[12:44] <smoser> jgarr, seedfrom was busted recently.
[12:44] <smoser> https://bugs.launchpad.net/cloud-init/+bug/1455233
[12:46] <smoser> anotheral, cna you give me example of what you're p utting in there?
[13:24] <openstackgerrit> Claudiu Popa proposed stackforge/cloud-init: Add the data source base classes and the HTTP OpenStack implementation  https://review.openstack.org/188327
[13:48] <openstackgerrit> Claudiu Popa proposed stackforge/cloud-init: Add the data source base classes and the HTTP OpenStack implementation  https://review.openstack.org/188327
=== Odd_Bloke changed the topic of #cloud-init to: cloud-init || 2.0 reviews: https://review.openstack.org/#/q/project:stackforge/cloud-init,n,z
=== zz_natorious is now known as natorious
=== natorious is now known as zz_natorious
[15:27] <Odd_Bloke> smoser: I'm pretty much wholesale lifting your OpenStack talk abstract. :)
[15:28] <smoser> imitation is the greatest form of flattery
[15:28] <Odd_Bloke> /nick Odd_moser
=== zz_natorious is now known as natorious
[16:05] <ByPasS> Odd_Bloke : has discussed with smoser curl http://169.254.169.254/openstack/latest/password is empty on linux VMs
[16:05] <ByPasS> Odd_bloke : on windows VM there is informations...
[16:05] <ByPasS> as type
[16:05] <ByPasS> as typo ! :D
[16:08] <Odd_Bloke> ByPasS: On the OpenStack I'm looking at, there is an admin_pass key in the config drive metadata.
[16:08] <Odd_Bloke> But not in the HTTP metadata.
[16:08] <Odd_Bloke> (You don't want it in the HTTP metadata; you're effectively granting anyone with access to your server root if it's available from there)
[16:11] <ByPasS> Odd_Bloke : the http metadata in windows is not really the password itself since there is mix with the password and the pub key is used to hash/encrypt it and in order to do nova get-password u need the private key sent..
[16:11] <ByPasS> Odd_Bloke : I will need to have a look at config drive metadata I never used it so far (yet)
[16:55] <ByPasS> Odd_Bloke : thanks using metada drive and this first boot command : mount /dev/sr0 /mnt; echo "ubuntu:"`cat meta_data.json  | python -mjson.tool | grep admin | awk -F'"' '{print $4}'` | chpasswd; umount /mnt it now sets the passwd properly :)
[16:55] <Odd_Bloke> ByPasS: :)
[16:56] <ByPasS> well forgot the full path for the cat but basicly u get the point
[16:56] <Odd_Bloke> ByPasS: We'd be happy to accept a patch in to cloud-init for this, if you wanted to try that. :)
[16:59] <ByPasS> Odd_Bloke : thatd be interesting, I will brag to my project manager about the successful passwd injection and will look into it if the company gimme some time or at least I can do it in spare time
[17:01] <harlowja> Odd_Bloke http://tinyurl.com/on6c32o try that sucker out
[17:01]  * harlowja nicer dashboard for gerrit
[17:01] <harlowja> * created from  https://review.openstack.org/192770
[17:03] <Odd_Bloke> OH YOU FANCY
[17:03] <harlowja> :-P
=== harlowja changed the topic of #cloud-init to: cloud-init || 2.0 reviews: http://tinyurl.com/on6c32o
[17:03] <harlowja> :)
=== Odd_Bloke changed the topic of #cloud-init to: cloud-init || cloud-init 2.0 reviews: http://tinyurl.com/on6c32o
[17:03]  * harlowja to fast, ha
[17:04] <Odd_Bloke> Needs more cowbell^Wcloud-init.
[17:04] <harlowja> lol
[17:45] <anotheral> smoser: actually, it looks like the cloud-init part is working, but that set_hostname: does not.  So i can use a shell script as a workaround.
[20:38] <harmw> harlowja: I'm not seeing very mucho reviews :p
[21:09] <harlowja> harmw make some more code ;)
[21:09] <harmw> that hurts :p
[21:10] <harlowja> ;)
[21:11]  * harlowja can't do as much as i want, got all these other code-bases to help with too
[21:11]  * harmw knows that problem :>
[21:12] <harlowja> and i'm not #1 so need to make that happen ( @ http://stackalytics.com/?metric=commits )
[21:12] <harlowja> *see bottom loll
[21:12] <harlowja> :-P
[21:12] <harlowja> taking my #1 place, bums
[21:12] <harlowja> lol
=== natorious is now known as zz_natorious
[22:47] <jgarr> do runcmd entries run as root? or as one of the defined users?
[22:48]  * jgarr is curious if that's a problem with his subscription-manager command that doesn't appear to run
[22:59] <ByPasS> jgarr : yes it does run as root / UID 0
[23:00] <ByPasS> jgarr : at least from my experience I can use mount with runcmd which requires root
[23:00] <jgarr> ByPasS: k thanks, then I have no idea why subscription-manager doesn't work as a runcmd :-/
[23:00] <jgarr> my only thought is it runs too early in the boot process and network isn't available or something
[23:02] <ByPasS> are you using list method or string ?
[23:03] <ByPasS> as far as I know bootcmd is executed really early in the boot process
[23:04] <ByPasS> runcmd set in a string should be executed on runlevel 3 while passing the info in a string will be written to a file and ran by sh (according to the doc)
[23:05] <ByPasS> I'd add a while loop checking for network connectivity right before the subscription manager cmd to validate your suspicion
[23:06] <jgarr> ByPasS: I see the command is in /var/lib/cloud/instance/runcmd but the machine isn't subscribed
[23:07] <jgarr> I also try catching sterr/stout via `subscription-manager register ... >> /root/cloudinit.log 2>&1` but it never shows up in the file
[23:08] <jgarr> I didn't think to try a loop. I'll give it a shot. thanks