zhangyue19921010 commented on code in PR #6133:
URL: https://github.com/apache/hudi/pull/6133#discussion_r1028255753
##########
hudi-timeline-service/src/main/java/org/apache/hudi/timeline/service/handlers/MarkerHandler.java:
##########
@@ -150,7 +168,50 @@ public boolean doesMarkerDirExist(String markerDir) {
* @param markerName marker name
* @return the {@code CompletableFuture} instance for the request
*/
- public CompletableFuture<String> createMarker(Context context, String
markerDir, String markerName) {
+ public CompletableFuture<String> createMarker(Context context, String
markerDir, String markerName,
+ String batchInterval, String
period, String maxAllowableHeartbeatIntervalInMs,
+ String basePath, String
earlyConflictDetectionEnable,
+ String
earlyConflictDetectionClassName) {
+ // Step1 do early conflict detection if enable
+ if (Boolean.parseBoolean(earlyConflictDetectionEnable)) {
+ try {
+ synchronized (earlyConflictDetectionLock) {
+ if (earlyConflictDetectionStrategy == null) {
+ earlyConflictDetectionStrategy =
ReflectionUtils.loadClass(earlyConflictDetectionClassName);
+ }
+
+ if (!markerDir.equalsIgnoreCase(currentMarkerDir)) {
+ this.currentMarkerDir = markerDir;
+ Set<String> actions = CollectionUtils.createSet(COMMIT_ACTION,
DELTA_COMMIT_ACTION, REPLACE_COMMIT_ACTION);
+ Set<HoodieInstant> oldInstants =
viewManager.getFileSystemView(basePath)
+ .getTimeline()
+ .filterCompletedInstants()
+ .filter(instant -> actions.contains(instant.getAction()))
+ .getInstants()
+ .collect(Collectors.toSet());
+
+ earlyConflictDetectionStrategy.fresh(batchInterval, period,
markerDir, basePath, maxAllowableHeartbeatIntervalInMs, fileSystem,
+ this, oldInstants);
+ }
+ }
+
+ if (earlyConflictDetectionStrategy.hasMarkerConflict()) {
+ earlyConflictDetectionStrategy.resolveMarkerConflict(basePath,
markerDir, markerName);
+ }
Review Comment:
As we discuss before, checker is doing the detection in a batching and async
way. So that it's not related to specific batches requests. For example
timeline service is deal with marker1, marker2 and marker3. But this checker
find a conflict on marker0.
So could we do this check and get check result at api level? what do you
think.
##########
hudi-timeline-service/src/main/java/org/apache/hudi/timeline/service/handlers/MarkerHandler.java:
##########
@@ -150,7 +168,50 @@ public boolean doesMarkerDirExist(String markerDir) {
* @param markerName marker name
* @return the {@code CompletableFuture} instance for the request
*/
- public CompletableFuture<String> createMarker(Context context, String
markerDir, String markerName) {
+ public CompletableFuture<String> createMarker(Context context, String
markerDir, String markerName,
+ String batchInterval, String
period, String maxAllowableHeartbeatIntervalInMs,
+ String basePath, String
earlyConflictDetectionEnable,
+ String
earlyConflictDetectionClassName) {
+ // Step1 do early conflict detection if enable
+ if (Boolean.parseBoolean(earlyConflictDetectionEnable)) {
+ try {
+ synchronized (earlyConflictDetectionLock) {
+ if (earlyConflictDetectionStrategy == null) {
+ earlyConflictDetectionStrategy =
ReflectionUtils.loadClass(earlyConflictDetectionClassName);
+ }
+
+ if (!markerDir.equalsIgnoreCase(currentMarkerDir)) {
+ this.currentMarkerDir = markerDir;
+ Set<String> actions = CollectionUtils.createSet(COMMIT_ACTION,
DELTA_COMMIT_ACTION, REPLACE_COMMIT_ACTION);
+ Set<HoodieInstant> oldInstants =
viewManager.getFileSystemView(basePath)
+ .getTimeline()
+ .filterCompletedInstants()
+ .filter(instant -> actions.contains(instant.getAction()))
+ .getInstants()
+ .collect(Collectors.toSet());
+
+ earlyConflictDetectionStrategy.fresh(batchInterval, period,
markerDir, basePath, maxAllowableHeartbeatIntervalInMs, fileSystem,
+ this, oldInstants);
+ }
+ }
+
+ if (earlyConflictDetectionStrategy.hasMarkerConflict()) {
+ earlyConflictDetectionStrategy.resolveMarkerConflict(basePath,
markerDir, markerName);
+ }
Review Comment:
As we discuss before, checker is doing the detection in a batching and async
way. So that it's not related to specific batches requests. For example
timeline service is deal with marker1, marker2 and marker3. But this checker
find a conflict on marker0.
So could we do this check and get check result at api level? what do you
think :)
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
