Due to a software bug in vSphere, it fails to handle ampersand in datacenter names.
The solution is to do what vSphere does (when browsing) and double-encode ampersands.
It is likely other characters need special treatment like this as well, haven't found any.
This change is related to 2 issues;
- The API does not return destroyed VMs if zone ID is passed for CS version < 4.5.2. Also see CLOUDSTACK-8578. This only affects domain and root admins.
- The instance name must be unique across all zones. If we pass the zone ID to find a VM, it will not be found if it is in a different zone but a deployment with the name would fail.
This fixes::
Traceback (most recent call last):
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080916.98-133068627776311/lxc_container", line 3355, in <module>
main()
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080916.98-133068627776311/lxc_container", line 1724, in main
lxc_manage.run()
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080916.98-133068627776311/lxc_container", line 1605, in run
action()
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080916.98-133068627776311/lxc_container", line 1145, in _started
self._config()
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080916.98-133068627776311/lxc_container", line 714, in _config
_, _value = option_line.split('=')
ValueError: too many values to unpack
With such a task::
tasks:
- lxc_container:
name: buildbot-master
container_config:
- "lxc.mount.entry = {{ cwd }} srv/peopletest none defaults,bind,uid=0,create=dir 0 0"
This fixes::
Traceback (most recent call last):
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080800.61-38257321141340/lxc_container", line 3353, in <module>
main()
File "/home/jpic/.ansible/tmp/ansible-tmp-1435080800.61-38257321141340/lxc_container", line 1712, in main
if not HAS_LXC:
NameError: global name 'HAS_LXC' is not defined
Due to a spurious newline we corrupted the payload. It depends on the order of the headers and if there were headers added by vSphere.
The Accept header was also not needed.
