imbajin commented on code in PR #814:
URL:
https://github.com/apache/hugegraph-toolchain/pull/814#discussion_r4274494768
##########
hugegraph-hubble/hubble-be/src/main/java/org/apache/hugegraph/service/load/LoadTaskService.java:
##########
@@ -350,9 +398,7 @@ public LoadTask resume(int taskId, String token) {
task.getOptions().incrementalMode = true;
task.reconnect(token);
task.setStatus(LoadStatus.RUNNING);
- this.update(task);
- this.taskExecutor.execute(task, () -> this.update(task));
- this.runningTaskContainer.put(taskId, task);
+ this.updateAndExecute(task);
Review Comment:
Fixed in the updated head. Embedded Loaders now take an immutable
progress-persistence policy: Hubble's initial and reconnect constructors select
false, so normal cleanup and the JVM hook close all resources without writing
parsed offsets. Existing standalone constructors retain their previous
behavior, and old checkpoints are preserved.
The actual CSV/parser/native semaphore regression pauses with the final
batch waiting for a permit, then resumes through the task's native reconnect
path. The predecessor leaves two rows missing; the fix inserts every row. A
separate actual single-insert failure/Resume control similarly changes from one
missing row to complete data. These controls use a controlled HTTP insertion
endpoint; they are not a real-Server/UI acceptance claim. The README explains
replay and Stop/Retry when an older checkpoint may be unsafe. Independent Java
17 unit/compile/style checks and three final source reviews pass.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
