When haproxy is configured to track the status of a server using a
server in another backend it will not report all status as expected.
Using substring matching will solve this.
The CSV will output the server status in the following way:
MAINT: MAINT (via pxname/svname)
UP: UP
DOWN: DOWN
Haproxy doc on trach: https://cbonte.github.io/haproxy-dconv/1.8/configuration.html#5.2-track
* HAProxy: Support waiting for nodes to drain before maint mode
A common task when working with a HAProxy-managed node is to first "drain" it
and then place it into maintenance mode (to be repaired or redeployed).
(Draining such a node consists of preventing new connections from being
established while waiting for active sessions to expire/close.) This commit
creates a new `drain' parameter for Ansible's HAProxy module, which, when set to
`yes` in conjunction with `state: disabled` and `wait: yes`, causes the module
to attempt to set a node to drain, wait for it to finish draining, and put the
node into maintenance mode. The action is recorded as a success if these steps
are completed before a maximum wait timeout is reached.
Implements: https://github.com/ansible/ansible-modules-extras/issues/521
Acknowledgments: This is based on user krislindgren's pull request #21420
* Correct how drain option is handled
Previously the `drain` parameter would have no effect, since the `disabled`
state would still immediately place backend servers into maintenance mode.
* Split modules/network into two parts
Given the dedicated team we we have working on Ansible Networking a
clearer split is needed between Networking modules and "things that
happen to use the network"
* nmcli to net_tools
* nmcli moved
2017-04-27 16:42:41 +01:00
Renamed from lib/ansible/modules/network/haproxy.py (Browse further)
