Fixes#10003
When endpoints are removed from an EndpointSlice resource, the destination controller builds a list of addresses to remove. However, if any of the removed endpoints have a Pod as their targetRef, we will attempt to fetch that pod to build the address to remove. If that pod has already been removed from the informer cache, this will fail and the endpoint will be skipped in the list of endpoints to be removed. This results in stale endpoints being stuck in the address set and never being removed.
We update the endpoint watcher to construct only a list of endpoint IDs for endpoints to remove, rather than fetching the entire pod object. Since we no longer attempt to fetch the pod, this operation is now infallible and endpoints will no longer be skipped during removal.
We also add a `TestEndpointSliceScaleDown` test to exercise this.
Signed-off-by: Alex Leong <alex@buoyant.io>
Since Go 1.13, errors may "wrap" other errors. [`errorlint`][el] checks
that error formatting and inspection is wrapping-aware.
This change enables `errorlint` in golangci-lint and updates all error
handling code to pass the lint. Some comparisons in tests have been left
unchanged (using `//nolint:errorlint` comments).
[el]: https://github.com/polyfloyd/go-errorlint
Signed-off-by: Oliver Gould <ver@buoyant.io>
Go's test runner (`go test`) can be non-deterministic with the order in
which it runs the tests. Tests in Go seem to be always
run in parallel, but the specifics here differ depending on the
available CPU.
We can take advantage of parallelism here to get better timing on our
tests, however, we need to block the start of each test until the
control plane (or extension) pods are ready. In each `TestMain`, we
block until the pods are ready.
Signed-off-by: Matei David <matei@buoyant.io>
This change re-works how tests are organised, by splitting each
integration test based on its functionality. Each "suite", is a
directory that compromises tests that will be specific to that suite;
deep tests will test the control plane, viz tests will target the
monitoring stack, and so on.
As part of this change, aside from moving the tests, the GH workflow
file has also been changed to remove old test names, and add (or
replace) new ones where appropriate.
To have a fully functioning viz suite, a new flag has been added to the
TestHelper to discern between tests that make use of the monitoring
stack and tests that do not. The external suite also has a duplicated
`edges` test so we can make sure interactions with external prometheus
don't break.
Signed-off-by: Matei David <matei@buoyant.io>
Alex Leong
Krzysztof Dryś
Oliver Gould
Matei David