] ha [17:51] should use the native enum, and a close-enough equivalent on py2 [17:57] harlowja, if six.PY3 ? [17:58] if enum is None [17:58] ? [17:58] ya, do that, i was just doing that approach cause wanted to make sure my mac tested the other path [17:58] and the venv i run in has enum lib installed, ha [17:58] magicalChicken, ^ [17:59] that is for enum stuff. harlowja thinks our namespace is for pathetic lusers [17:59] lol [17:59] it at least mirrors more closely the py3 stuff [18:00] so the basics should work across both [18:00] smoser: Yeah that amkes sense haha. Using __getattr__ is kinda not good [18:01] I'll get that pulled into the storage-config branch real, quick, thanks harlowja [18:03] np [18:03] do is enum not None though [18:03] as smoser pointed out [18:03] it probably is missing some things, but should work for basic things [18:05] soooo that gets me enough brownie points to get https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor merged right [18:05] lol [18:05] smoser in that i'm just for that merge not making the net/ folder any less copy/pasteable [18:05] think change big enough already [18:06] can do that later or something [18:06] or ummm, neterator [18:06] lol [18:08] harlowja, it does in deed. [18:08] oh goodie [18:09] does it also get enough for https://gist.github.com/harlowja/d63a36de0b405d83be9bd3222a5454a7 after that to also merge [18:09] lol [18:09] actually i gotta change that due to the nova bug [18:09] (with naming things dumb) [18:11] but, bb, gotta do oslo ptl stuff for a little