Using net.JoinHostPort to format host:port strings

[mialinx]

IPv6 based addrs shoul look like [host]:port instead of host:port

A Pull Request should be associated with an Issue.

We wish to have discussions in Issues. A single issue may be targeted by multiple PRs.
If you're offering a new feature or fixing anything, we'd like to know beforehand in Issues,
and potentially we'll be able to point development in a particular direction.

Related issue: https://github.com/github/orchestrator/issues/0123456789

Thank you! We are open to PRs, but please understand if for technical reasons we are unable to accept each and any PR

Description

This PR [briefly explain what is does]

In case this PR introduced Go code changes:

• contributed code is using same conventions as original code
• code is formatted via gofmt (please avoid goimports)
• code is built via ./build.sh
• code is tested via go test ./go/...

[shlomi-noach]

Thank you. This will unfortunately break a lot of things. Building the correct [host]:port is one thing, but there's a myriad of places where we assume the format is host:port and not [host]:port.

It is very unfortunate, indeed, but the change to ipv6 would need to be massive.

[shlomi-noach]

Hi @mialinx, I see you keep pushing changes. Would you like to perhaps explain what you are doing? Like I wrote earlier, there's a LOT to change if we want to go ipv6 that it's difficult to even find where to begin with.

[mialinx]

Hi @shlomi-noach,

Yep, I'm trying to find all suspicious places like using ':', ":", "%s:" in the code, check if they are related to hostname, port handling and replace them with IPv6 aware code.

As far as I understood, there are two main points:

• Check that all services may be conneсted using IPv6 addr. I have tested using IPv6 for specifing backend MySQL database, RAFT nodes, Zookeper store, and discovering mysql clusters (even @@hostname and show slave hosts are IPv6 addrs). And it works to me.
• Using host:port as identifier for comparing instances. I think it should also work if we carefully change all occurrences of host:port. Anyway for IPv4 and hostnames identifiers should stay the same.

It's very likely that I missed somtheing :( What other difficulties do you see ?

[shlomi-noach]

Your changes are thorough, indeed. What concerns me are all the places where we parse host:port. I'm yet to consider all the cases, my focus right now is on a few other high priority issues. Meanwhile I applaud you for this undertaking.

[mialinx]

It seems that integration tests flap as they depened on map keys order:
#785

[pznamensky]

@mialinx thanks for your work!
The only reason we don't use orchestrator is IPv6 related issues like this: #616
I hope this PR will be approved.

[shlomi-noach]

I completely dropped the ball on this PR. It'll take me time to re0review it.

[d3mash]

could you please take a look at this PR?

[mainameiz]

We really really need this PR. Please take a look 🙏

[shlomi-noach]

I apologize for not giving this attention. I very much appreciate the contribution and your work. It'll take a while before I've fully reviewed this. I'll meanwhile send for production testing in staging environment.