[BugFix] Fix transactional stream load with warehouse property not work rightly #56464
+46
−22
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…rk rightly Signed-off-by: drake_wang <[email protected]>
![starrocks-cr[bot]](https://avatars.githubusercontent.com/in/404334?s=60&v=4){kind=small}
| } | ||
| throw new StarRocksException("No compute node alive in warehouse: " + warehouseId); | ||
| } | ||
|
|
||
| public Long seqChooseBackendOrComputeId() throws StarRocksException { | ||
| List<Long> backendIds = seqChooseBackendIds(1, true, false, null); | ||
| if (CollectionUtils.isNotEmpty(backendIds)) { |
There was a problem hiding this comment.
The most risky bug in this code is:
A null pointer exception could be thrown if GlobalStateMgr.getCurrentState().getWarehouseMgr().getWarehouse(warehouseId) returns null.
You can modify the code like this:
public Long seqChooseComputeIdFromWarehouse(long warehouseId) throws StarRocksException {
Warehouse warehouse = GlobalStateMgr.getCurrentState().getWarehouseMgr().getWarehouse(warehouseId);
if (warehouse == null) {
throw new StarRocksException("Warehouse not found for ID: " + warehouseId);
}
List<ComputeNode> aliveComputeNodes =
GlobalStateMgr.getCurrentState().getWarehouseMgr().getAliveComputeNodes(warehouseId);
if (CollectionUtils.isNotEmpty(aliveComputeNodes)) {
List<Long> computeNodes = seqChooseNodeIds(1, false, null, aliveComputeNodes);
if (CollectionUtils.isNotEmpty(computeNodes)) {
return computeNodes.get(0);
}
}
throw new StarRocksException("No compute node alive in warehouse: " + warehouseId);
}
| // txnNodeMap is LRU cache, it atomic remove unused entry | ||
| accessTxnNodeMapWithWriteLock(txnNodeMap -> txnNodeMap.put(label, chosenNodeId)); | ||
| accessTxnNodeMapWithWriteLock(txnNodeMap -> txnNodeMap.put(label, nodeId)); | ||
| } else { | ||
| nodeId = accessTxnNodeMapWithReadLock(txnNodeMap -> txnNodeMap.get(label)); | ||
| } |
There was a problem hiding this comment.
The most risky bug in this code is:
Potential NullPointerException when warehouseName is empty.
You can modify the code like this:
private Long getNodeId(TransactionOperation txnOperation, String label, String warehouseName) throws StarRocksException {
Long nodeId = null; // Initialize to null for safety
// save label->be hashmap when begin transaction, so that subsequent operator can send to same BE
if (TXN_BEGIN.equals(txnOperation)) {
if (StringUtils.isNotEmpty(warehouseName)) {
Warehouse warehouse = GlobalStateMgr.getCurrentState().getWarehouseMgr().getWarehouse(warehouseName);
if (warehouse != null) { // Check if the warehouse exists
nodeId = GlobalStateMgr.getCurrentState().getNodeMgr()
.getClusterInfo().getNodeSelector().seqChooseComputeIdFromWarehouse(warehouse.getId());
}
// Handle case where warehouse is not found
if (nodeId == null) {
throw new StarRocksException("Warehouse " + warehouseName + " is not valid");
}
}
if (nodeId == null) { // If nodeId was not set, fall back to default behavior
nodeId = GlobalStateMgr.getCurrentState().getNodeMgr()
.getClusterInfo().getNodeSelector().seqChooseBackendOrComputeId();
}
// txnNodeMap is LRU cache, it atomic remove unused entry
accessTxnNodeMapWithWriteLock(txnNodeMap -> txnNodeMap.put(label, nodeId));
} else {
nodeId = accessTxnNodeMapWithReadLock(txnNodeMap -> txnNodeMap.get(label));
}
if (nodeId == null) {
throw new StarRocksException("Could not obtain a valid node ID for label: " + label);
}
return nodeId;
}
|
[Java-Extensions Incremental Coverage Report]✅ pass : 0 / 0 (0%) |
[BE Incremental Coverage Report]✅ pass : 0 / 0 (0%) |
kevincai
reviewed
Mar 4, 2025
| return computeNodes.get(0); | ||
| } | ||
| } | ||
| throw new StarRocksException("No compute node alive in warehouse: " + warehouseId); |
There was a problem hiding this comment.
Suggested change
| throw new StarRocksException("No compute node alive in warehouse: " + warehouseId); | |
| throw new StarRocksException("No compute node alive in warehouse: " + warehouse.getName()); |
| @@ -250,7 +250,6 @@ public static Optional<Long> getWarehouseIdByNodeId(SystemInfoService systemInfo | |||
| LOG.warn("failed to get warehouse id by node id: {}", nodeId); | |||
| return Optional.empty(); | |||
| } | |||
|
|
|||
| @@ -49,7 +49,8 @@ public ResultWrapper handle(BaseRequest request, BaseResponse response) throws S | |||
| Long timeoutMillis = txnOperationParams.getTimeoutMillis(); | |||
| String label = txnOperationParams.getLabel(); | |||
| Channel channel = txnOperationParams.getChannel(); | |||
| LOG.info("Handle transaction with channel info, label: {}", label); | |||
| LOG.info("Handle transaction with channel info, label: {}, warehouse: {}", label, | |||
There was a problem hiding this comment.
this log may not be accurate, because only the begin txn will take the warehouse name from parameter, subsequent ops are not using the warehousename parameter at all.
banmoy
reviewed
Mar 4, 2025
| Long chosenNodeId = GlobalStateMgr.getCurrentState().getNodeMgr() | ||
| .getClusterInfo().getNodeSelector().seqChooseBackendOrComputeId(); | ||
| nodeId = chosenNodeId; | ||
| if (StringUtils.isNotEmpty(warehouseName)) { |
There was a problem hiding this comment.
we can provide a utility function to unify the same logic in LoadAction
|
should add some tests |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Why I'm doing:
What I'm doing:
Fixes #issue
What type of PR is this:
Does this PR entail a change in behavior?
If yes, please specify the type of change:
Checklist:
Bugfix cherry-pick branch check: