Skip to content
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.

Sign up for GitHub

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

Jump to bottom

[Backport stable/8.0] Only check that the topology reports all replicas #10084

Merged
merged 9 commits into from
Aug 17, 2022
Merged

[Backport stable/8.0] Only check that the topology reports all replicas #10084

merged 9 commits into from
Aug 17, 2022

Commits on Aug 16, 2022

  1. feat(test-util): backport TopologyAssert and tests 

    Backports the expanded TopologyAssert from the main branch to
stable/8.0, as well as its tests.
    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    c369558 View commit details
    Browse the repository at this point in the history

  2. test(qa): only check that the topology reports all replicas 

    Instead of using the helper `isComplete`, manually check that the
topology contains all replicas, omitting the check for the partition
health. While this would be safer, there is a race condition where the
partitions may be unhealthy for a short time on start up. For this test,
we only really want to check that the nodes find each other, not that
processing works, as it should be largely unaffected.
    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    05539f7 View commit details
    Browse the repository at this point in the history

  3. test(qa): fix IPv6 flakiness 

    The IPv6 test is sometimes flaky because it may take a while for a full
cluster to start: 3 brokers and one gateway. To test IPv6, however, we
don't really need so many nodes, and having 1 broker and 1 gateway is
enough: all ports are tested properly (except the management port, which
is tested already by Spring).

Additionally, this fixes that if the test fails once, the network may
not be removed properly afterwards, which causes further iterations to
fail. We ensure the network is closed here, and also we print out a
proper error if we fail to start the network on the next run.

Finally I took the time to clean up the test using new APIs provided by
zeebe-test-container.
    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    a95f31d View commit details
    Browse the repository at this point in the history

  4. test: let embedded broker rule wait for healthy leader 

    Previously, the rule only waited for a leader of partition 1 before
returning from `startBroker`. With this change, the leader must also
report as healthy.

This is intended to fix some flaky tests where after the startup,
commands send to the leader of partition 1 failed because no command-api
handler was registered. Since registration of handlers is required for
a healthy status, this should hopefully prevent these kinds of flaky
tests.
    @lenaschoenburg @npepinpe
    lenaschoenburg authored and npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    80f5eea View commit details
    Browse the repository at this point in the history

  5. refactor: use TopologyAssert to wait for healthy topology

    @lenaschoenburg @npepinpe
    lenaschoenburg authored and npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    72065d2 View commit details
    Browse the repository at this point in the history

  6. refactor(qa): fix flakiness on startup of ClusterRule 

    The `ClusterRule` uses the topology to figure out if the cluster is
ready before a test starts. There was a flaky condition where it might
start when all partitions are present, but before they are healthy. This
change was pulled from #9502 - however in that PR, it was only done in
the broker module's `EmbeddedBrokerRule`.

At the same time, the rule is updated such that everywhere it was
waiting for the topology, now also uses `TopologyAssert`.
    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    af34d1f View commit details
    Browse the repository at this point in the history

  7. test: update usage of refactored TopologyAssert

    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    0f01b41 View commit details
    Browse the repository at this point in the history

  8. refactor(qa): remove dead code and fix some warnings

    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    3fe3bb1 View commit details
    Browse the repository at this point in the history

  9. test(qa): avoid race condition when polling for topology by ignoring … 

    …gRPC exceptions
    @npepinpe
    npepinpe committed Aug 16, 2022
    Configuration menu
    Copy the full SHA
    0da5eff View commit details
    Browse the repository at this point in the history