Taking VMs down
Table of contents
Scenario
Let’s say that you’d like to make sure that if one of the VMs in your cluster goes down (let’s say one that’s running master nodes), some service running on them still works well.
Well, PowerfulSeal can do that for you.
Cloud drivers
If you’d like to modify the state of VMs, you’re going to need to configure a cloud driver. You can learn how to do that here.
Policy
- name: Test load-balancing on master nodes
steps:
# take down a single master VM
- nodeAction:
# pick all master VMs
matches:
- property:
name: "name"
value: ".*master.*"
filters:
# only pick the VMs which are in UP state
- property:
name: "state"
value: "UP"
# and only pick a single VM
- randomSample:
size: 1
actions:
- stop:
force: false
# issue an HTTP probe to the url of the service we expect to see up
- probeHTTP:
target:
url: "http://load-balancer.example.com"
# we can allow some retries, if needed
retries: 3
# we can also make more requests to make sure we hit all instances
count: 100
# restart the node which are down
- nodeAction:
# pick all master VMs
matches:
- property:
name: "name"
value: ".*master.*"
filters:
# filter down to only the VMs which are in DOWN state
- property:
name: "state"
value: "DOWN"
actions:
- start
Auto-restart
By default, VMs stopped by the Seal will be auto-restarted at the end of scenario (or if there is a failure, in the cleanup). You can disable that by setting autoRestart to false.
- name: Take machines down, and don't bring them back up
steps:
# take down a single master VM
- nodeAction:
# pick all master VMs
matches:
- property:
name: "name"
value: ".*master.*"
filters:
- randomSample:
size: 1
actions:
- stop:
autoRestart: false