cloud-init removes SharedConfig.XML on new instances of Azure

[Ben Howard]

The Azure DS is removing /var/lib/waagent/SharedConfig.xml on first boot, causing walinuxagent to fail.

Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Attempting to remove /var/lib/waagent/SharedConfig.xml
Apr 14 11:46:24 ubuntu [CLOUDINIT] DataSourceAzure.py[INFO]: removed stale file(s) in '/var/lib/waagent': ['/var/lib/waagent/SharedConfig.xml']
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Writing to /var/lib/waagent/ovf-env.xml - wb: [384] 1632 bytes
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Running command hostname with allowed return codes [0] (shell=False, capture=True)
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Running command ['hostname', 'daily-vivid-0414-1ce494fec4'] with allowed return codes [0] (shell=False, capture=True)
Apr 14 11:46:24 ubuntu [CLOUDINIT] DataSourceAzure.py[DEBUG]: pubhname: publishing hostname [phostname=ubuntu hostname=daily-vivid-0414-1ce494fec4 policy=True interface=eth0]
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Reading from /proc/uptime (quiet=False)
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Read 11 bytes from /proc/uptime
Apr 14 11:46:24 ubuntu [CLOUDINIT] util.py[DEBUG]: Running command ['sh', '-xc', 'i=$interface; x=0; ifdown $i || x=$?; ifup $i || x=$?; exit $x'] with allowed return codes [0] (shell=False, capture=False)
Apr 14 11:46:26 ubuntu [CLOUDINIT] util.py[DEBUG]: Reading from /proc/uptime (quiet=False)
Apr 14 11:46:26 ubuntu [CLOUDINIT] util.py[DEBUG]: Read 11 bytes from /proc/uptime
Apr 14 11:46:26 ubuntu [CLOUDINIT] util.py[DEBUG]: publishing hostname took 2.119 seconds (2.12)
Apr 14 11:46:26 ubuntu [CLOUDINIT] DataSourceAzure.py[DEBUG]: invoking agent: ['service', 'walinuxagent', 'start']
Apr 14 11:46:26 ubuntu [CLOUDINIT] util.py[DEBUG]: Running command ['service', 'walinuxagent', 'start'] with allowed return codes [0] (shell=False, capture=True)
Apr 14 11:47:27 ubuntu [CLOUDINIT] util.py[DEBUG]: waiting for files took 60.959 seconds
Apr 14 11:47:27 ubuntu [CLOUDINIT] DataSourceAzure.py[WARNING]: Did not find files, but going on: {'/var/lib/waagent/SharedConfig.xml'}
Apr 14 11:47:27 ubuntu [CLOUDINIT] DataSourceAzure.py[WARNING]: SharedConfig.xml missing, using static instance-id

[Scott Moser (smoser)]

so, to give an idea on what happens here.
cloud-init boots, finds its on azure, and then runs:
 service walinuxagent start
it expects that that will produce some files in some amount of time (60 seconds).
if it does not find those, it complains and goes on.

/var/log/waagent.log has similar messages about missing sharedconfig.xml

2015/04/14 16:13:44 Finished processing ExtensionsConfig.xml
2015/04/14 16:13:44 ERROR:Traceback (most recent call last):
2015/04/14 16:13:44 ERROR: File "/usr/sbin/waagent", line 4600, in AddSystemInfo
2015/04/14 16:13:44 ERROR: sharedConfig = xml.dom.minidom.parse("/var/lib/waagent/SharedConfig.xml").childNodes[0]
2015/04/14 16:13:44 ERROR: File "/usr/lib/python2.7/xml/dom/minidom.py", line 1918, in parse
2015/04/14 16:13:44 ERROR: return expatbuilder.parse(file)
2015/04/14 16:13:44 ERROR: File "/usr/lib/python2.7/xml/dom/expatbuilder.py", line 922, in parse
2015/04/14 16:13:44 ERROR: fp = open(file, 'rb')
2015/04/14 16:13:44 ERROR:IOError: [Errno 2] No such file or directory: '/var/lib/waagent/SharedConfig.xml'
2015/04/14 16:13:44 ERROR:

[Scott Moser (smoser)]

i'm pretty sure this is a systemd loop caused by cloud-init and wallinuxagent

 * cloud-init's cloud-init.service [1] is issuing 'service walinuxagent start'
 * walinux-agent's systemd job:
   | +After=network.target
   | +Wants=sshd.service sshd-keygen.service
 * Also, be aware of bug 1440180 that fixed cloud-init's (mis)use of network.target.

--
[1] http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/systemd/cloud-init.service

[Scott Moser (smoser)]

bug 1425376 has more information than bug 1440180 about network.target, but for some reason slangasek closed 1440180 in the cloud-init changelog wiht his fix.

[Launchpad Janitor (janitor)]

This bug was fixed in the package walinuxagent - 2.0.12-0ubuntu2

---------------
walinuxagent (2.0.12-0ubuntu2) vivid; urgency=medium

  * Fixed systemd unit file which caused SharedConfig.xml to be deleted by
    Cloud-init (LP: #1444086).
 -- Ben Howard <email address hidden> Wed, 15 Apr 2015 10:59:38 -0600