Copilot commented on code in PR #8676:
URL: https://github.com/apache/hbase/pull/8676#discussion_r4076433875
##########
dev-support/read-replica/python/src/hbase_docker_client.py:
##########
@@ -141,26 +136,27 @@ def _get_pid_from_jps(self, process_name: str) -> int |
None:
return None
def wait_for_hbase_ui(self) -> bool:
- """Checks for a 200 OK on the HBase Master UI."""
- # Read HBASE_HOST from environment, falling back to 'localhost' for
host-native execution
- url = f"http://{self._hbase_host}:{self._hbase_ui_port}";
- logger.info(f"Waiting for HBase UI: {self._cluster_name} on {url}")
- last_exception = None
+ """
+ Checks for a 200 OK on the HBase Master UI inside the container using
curl.
+ """
+ check_cmd = f"curl -s -f --max-time 3
http://{self._container_name}:{self._hbase_ui_port} > /dev/null"
Review Comment:
The final image explicitly installs `wget` but not `curl`
(Dockerfile:90-102), so this readiness command can fail with `curl: command not
found` inside both test containers and exhaust all retries before the suite
starts. Use the already-installed `wget` or add `curl` to the final image.
##########
dev-support/read-replica/README.md:
##########
@@ -113,13 +113,13 @@ read-replica/
## CI: Jenkins Nightly Pipeline
**Files:**
-- `dev-support/Jenkinsfile` — stage definition (`hbase read-replica feature
checks`)
-- `dev-support/hbase_nightly_read_replica_test.sh` — test driver script
+- `dev-support/read-replica/Jenkinsfile` — pipeline definition (`hbase
read-replica feature checks`)
+- `dev-support/read-replica/hbase_nightly_read_replica_test.sh` — test driver
script
### When It Runs
-The read-replica stage runs as part of the HBase nightly build on the `master`
and `branch-3`
-branches. It executes in parallel alongside the other nightly check stages
(Yetus, JDK8/11/17).
+The read-replica tests run as their own standalone nightly pipeline on the
`master` and `branch-3`
+branches, separate from the main HBase nightly build.
Review Comment:
These new pipeline notes are followed by stale instructions that still say
the driver cleans and archives `cluster1/logs` and `cluster2/logs` (including
the mounted-volumes section), but this change removes those mounts and cleanup
variables. Please update those sections too, otherwise the documentation
directs operators to log directories that the test no longer creates or
archives.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
