[
https://issues.apache.org/jira/browse/FLINK-15669?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aljoscha Krettek reopened FLINK-15669:
--------------------------------------
I partially reverted these changes:
Only the tests are reverted because the fix is still good. The added
tests don't work because of a race condition: if the query finishes
before the test cancels them the job status will be FINISHED, and not
CANCELLED, as the text expects.
revert on release-1.10: 4518d18a726b35de9ff802d155fd8100dc711a63
revert on master: cbaea2924a90a1141d9794f6cd8f561e77ddc11a
> SQL client can't cancel flink job
> ---------------------------------
>
> Key: FLINK-15669
> URL: https://issues.apache.org/jira/browse/FLINK-15669
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Client
> Affects Versions: 1.10.0
> Reporter: godfrey he
> Assignee: godfrey he
> Priority: Critical
> Labels: pull-request-available
> Fix For: 1.10.1, 1.11.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> in sql client, CLI client do cancel query operation through {{void
> cancelQuery(String sessionId, String resultId)}} method in {{Executor}}.
> However, the {{resultId}} is a random UUID, is not the job id. So CLI client
> can't cancel a running job.
> related code in {{LocalExecutor}}:
> {code:java}
> private <C> ResultDescriptor executeQueryInternal(String sessionId,
> ExecutionContext<C> context, String query) {
> ......
> // store the result with a unique id
> final String resultId = UUID.randomUUID().toString();
> resultStore.storeResult(resultId, result);
> ......
> // create execution
> final ProgramDeployer deployer = new ProgramDeployer(
> configuration, jobName, pipeline);
> // start result retrieval
> result.startRetrieval(deployer);
> return new ResultDescriptor(
> resultId,
> removeTimeAttributes(table.getSchema()),
> result.isMaterialized());
> }
> private <T> void cancelQueryInternal(ExecutionContext<T> context, String
> resultId) {
> ......
> // stop Flink job
> try (final ClusterDescriptor<T> clusterDescriptor =
> context.createClusterDescriptor()) {
> ClusterClient<T> clusterClient = null;
> try {
> // retrieve existing cluster
> clusterClient =
> clusterDescriptor.retrieve(context.getClusterId()).getClusterClient();
> try {
> // ======== cancel job through resultId =======
> clusterClient.cancel(new
> JobID(StringUtils.hexStringToByte(resultId))).get();
> } catch (Throwable t) {
> // the job might has finished earlier
> }
> } catch (Exception e) {
> throw new SqlExecutionException("Could not retrieve or
> create a cluster.", e);
> } finally {
> try {
> if (clusterClient != null) {
> clusterClient.close();
> }
> } catch (Exception e) {
> // ignore
> }
> }
> } catch (SqlExecutionException e) {
> throw e;
> } catch (Exception e) {
> throw new SqlExecutionException("Could not locate a cluster.",
> e);
> }
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
