@@ -66,6 +66,24 @@ properties:
6666 until then, the default will be the recommended 60 seconds.
6767 see also: https://docs.yugabyte.com/latest/manage/upgrade-deployment
6868 default: 60
69+ post_deploy_delay_factor:
70+ description: |
71+ must be a whole number.
72+ consider removing this property altogether a TODO.
73+ multiplication factor used in the post-deploy hook to calculate how long each node should
74+ sleep before attempting to run itself. this number is multipled by the node index to
75+ calculate how long, in seconds, to sleep.
76+ for example, given a delay factor of 5, on node index 0,
77+ the node will sleep 0 seconds before running its post-deploy. meanwhile on node index 1,
78+ the node will sleep for 5 seconds before running its post-deploy. and on node index 2, the
79+ node will sleep for 10 seconds before running its post-deploy.
80+ and for example, given a delay factor of 8, node index 0 sleeps for 0 seconds, index 1
81+ sleeps for 8 seconds, index 2 sleeps for 16 seconds, and so on.
82+ this is not the most incredible solution in the world, but it is a relatively okay
83+ way to prevent nodes from stumbling over each other.
84+ in the future it may be preferrable to have only a subset of nodes try to perform
85+ post-deploy at all, but this should be an acceptable solution for the time being.
86+ default: 10
6987
7088 tls.allow_insecure_connections:
7189 description: |
0 commit comments