[FLINK-27160] Add e2e tests for session job #179
Conversation
|
FYI, I found that the test can not completely run on the Mac now. Because the minikube ip can not be reached from the local machine. |
|
The result of |
I use the |
e2e-tests/test_last_state_upgrade.sh
Outdated
| @@ -47,6 +48,7 @@ function wait_for_jobmanager_running() { | |||
| function assert_available_slots() { | |||
| expected=$1 | |||
| ip=$(minikube ip) | |||
| curl http://$ip/default/${CLUSTER_ID}/overview | |||
There was a problem hiding this comment.
Why do we need this line?
e2e-tests/test_sessionjob_ha.sh
Outdated
| retry_times 30 3 "kubectl get deploy/${CLUSTER_ID}" || exit 1 | ||
|
|
||
| kubectl wait --for=condition=Available --timeout=${TIMEOUT}s deploy/${CLUSTER_ID} || exit 1 | ||
| jm_pod_name=$(kubectl get pods --selector="app=${CLUSTER_ID},component=jobmanager" -o jsonpath='{..metadata.name}') | ||
|
|
||
| echo "Waiting for jobmanager pod ${jm_pod_name} ready." | ||
| kubectl wait --for=condition=Ready --timeout=${TIMEOUT}s pod/$jm_pod_name || exit 1 | ||
|
|
||
| wait_for_logs $jm_pod_name "Rest endpoint listening at" ${TIMEOUT} || exit 1 |
There was a problem hiding this comment.
Could be deduplicated by having wait_for_jobmanager_running in utils.sh.
e2e-tests/test_sessionjob_ha.sh
Outdated
| SESSION_CLUSTER_IDENTIFIER="flinkdep/session-cluster-1" | ||
| SESSION_JOB_IDENTIFIER="sessionjob/flink-example-statemachine" | ||
|
|
||
| function cleanup_and_exit() { |
There was a problem hiding this comment.
Could be extracted to util.sh.
| wait_for_logs $jm_pod_name "Rest endpoint listening at" ${TIMEOUT} || exit 1 | ||
| } | ||
|
|
||
| function assert_available_slots() { |
There was a problem hiding this comment.
Could be extracted to util.sh and deduplicated.
|
I have addressed your comments, please take a look again @wangyang0918 |
wangyang0918
force-pushed
the
e2e-test-0422
branch
from
d22ebba to
fb13a14
Compare
This PR is meant to add the e2e test for the session job. It mostly aligned to the flinkdep tests.