Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

proposal: TLS by default for all etcd peer communication #4022

Closed
philips opened this issue Dec 18, 2015 · 6 comments
Closed

proposal: TLS by default for all etcd peer communication #4022

philips opened this issue Dec 18, 2015 · 6 comments

Comments

@philips
Copy link
Contributor

philips commented Dec 18, 2015

Problem: We want more people using TLS for etcd in peer communication. We don't want this to happen: https://blog.shodan.io/its-still-the-data-stupid/

Solution: Make it easier to get etcd into a TLS mode and operate the lifecycle of the cluster security and public keys. To accomplish this goal we should make TLS the default for etcd.

Proposal document: https://docs.google.com/document/d/11nC7y42N0TwJnkKvIBpNRURAfOJYiSHAGPCX-VSke2c/edit#heading=h.xubaa8vzokts

@xiang90
Copy link
Contributor

xiang90 commented Dec 18, 2015

Problem: We want more people using TLS for etcd in peer communication.

How will the TLS on peer help with the unauthed access? TLS+Auth client can only prevent outsider to corrupt the cluster. Shall we suggest people to only expose peer address in VPC or local network if possible?

@philips
Copy link
Contributor Author

philips commented Dec 18, 2015

It is very possible that someone could use unencrypted peer communication to extract cluster secrets on the wire and unverified communication to get a snapshot of the database. I have another proposal I am writing for the client side but I think the peer side is a clear first step.

@xiang90
Copy link
Contributor

xiang90 commented Dec 18, 2015

@philips Yea, I agree it is helpful. I mean this proposal will not prevent the exact same issue described https://blog.shodan.io/its-still-the-data-stupid/. We need client side protection.

@philips
Copy link
Contributor Author

philips commented Dec 18, 2015

@xiang90 agreed on client side, that is the next proposal :)

@markhowells
Copy link

@xiang90 @philips I agree with the goal of getting making TLS more widely used and using TLS for peer comms is still difficult for many users. However, it's the client side that I'm concerned about - making clients use TLS natively is problematic (especially for those of us that have to use legacy apps), but using the etcd proxy to terminate TLS for such clients is broken - #3894

@gyuho
Copy link
Contributor

gyuho commented Mar 21, 2018

Moving discussion to #9475.

@gyuho gyuho closed this as completed Mar 21, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Development

No branches or pull requests

4 participants