=== natorious is now known as zz_natorious
=== zz_natorious is now known as natorious
[17:28] <mrkz> hello, Is there a [full] documentation of  all the cloud-init configuration options (/etc/cloud/cloud.cfg)?
[17:28] <mrkz> looking atm http://cloudinit.readthedocs.org/en/latest/index.html but looks incomplete to me
[18:42] <harmw> just out of python curiosity, how current is Flask for writing a simple REST service?
=== tmclaugh[work]_ is now known as tmclaugh[work]
[20:53] <robjo> Trying to mimic a config drive outside of OpenStack using qemu but am not having much luck, could use some help
[20:54] <robjo> I have the config in ISO format and am attaching the iso file to the VM with the -cdrom option for qemu
[20:55] <robjo> The device shows up as /dev/sr0 and thus I figured it should just work, but it does not appear to be picked up, any ideas?
=== tmclaugh[work]_ is now known as tmclaugh[work]
=== tmclaugh[work]_ is now known as tmclaugh[work]
[21:24] <mrkz> robjo: maybe try #qemu ?
[21:25] <robjo> mrkz: I am trying qemu
[21:25] <robjo> qemu-kvm -netdev tap,helper=/usr/lib/qemu-bridge-helper,id=hn0 -device rtl8139,netdev=hn0,mac=00:16:3e:7e:18:36 -m 2048  -cdrom /work/tmp/configDrive/configdrive.iso openSUSE-13.1-OS-guest.x86_64-0.0.6.raw
[21:26] <robjo> Inside the vm the ISO shows up as /dev/sr0, that based on my understanding of the code should be all there is to it
[21:26] <mrkz> I mean #qemu channel, this one looks a bit inactive :/
[21:26] <robjo> oh, sorry
[21:28] <mrkz> robjo: I just joined and ask, but no answer so far
[21:29] <robjo> it's more of a cloud-init than a qemu thing
[21:32] <mrkz> robjo: so, if you're trying to mimic an openstack instance boot, shouldn't you look into nova instead of cloud-init?
[21:33] <robjo> cloud-init initializes the VM and I want to test some changes to cloud init without having to set up OpenSTack
[21:36] <mrkz> I'm kinda new to this, but afaik nova boot up the VM and then cloud-init runs @ boot time inside the VM and then performs all you setup in config file
[21:36] <mrkz> dunno if I'm right tho
[21:37] <robjo> mrkz: Yes that's how it works
[21:38] <mrkz> robjo: so I quite don't understand your issue
[21:39] <mrkz> also, what I do to test is rm /var/lib/cloud and then re-run cloud-init init inside VM
[21:39] <robjo> When cloud-init runs it gets data from a so called data source, this can be a meta data server or can also be a so called ConfigDrive
[21:40] <robjo> ConfigDrive is one of the options offered with OpenSTack and is used for example when the environment cannot have a DHCP server
[21:40] <robjo> when using ConfigDrive cloud-init can write the network configuartion file and then bring upi the network
[21:41] <robjo> But this implies that the config drive is properly recognized by cloud-init
[21:42] <robjo> Theoretically having the config drive info in an ISO file that is attached to the VM and shows up a /dev/sr0 should be sufficient, however in my case cloud-init does not appear to find the config drive
[21:42] <robjo> thus I cannot test the writing of the network configuration file
[21:48] <dbuechler> Robjo: Unfortunately, I don't know anything about ConfigDrive, other than cloud-init seems to have a datasource for it.  What version of cloud-init are you running?
[21:48] <robjo> 0.7.6
[21:49] <dbuechler> What OS is the VM?
[21:51] <robjo> dbuechler. it really doesn't matter, but it is openSUSE 13.1
[21:51] <robjo> anyway I had another ISO and that one works as config drive :)
[21:51] <robjo> Looks like the second ISO I generated is not quite what cloud-init expected
[21:51] <robjo> I am all set
[21:52] <mrkz> robjo: have you looked @ /var/log/cloud-init.log in both VM instances?
[21:55] <dbuechler> Cool.
[21:56] <dbuechler> robjo: Glad to hear that you got it figured out.
[21:57] <dbuechler> I've got a problem of my own... cloud-init 0.7.7, built as an rpm for CentOS 7 using 'make rpm' runs on boot but bombs with:
[21:58] <dbuechler> cloud-init[1260]: import cloudinit.util as util
[21:58] <dbuechler> cloud-init[1260]: File "/usr/lib/python2.7/site-packages/cloudinit/util.py", line 49, in <module>
[21:58] <dbuechler> cloud-init[1260]: from six.moves.urllib import parse as urlparse
[21:58] <dbuechler> cloud-init[1260]: ImportError: No module named urllib
[21:59] <dbuechler> I have python-urllib3 1.5 installed, so I'm at a loss to explain it.
[21:59] <robjo> Couple of sources for the trouble would be that the dependencies in the rpm are not correct and when installing cloud init python-six is not pulled in
[22:00] <robjo> but six.moves.urllib is imported, not urllib
[22:00] <mrkz> dbuechler: but does $ python -c "import urllib; print urllib" work?
[22:01] <dbuechler> Good question!  Wait 1.
[22:01] <mrkz> no, as robjo said before, try this instead
[22:01] <mrkz> $ python -c "from six.moves.urllib import parse as urlparse; print urlparse"
[22:02] <dbuechler> Ok.  One sec and I'll find out.
[22:03] <dbuechler> Returns No module named urllib
[22:04] <dbuechler> I have python-six 1.3.0 installed.
[22:04] <dbuechler> ...from rpm - probably EPEL.
[22:05] <mrkz> so, I might be wrong, but looks like your six package lacks that
[22:06] <dbuechler> Errr, looks like a stock CentOS package.
[22:07] <dbuechler> Hrm.  Ok.  I'll check and see if EPEL has a six package.
[22:07] <robjo> I would concur, python-six-1.3.0 is most likely too old
[22:08] <robjo> I have 1.8.0 on my system and things work
[22:09] <dbuechler> Ok.  I'll take a look around and report back.  I appreciate the input.
[22:47] <dbuechler> Reporting back on my urllib issue - updating six in my CentOS 7 image with "pip install --upgrade six" DID solve the problem.  python-six that ships with CentOS 7 is too old.