From 0bf9647047ae8bb8962ae8470c964492a04aab3f Mon Sep 17 00:00:00 2001
From: Siyang Tang <82279870+TangSiyang2001@users.noreply.github.com>
Date: Thu, 10 Oct 2024 22:20:47 +0800
Subject: [PATCH] [fix](delete) Fix potential delete job stuck util timeout if
 exception happend in FE DeleteJob execution (#41672)

## Proposed changes

Fail task should also count down for the count down latch to prevent job
stuck.
---
 .../src/main/java/org/apache/doris/master/MasterImpl.java      | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fe/fe-core/src/main/java/org/apache/doris/master/MasterImpl.java b/fe/fe-core/src/main/java/org/apache/doris/master/MasterImpl.java
index 3e63a5421f798f..27469301e17615 100644
--- a/fe/fe-core/src/main/java/org/apache/doris/master/MasterImpl.java
+++ b/fe/fe-core/src/main/java/org/apache/doris/master/MasterImpl.java
@@ -423,6 +423,9 @@ private void finishRealtimePush(AgentTask task, TFinishTaskRequest request) thro
         } catch (MetaNotFoundException e) {
             AgentTaskQueue.removeTask(backendId, TTaskType.REALTIME_PUSH, signature);
             LOG.warn("finish push replica error", e);
+            if (pushTask.getPushType() == TPushType.DELETE) {
+                pushTask.countDownLatch(backendId, pushTabletId);
+            }
         } finally {
             olapTable.writeUnlock();
         }