You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@ignite.apache.org by "Cyrill (via GitHub)" <gi...@apache.org> on 2025/12/19 17:08:40 UTC
Re: [PR] IGNITE-27255 Fixed TxIdMismatchException due to write attempt by RW transaction after write intent resolution [ignite-3]
Cyrill commented on code in PR #7263:
URL: https://github.com/apache/ignite-3/pull/7263#discussion_r2635752604
##########
modules/partition-replicator/src/main/java/org/apache/ignite/internal/partition/replicator/handlers/TxStateCommitPartitionReplicaRequestHandler.java:
##########
@@ -106,19 +132,53 @@ private CompletableFuture<TransactionMeta> triggerTxRecoveryOnTxStateResolutionI
TxMeta txMeta = txStatePartitionStorage.get(txId);
if (txMeta == null) {
+ InternalClusterNode coordinator = (txStateMeta == null || txStateMeta.txCoordinatorId() == null)
+ ? null
+ : clusterNodeResolver.getById(txStateMeta.txCoordinatorId());
+
// This means the transaction is pending and we should trigger the recovery if there is no tx coordinator in topology.
if (txStateMeta == null
|| txStateMeta.txState() == ABANDONED
|| txStateMeta.txCoordinatorId() == null
- || clusterNodeResolver.getById(txStateMeta.txCoordinatorId()) == null) {
+ || coordinator == null) {
// This means that primary replica for commit partition has changed, since the local node doesn't have the volatile tx
// state; and there is no final tx state in txStateStorage, or the tx coordinator left the cluster. But we can assume
// that as the coordinator (or information about it) is missing, there is no need to wait a finish request from
// tx coordinator, the transaction can't be committed at all.
return txRecoveryEngine.triggerTxRecovery(txId, localNode.id())
.handle((v, ex) ->
CompletableFuture.<TransactionMeta>completedFuture(txManager.stateMeta(txId)))
- .thenCompose(v -> v);
+ .thenCompose(Function.identity());
+ } else if (coordinator != null) {
+ // If there is coordinator in the cluster we should fallback to coordinator request. It's possible that coordinator
+ // was not seen in topology on another node which requested the state from commit partition, but can be seen here.
+ HybridTimestamp timestamp = readTimestamp == null ? HybridTimestamp.MIN_VALUE : readTimestamp;
+
+ return txMessageSender.resolveTxStateFromCoordinator(
+ coordinator,
+ txId,
+ timestamp,
+ senderCurrentConsistencyToken,
+ senderGroupId
+ )
+ .handle((response, e) -> {
+ if (e == null && response.txStateMeta() != null) {
+ TransactionMetaMessage transactionMetaMessage = Objects.requireNonNull(response.txStateMeta(),
Review Comment:
requireNonNull seems excessive here
--
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: notifications-unsubscribe@ignite.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org