You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@fluss.apache.org by "swuferhong (via GitHub)" <gi...@apache.org> on 2025/09/10 01:39:46 UTC
[PR] [client] Fix the MetadataRequest sent by LogFetcher timeout exception when upgrading cluster [fluss]
swuferhong opened a new pull request, #1666:
URL: https://github.com/apache/fluss/pull/1666
<!--
*Thank you very much for contributing to Fluss - we are happy that you want to help us improve Fluss. To help the community review your contribution in the best possible way, please go through the checklist below, which will get the contribution into a shape in which it can be best reviewed.*
## Contribution Checklist
- Make sure that the pull request corresponds to a [GitHub issue](https://github.com/apache/fluss/issues). Exceptions are made for typos in JavaDoc or documentation files, which need no issue.
- Name the pull request in the format "[component] Title of the pull request", where *[component]* should be replaced by the name of the component being changed. Typically, this corresponds to the component label assigned to the issue (e.g., [kv], [log], [client], [flink]). Skip *[component]* if you are unsure about which is the best component.
- Fill out the template below to describe the changes contributed by the pull request. That will give reviewers the context they need to do the review.
- Make sure that the change passes the automated tests, i.e., `mvn clean verify` passes.
- Each pull request should address only one issue, not mix up code from multiple issues.
**(The sections below can be removed for hotfixes or typos)**
-->
### Purpose
<!-- Linking this pull request to the issue -->
Linked issue: close #1665
<!-- What is the purpose of the change -->
This pr is aims to Fix the TimeoutException of MetadataRequest sent by LogFetcher when upgrading cluster. The root cause of this issue is that both `LogFetcher#sendFetches()` and `LogFetcher#handleFetchLogResponse()` hold the synchronized lock simultaneously. Specifically:
1. `sendFetches()` initiates a metadata operation that sends a `MetadataRequest` while holding the lock.
2. `handleFetchLogResponse()` is triggered in the `whenComplete()` method of `FetchLogRequest`. The caller of `whenComplete()` – `ServerConnection.ResponseCallback#onRequestResult(int requestId, ApiMessage response)` – blocks the Netty client worker thread.
Deadlock Scenario:
- When `LogFetcher` triggers a `MetadataRequest`, it holds the synchronized lock of `LogFetcher`.
- Simultaneously, multiple `FetchLogResponses` return to `Netty client workers` (whose count is controlled by `client.netty.client.num-network-threads`).
- These responses invoke `ResponseCallback#onRequestResult`, which occupies the `Netty worker threads` and waits for the same `LogFetcher` lock.
- Result: Deadlock occurs because:
a) The `MetadataResponse` cannot acquire a `Netty worker thread` (blocked by `FetchLogResponse` processing)
b) `FetchLogResponse` processing waits indefinitely for the `LogFetcher` lock (held by `MetadataRequest`).
- Eventually, the `MetadataRequest` times out. After timeout, `ResponseCallback#onRequestResult` finally acquires the lock and recovers.
### Brief change log
<!-- Please describe the changes made in this pull request and explain how they address the issue -->
### Tests
<!-- List UT and IT cases to verify this change -->
### API and Format
<!-- Does this change affect API or storage format -->
### Documentation
<!-- Does this change introduce a new feature -->
--
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: issues-unsubscribe@fluss.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [client] Fix the MetadataRequest sent by LogFetcher timeout exception when upgrading cluster [fluss]
Posted by "wuchong (via GitHub)" <gi...@apache.org>.
wuchong commented on code in PR #1666:
URL: https://github.com/apache/fluss/pull/1666#discussion_r2348627118
##########
fluss-client/src/main/java/org/apache/fluss/client/table/scanner/log/LogFetcher.java:
##########
@@ -243,18 +273,21 @@ private TableOrPartitions getTableOrPartitionsInFetchRequest(FetchLogRequest fet
return new TableOrPartitions(tableIdsInFetchRequest, tablePartitionsInFetchRequest);
}
- private static class TableOrPartitions {
+ /** A helper class to hold table ids or table partitions. */
+ @VisibleForTesting
+ public static class TableOrPartitions {
private final @Nullable Set<Long> tableIds;
private final @Nullable Set<TablePartition> tablePartitions;
- private TableOrPartitions(
+ TableOrPartitions(
@Nullable Set<Long> tableIds, @Nullable Set<TablePartition> tablePartitions) {
this.tableIds = tableIds;
this.tablePartitions = tablePartitions;
}
}
- private void invalidTableOrPartitions(TableOrPartitions tableOrPartitions) {
+ @VisibleForTesting
+ public void invalidTableOrPartitions(TableOrPartitions tableOrPartitions) {
Review Comment:
Change them to package visible, it is only accessed in the same test class.
##########
fluss-client/src/main/java/org/apache/fluss/client/table/scanner/log/LogFetcher.java:
##########
@@ -243,18 +273,21 @@ private TableOrPartitions getTableOrPartitionsInFetchRequest(FetchLogRequest fet
return new TableOrPartitions(tableIdsInFetchRequest, tablePartitionsInFetchRequest);
}
- private static class TableOrPartitions {
+ /** A helper class to hold table ids or table partitions. */
+ @VisibleForTesting
+ public static class TableOrPartitions {
Review Comment:
Change them to package visible, it is only accessed in the same test class.
##########
fluss-client/src/test/java/org/apache/fluss/client/table/scanner/log/LogFetcherTest.java:
##########
@@ -183,6 +188,37 @@ void testFetchWhenDestinationIsNullInMetadata() throws Exception {
assertThat(records.get(tb0).size()).isEqualTo(10);
}
+ @Test
+ void testFetchWithInvalidTableOrPartitions() throws Exception {
+ MetadataUpdater metadataUpdater1 =
+ new MetadataUpdater(clientConf, FLUSS_CLUSTER_EXTENSION.getRpcClient());
+ logFetcher =
+ new LogFetcher(
+ DATA1_TABLE_INFO,
+ null,
+ logScannerStatus,
+ clientConf,
+ metadataUpdater1,
+ TestingScannerMetricGroup.newInstance(),
+ new RemoteFileDownloader(1));
+
+ ExecutorService executor = Executors.newSingleThreadExecutor();
+ Future<?> future =
+ executor.submit(
+ () -> {
+ // If this test blocked, please checking whether it was blocked with
+ // the same reason as https://github.com/apache/fluss/pull/1666
+ for (int i = 0; i < 1000; i++) {
+ logFetcher.sendFetches();
+ logFetcher.invalidTableOrPartitions(
+ new LogFetcher.TableOrPartitions(
+ Collections.singleton(tableId), null));
+ }
+ });
+
+ future.get(2, TimeUnit.MINUTES);
Review Comment:
assert future is done in the end, and close the executor
##########
fluss-client/src/main/java/org/apache/fluss/client/utils/MetadataUtils.java:
##########
@@ -63,9 +63,10 @@ public class MetadataUtils {
* add those table into cluster.
*/
public static Cluster sendMetadataRequestAndRebuildCluster(
- AdminReadOnlyGateway gateway, Set<TablePath> tablePaths)
+ AdminReadOnlyGateway gateway, Set<TablePath> tablePaths, long metadataRequestTimeoutMs)
Review Comment:
Please remove the timeout parameter, and use the default 30s in the test.
--
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: issues-unsubscribe@fluss.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [client] Fix the MetadataRequest sent by LogFetcher timeout exception when upgrading cluster [fluss]
Posted by "wuchong (via GitHub)" <gi...@apache.org>.
wuchong merged PR #1666:
URL: https://github.com/apache/fluss/pull/1666
--
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: issues-unsubscribe@fluss.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org