[chore] Update google.golang.org/grpc to v1.59.0

[tmshort]

require grpc@v1.59.0 was replaced by v1.40.0, which is now out-of-date.

This will allow additional dependency updates.

Reviewer Checklist

• Implementation matches the proposed design, or proposal is updated to match implementation
• Sufficient unit test coverage
• Sufficient end-to-end test coverage
• Bug fixes are accompanied by regression test(s)
• e2e tests and flake fixes are accompanied evidence of flake testing, e.g. executing the test 100(0) times
• tech debt/todo is accompanied by issue link(s) in comments in the surrounding code
• Tests are comprehensible, e.g. Ginkgo DSL is being used appropriately
• Docs updated or added to /doc
• Commit messages sensible and descriptive
• Tests marked as [FLAKE] are truly flaky and have an issue
• Code is properly formatted

[grokspawn]

Is this going to result in a bunch of dependabot/synk/other vuln scans to throw up because of these earlier versions? Would we instead need to change the replace to v1.59.0 here?

[tmshort]

I don't think so, because the "required" version used is v1.59.0, and that's the version within vendor. We ought to remove the replaces line, and putting in the same version doesn't make much sense.

[joelanford]

Oh interesting. When I was investigating #3146, I didn't see the replace, so my toy client was using 1.59.0.

It looks like the behavior of the client changes quite a bit from 1.40 to 1.59 in that:

• in 1.40, the status changes from READY to CONNECTING/TRANSIENT_FAILURE when the catalog pod is killed, and eventually goes back to READY when the new catalog pod comes back up.
• in 1.59, the status changes from READY to IDLE when the catalog pod is killed and does not attempt reconnection.

That might explain why we're using a replace. Ultimately, I agree we should drop the replace and resolve these underlying issues.

[joelanford]

I did a little bit of testing and found that we can get the client to reconnect when we detect the idle state by adding the following code snippet in pkg/controller/registry/grpc/source.go in the watch() function (around line 231, just before the notify channel send)

// always try to reconnect if the connection goes idle
if newState == connectivity.Idle {
    source.Conn.Connect()
}

My guess is that this will resolve the issues with e2es in this PR and it may resolve the issue described in #3146 as well.

[joelanford]

Looks like it all passed, and you've also accounted for @stevekuznetsov's comment from my PR.

Shall we go ahead and merge this one? (I mainly put mine up just to get a quicker signal earlier this afternoon)

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: joelanford, tmshort

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [joelanford,tmshort]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[joelanford]

#3147 just merged, so this one is no longer necessary.