docs(configuration): add cluster configuration #851
Conversation
|
@robszumski @colegleason @xiangli-cmu |
| @@ -41,6 +41,9 @@ configuration files. | |||
| * `-peer-election-timeout` - The number of milliseconds to wait before the leader is declared unhealthy. | |||
| * `-peer-heartbeat-interval` - The number of milliseconds in between heartbeat requests | |||
| * `-snapshot=false` - Disable log snapshots. Defaults to `true`. | |||
| * `-cluster-active-size` - The number of expected peer-mode instances in the cluster. It is used only when the instance creates a new cluster. | |||
There was a problem hiding this comment.
It is used only when the instance creates a new cluster.
Is there a better way to explain this? Basically it sets the value in the cluster and then the cluster value is used from then onwards?
There was a problem hiding this comment.
It sets the value if it decides to build a one-node cluster by itself, and does it in the init step.
The cluster uses this value from then onwards.
If user wants to change it later, they could only do it through HTTP API currently.
There was a problem hiding this comment.
OK. What about:
"Only applied if the etcd instance is the first peer in the cluster."
I think we also need a section of this that talks about and links to the cluster config API similar to my other PR: https://github.com/coreos/etcd/pull/850/files#diff-c2e93b5d997618cfe14f6ea88f78f0d3R15
There was a problem hiding this comment.
Agree with you.
Your PR is also on this file, so I would rely on that one to make it.
There was a problem hiding this comment.
Whoops, I didn't realize you wanted to keep both issues.
| @@ -41,6 +41,9 @@ configuration files. | |||
| * `-peer-election-timeout` - The number of milliseconds to wait before the leader is declared unhealthy. | |||
| * `-peer-heartbeat-interval` - The number of milliseconds in between heartbeat requests | |||
| * `-snapshot=false` - Disable log snapshots. Defaults to `true`. | |||
| * `-cluster-active-size` - The number of expected peer-mode instances in the cluster. Only applied if the etcd instance is the first peer in the cluster. | |||
There was a problem hiding this comment.
The last nit I have on this is that we don't use the term peer-mode very often. When we explain this to be people verbally we say something about how it participates or not in the consensus protocol. Can we mention that here?
"The expected number of instances participating in the consensus protocol. Only applied if the etcd instance is the first peer in the cluster."
|
@robszumski |
|
lgtm. Can you look over #850 again? |
docs(configuration): add cluster configuration
#849