[01:13] smoser: that patch totally fixed the problem (w/ growpart). [03:26] hm, is there a 'correct' way to shim/bootstrap the hostname function? i'd prefer to run my own little diy python script to set the hostname, but have it called through cloud-init [04:36] chucky_z: you could use the write_file module to write the script to disk, and then run it via a runcmd. Or you could just deliver the script as your only user-data, making sure it starts with #!/usr/bin/python. Or you could bundle it up with a cloud-cfg file in a MIME archive and pass that as your user-data. === rangerpbzzzz is now known as rangerpb [13:30] smoser: just FYI - really slow progress due to sick wife/kids and many external DPDK calls today - but the mock stuff worked fine [13:32] :-( [13:32] cpaelzer, yeah, and feel free to re-organize code for easier testing. [14:16] larsks, https://launchpad.net/cloud-utils/trunk/0.28 [14:17] smoser: thanks! [14:17] that was a long time coming. thank you for pushing [17:08] smoser rharper i'm also going to fix python2.6 in https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/+merge/293957 so expecct a few more changes [17:11] * harlowja will also add a few more test cases to the net rendering pipeline [17:44] since it appears that 0.7.x at some point lost 2.6 support :-/ [17:45] (and ya, i know some people still use rhel6) [18:12] what user does cloud-init run as? is that defined in the cloud.cfg? [18:13] root [18:13] hm, ok [18:15] i wrote a script to be run by runcmd, placed it into /etc/cloud/cloud.cfg.d/, but it doesn't appear to be doing anything (the script has no output, it's an alternate hostname setting) [18:28] ya, probably not the right place to put it [18:58] hm, couldn't get the runcmd to do anything, but placing the actual script under /var/lib/cloud/scripts/per-boot/ worked. :) [19:52] chucky_z, thats the right place to put it. [19:52] (/var/lib...) [19:52] if I understand correctly, a `runcmd` should go under /etc/cloud/cloud.cfg.d/ though? [19:56] no. [19:56] you could put a file like: [19:56] /etc/cloud/cloud.cfg.d/my-runcommands.cfg [19:56] with [19:56] runcmd: [19:57] - [sh, '-c', 'echo HI MOM'] [19:57] but then if you provide user-data that has runcmd in it it will override that. [20:40] smoser what can we do about not breaking py26 for 0.7.x [20:40] pyserial seems to be dead on 26 [20:40] and various other things in cloudinit [20:51] hm. [20:51] :-P [20:51] i fixing a bunch === rangerpb is now known as rangerpbzzzz [20:51] has it broken recently ? [20:51] ya [20:51] :-( [20:51] for a while i think, lol [20:51] anything using pyserial not work, lol [20:52] well, anything using newer versions of pyserial. [20:52] right? [20:52] surely pyserial from rhel-of-years-ago works [20:52] ya, i think 3.0+ [20:52] those seemed to remove py26 [20:55] would be nice to have even basic travis testing on cloudinit [20:56] maybe i'll setup something that can at leaset run nightly [20:59] we do haev a system that we can run some c-i on [20:59] can you see https://server-team-jenkins.canonical.com/ [20:59] i thin kthat is public [20:59] no connecting [20:59] :-P [20:59] hm.. [21:00] my route there goes all internal. [21:00] :-( [21:00] yeah, we should really get something setup [21:01] ya, i mean https://travis-ci.org/harlowja/cloud-init-1/ i can run if it will start up [21:01] using https://github.com/harlowja/cloud-init-1/blob/master/.travis.yml [21:01] which is the mirror of a mirror of bzr [21:01] lol [21:03] let's see if https://travis-ci.org/harlowja/cloud-init-1/builds/129542406 runs [21:03] at least i can force that hourly or something [21:03] ok. so i want to get this going and get to git and such. [21:03] ya [21:03] how about that :-P [21:03] i have a few bugs that i'm working on that need to be fixed in xenial [21:04] so they're high priority [21:04] k [21:04] fixhttps://travis-ci.org/harlowja/cloud-init-1/jobs/129542407 is the pyserial junk [21:04] * https://travis-ci.org/harlowja/cloud-init-1/jobs/129542407 [21:04] and then after i get that in i want to look at the nwere datasource model thing that i started. [21:04] after we fix 2.6 [21:04] lol [21:04] well, yeah [21:05] the big thing missing from datasources right now is the ability to find local information (such as a config drive) [21:05] but not be "processed" (and have '#include http://' from user-data rendered) before networking is up. [21:05] ie right now, for config drive you only get networking applied if your 'dsmode=local' [21:05] but if you are dsmode=local, then you cant use '#include http'... [21:06] woot, https://travis-ci.org/harlowja/cloud-init-1/builds/129542406 worked [21:06] so. that needs badly fixing. [21:06] except for 2.6 [21:06] pypy even works [21:06] lol [21:06] i like the little baby penguins there [21:06] they're so cute [21:06] :-p [21:06] * smoser has to go now [21:06] later. [21:06] peace [21:18] harlowja: pypy? nice; I was going to give that run, some folks on single-core arm have said "python is slow" and wanted to see what the speed-up might be using pypy [21:18] well it runs at least, at passes tests [21:19] https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/+merge/293957 that should fix all the py26 stuff also [21:19] mainly http://bazaar.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/revision/1229 [21:21] probably can be refined, but ya, at least it works now [21:21] where works == passes tests [21:21] right [21:21] that's what I was going to test first; step one done =) [21:21] :-P [21:22] heh, the pypy run was slower than the others [21:24] cool, wel there is my cloudinit changes for the day, back to openstack-land [21:28] ta