Start to kill the word "minion"

[thockin]

Mostly mechanican changes that were pretty obvious.

This series is sort of risky. I think we only want the first commit here and the rest later? Need to do docs, too, and that is pretty manual.

I don't think I changed any "minion" string where the context was salt (for which "minion" is correct")

[k8s-bot]

EXPERIMENTAL JENKINS PR BUILDER: e2e build succeeded.

[bgrant0607]

Awesome!

[bgrant0607]

Ref #1111

[k8s-bot]

EXPERIMENTAL JENKINS PR BUILDER: e2e build succeeded.

[bgrant0607]

Looks pretty reasonable to me. Did you run e2e?

[thockin]

Did not run e2e yet, no time today.

On Wed, Jun 3, 2015 at 10:30 PM, Brian Grant notifications@github.com
wrote:

Looks pretty reasonable to me. Did you run e2e?

—
Reply to this email directly or view it on GitHub
#9176 (comment)
.

[justinsb]

So to confirm: the new names are masters & nodes? And a master is not a node? (I have a PR coming for configuring security groups for AWS, and was going to give the masters & TNFKAM* different security groups; I might as well give them the right names if I can)

(* the nodes formerly known as minions)

[bgrant0607]

The current name is node.

The "master" is currently not a full schedulable node running kubelet, kube-proxy, etc., but that's the future -- self-hosting of the control plane (see #246 for some old discussion).

[thockin]

To surface a side discussion from a meeting - maybe we're over the hump of
confusing k8s "minion" with salt minions? Maybe we can just keep "minion"?

On Thu, Jun 4, 2015 at 3:04 PM, Brian Grant notifications@github.com
wrote:

The current name is node.

The "master" is currently not a full schedulable node running kubelet,
kube-proxy, etc., but that's the future -- self-hosting of the control
plane (see #246
#246 for some
old discussion).

—
Reply to this email directly or view it on GitHub
#9176 (comment)
.

[bgrant0607]

grep -r -i minion cluster cmd contrib docs examples hack pkg plugin test www |wc -l
1713
grep -r -i node cluster cmd contrib docs examples hack pkg plugin test www |wc -l
5411

[wojtek-t]

@thockin - what is the plan for this PR? Are we going to proceed with it or should we close it?

[k8s-bot]

GCE e2e build/test failed for commit cefdc5d.

• Build Log
• Test Artifacts
• Internal Jenkins Results

[k8s-github-robot]

Labelling this PR as size/XL

[wojtek-t]

@thockin - what is the plan for this PR? Are we going to proceed with it or should we close it?

[thockin]

We should still do it, but this PR is pretty obsolete.

On Tue, Sep 8, 2015 at 5:08 AM, Wojciech Tyczynski <notifications@github.com

wrote:

@thockin https://github.com/thockin - what is the plan for this PR? Are
we going to proceed with it or should we close it?

—
Reply to this email directly or view it on GitHub
#9176 (comment)
.

[wojtek-t]

@thockin can we close this one? It might be easier to start it from scratch than to rebase this one :)

[wojtek-t]

@thockin - thanks. I can try to redo some if it in my free cycles.

[davidopp]

I still think we should use "minion" in cases where we are referring to something that can only be a worker node.

Even once we have self-hosting I think it will be a useful distinction as some people won't want to share their master nodes with user containers for security and performance isolation reasons.

[wojtek-t]

@davidopp can you please give an example of where you think we should leave minion? From what you wrote it seems to me that you're referring to places that are exposed to user (e.g. kube-up script?).
My thinking is that since we explicitly name it "Node" in v1 api, all places that refer to it (e.g. pkg/registry) should also have "node"-related names (not "minion" ones).