test: Add non-aggregated CDS-over-gRPC integration test#5228
test: Add non-aggregated CDS-over-gRPC integration test#5228htuch merged 54 commits intoenvoyproxy:masterfrom
Conversation
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
htuch
left a comment
htuch
left a comment
There was a problem hiding this comment.
Great, this should be super useful to have on hand, much easier to understand than the ADS integration test.
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
|
🙀 Error while processing event: |
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
|
Finally got it working, this time I'm pretty sure without any hacky workarounds or misunderstandings! It looks like the big issue was needing a way to defer the framework's check that listeners are listening. PTAL. Also, a minor thing, while I'm making changes in IntegrationTestServer: would people be ok with renaming IntegrationTestServer::create() to createAndStart()? The first couple of times tracing through the code, I found myself surprised by how much create() was actually doing - i.e., calling start(), which does a lot. |
Signed-off-by: Fred Douglas <[email protected]>
fredlas
changed the title
|
Wait, sorry, it's still sending the test request to the CDS server, because the static listener's route points to my_cds_cluster (in order to get the listener through initialization). I think I'm going to need to manually add a listener outside/after the normal initialization flow. |
…t to Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
…nstream Signed-off-by: Fred Douglas <[email protected]>
htuch
left a comment
htuch
left a comment
There was a problem hiding this comment.
LGTM modulo one final nit. Thanks for all the great cleanup work you've done in this PR.
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
Signed-off-by: Fred Douglas <[email protected]>
|
Sorry, noticed some last-minute cleanup to be done! |
|
Forgot to tag this as related to #4991. |
) While trying to implement incremental CDS support in Envoy, I found it extremely difficult to work with the integration tests, which I hoped to adapt/copy for incremental. The only test using gRPC was ADS. I decided that a more manageable first step was to add a test for just (non-incremental) CDS over gRPC. Risk Level: none, just a new test Signed-off-by: Fred Douglas <[email protected]> Signed-off-by: Fred Douglas <[email protected]>
3 participants
Description: While trying to implement incremental CDS support in Envoy, I found it extremely difficult to work with the integration tests, which I hoped to adapt/copy for incremental. The only test using gRPC was ADS. I decided that a more manageable first step was to add a test for just (non-incremental) CDS over gRPC.
Risk Level: none, just a new test