-
Notifications
You must be signed in to change notification settings - Fork 4.8k
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
release: flipping deprecated features to fatal-by-default #6509
Conversation
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM but would like to discuss hosts
a bit more with the group before we merge.
"envoy.deprecated_features.route.proto:enabled", | ||
"envoy.deprecated_features.tcp_proxy.proto:deprecated_v1", | ||
"envoy.deprecated_features.fault.proto:type", | ||
"envoy.deprecated_features.cds.proto:hosts", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
After further discussions with @htuch this morning my current feeling is we should just undeprecate this. IMO it's not worth the pain. Would like to hear what others think.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have no objections for making hosts
fatal-by-default, but un-deprecating sounds quite appealing, since Istio is still using hosts
(though, I don't believe there are any blockers for migrating away from it).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
"envoy.deprecated_features.tcp_proxy.proto:deprecated_v1" is used in Istio.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there a reasonable replacement for deprecated_v1
or is it still not ready? If the former, we can move forward with the deprecation per the policy. If not, we should undeprecate this field. cc @PiotrSikora
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
tcp_proxy.deprecated_v1
is not used in Istio, so we don't have objections removing that.
Having said that, Envoy still doesn't have a proper replacement for this feature, since no one implemented #4457 yet.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for @PiotrSikora. I think in that case we also need to undeprecate tcp_proxy.deprecated_v1
until we have a complete replacement.
"envoy.deprecated_features.config_source.proto:UNSUPPORTED_REST_LEGACY", | ||
"envoy.deprecated_features.ext_authz.proto:use_alpha", | ||
"envoy.deprecated_features.route.proto:enabled", | ||
"envoy.deprecated_features.tcp_proxy.proto:deprecated_v1", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Update: Istio is not using tcp_proxy.deprecated_v1
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Link? I couldn't find any references to it last time I looked (few weeks ago).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry, I don't find it either. Update my original comment
tcp_proxy deprecated_v1 doesn't seem to be used - only has "Routes" which we're not using. |
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Deprecate all the features!
"envoy.deprecated_features.ext_authz.proto:use_alpha", | ||
"envoy.deprecated_features.route.proto:enabled", | ||
"envoy.deprecated_features.fault.proto:type", | ||
"envoy.deprecated_features.route.proto:runtime_key", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nit: this could be sorted alphabetically.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think we could locally, but tdb if we want to have these grouped by release rather than alphabetical overall. I'm going to merge without fixing just because I want it have a day or two where they're flipped and I'm around for questions and we can sort out ordering next release or in a follow-up?
Somewhat ironically, bind_to_port isn't currently flagged as deprecated but use_original_dst was, so manually extended its lifetime per discussion on #5355.
It wasn't clear to me if the tcp_proxy deprecated_v1 fields were also part of #5355 so tagging Piotr as a reviewer.
I'll envoy-announce once we've nailed down the fields and gotten approval from all.
Risk Level: High (first time using this process - it will likely cause problems for someone)
Testing: tests pass.
Docs Changes: n/a
Release Notes: Should we relnote these? They're already in DEPRECATED.md