api: add listener_address_restriction_hint in Node to hint known listening ports on the node#10370
Conversation
|
@htuch pls review |
|
To be more clear about the use-case: This is for cases where the server has already decided which ports it will listen on and needs configuration for those ports. |
|
Looked into the For some reason |
|
This was suggested by @htuch, so I assume he'll review. |
|
@htuch Just FYI I did try to test the changes on my Mac but there are infra/tooling problems I am unable to debug. Pls let me know if the failures are related to my changes. |
htuch
left a comment
There was a problem hiding this comment.
LGTM, but you need to run fix_format (to forward propagate into v3) and would be good to hear from at least one other @envoyproxy/api-shepherds
/wait
api/envoy/api/v2/core/base.proto
Outdated
There was a problem hiding this comment.
This PR looks good, just trying to see if we can maybe name this a bit more descriptively (I know I was responsible for the original suggestion :P)
How about listener_bound_addresses, listening_addresses, listener_available_addresses?
There was a problem hiding this comment.
I'll pick listening_addresses as the shortest of them all.
So I ran And that finished successfully and generated following changes: Should I commit all these changes? And more importantly, why is |
htuch
left a comment
There was a problem hiding this comment.
@sanjaypujare good catch on the v3 file move:
- Yes, please add all the new shadow files.
- See my comment on how to ensure the new socket options file is in the right place.
/wait
There was a problem hiding this comment.
You'll need to add option (udpa.annotations.file_migrate).move_to_package = "envoy.config.core.v3"; (as done in other files in this package) for the correct file movement in v3.
`
mattklein123
left a comment
There was a problem hiding this comment.
LGTM at a high level as an alternate API reviewer. Thank you!
@htuch done. PTAL. |
|
@sanjaypujare I think you ended up colliding with a known master breakage. Can you sync past #10426, merge master and try again? You will also need to fix DCO: https://github.com/envoyproxy/envoy/blob/master/CONTRIBUTING.md#fixing-dco |
…t about known listening ports on the node Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
af49960 to
15ff0cc
Compare
Done both. PTAL. |
|
@sanjaypujare see the docs warning: I suggest adding to https://github.com/envoyproxy/envoy/blob/master/docs/root/api-v3/common_messages/common_messages.rst and https://github.com/envoyproxy/envoy/blob/master/docs/root/api-v2/common_messages/common_messages.rst as appropriate. |
Signed-off-by: Sanjay Pujare <sanjaypujare@users.noreply.github.com>
|
@htuch v3 common_messages.rst updated (v2 was already done). |
@htuch thanks! Once this is merged the downstream design/implementation activities will continue. |
Description: Added an optional field in
Nodeas a generic hint to the management server about known listening ports on the nodeRisk Level: Low
Testing: manually tested by compiling protos into generated Java code and using the code in a Java project
Docs Changes: N/A
Release Notes: N/A
Signed-off-by: Sanjay Pujare sanjaypujare@users.noreply.github.com