[16:33] <smoser> hey all
[16:33] <smoser> i'm hoping to mark 0.7.3 today.
[16:45] <pedroalvarez> cool
[16:53] <pedroalvarez> smoser: I don't know why I cannot make cloud-init receive data from OpenStack, I have a log this time. http://paste.ubuntu.com/6205664/
[16:59] <smoser> pedroalvarez, what are you expecting to find ?
[16:59] <smoser> configdrive or ec2 metadata service ?
[17:01] <pedroalvarez> i have installed ubuntu in openstack, and the cloud-config simply worked
[17:01] <pedroalvarez> I'm using the same configuration in other operating system, and I don't know why doesn't work
[17:02] <smoser> ah.
[17:02] <smoser> is ubuntu using cloud-config or ec2 
[17:02] <smoser> er.r..
[17:02] <smoser> config-drive or ec2 i meant.
[17:02] <pedroalvarez> how can I know? where in the log?
[17:04] <pedroalvarez> smoser: forget it for today, I have to go home. I'll ask you again tomorrow
[17:05] <smoser> cloud-init output should say.
[17:06] <smoser> but basically either you've configured openstack to offer config-drive (or requested it at launch) or to offer the ec2 metadata service.
[17:06] <smoser> if you don't know, then you're prbably doing the MD service.
[17:06] <smoser> but i have less guessses as to what is going wrong with your install for MD than I would for config-drive.
[17:07] <smoser> for config-drive, i'd suspect that enough drivers are not loaded to see the disk or a race condition in it becoming avialable.
[18:10] <smoser> harlowja, around ?
[18:10] <harlowja> yo
[18:10] <harlowja> smoser poke
[18:11] <smoser> you hav eanything really important that you'd like to have in to 0.7.3 ?
[18:11] <smoser> cause i'm about to push "go" on that.
[18:11] <harlowja> not afaik :)
[18:11] <smoser> k.
[18:11] <harlowja> that windows stuff?
[18:11] <harlowja> lol
[18:11] <smoser> yeah. windows. :)
[18:12] <harlowja> alexpilotti ;)
[20:56] <utlemming> smoser: problem with the latest uplaod to lp:cloud-init
[20:57] <smoser> utlemming, ?
[20:57] <utlemming> smoser: ephemeral0 is not mapping to the first partition
[20:57] <utlemming> smoser: on parted disks
[20:57] <utlemming> smoser: http://paste.ubuntu.com/6206712/
[20:58] <utlemming> smoser: http://paste.ubuntu.com/6206721/ 
[20:58] <smoser> i thought you had tested this.
[20:59] <utlemming> smoser: yeah, I did
[20:59] <utlemming> smoser: which is why I am confused
[20:59] <utlemming> smoser: because I took the branch you pinged me and tested against that
[21:02] <smoser> well i merged your branch
[21:06] <smoser> utlemming, please figure out what needs fixing. i just released 0.7.3 so that sucks.
[21:07] <utlemming> smoser: yeah, this blows hard
[21:07] <utlemming> smoser: I'll get this ASAP
[21:07] <smoser> i'll be back in in a couple hours.
[21:15] <utlemming> smoser: lp:~utlemming/cloud-init/lp1236594-bad_ephemeral0_mnt
[23:33] <smoser> utlemming, still there?
[23:33] <utlemming> smoser: yup
[23:33] <smoser> I did the "if not partition" by design.
[23:33] <smoser> because input to devnode_for_dev_part
[23:34] <smoser> should be a string or None
[23:34] <smoser> bool("0") == True
[23:34] <smoser> oh. but i guess "" is not true. that must be what was occuring there?
[23:34] <utlemming> right
[23:35] <smoser> no.
[23:36] <smoser> expand_dotted_devname("ephemeral0")
[23:36] <smoser> would return None
[23:36] <smoser> hm..
[23:36] <utlemming> the issue is "ephemeral0" is not selecting ephemeral0 or epehemeral0.1 because of that "if partition"
[23:39] <smoser> right.
[23:41] <smoser> utlemming, so i think the right place for that change is in sanitiz_devname.
[23:41] <smoser> i think . as its there that we're saying "ephemeral0" == ephemeral0.0 or ephemeral0.1
[23:43] <utlemming> hrm...
[23:44] <utlemming> but isn't that the same things as my patch? 
[23:46] <smoser> well mostly, except for then there is no way to call devnode_foor_dev_part with "do not expand this"
[23:46] <smoser> err.. with "do not allow part 0 or part 1"
[23:47] <utlemming> smoser: since this is a specific function to cc_mounts, I think adding a flag that controls that behavior seems sane
[23:48] <utlemming> otherwise we have two functions that only get halfway there
[23:48] <utlemming> i.e. def ...(..., literal=False)
[23:49] <utlemming> which controls the beavhior
[23:49] <utlemming> but again, this is one function that is called by one module
[23:49] <smoser> i think you're right
[23:52] <smoser> utlemming, you're changes are correct.
[23:52] <smoser> i'mpretty sure.
[23:58] <utlemming> smoser: I tested those changes on SmartOS, EC2 and Azure