(phoenix) branch 5.3 updated: PHOENIX-7493 Addendum Only making HAGroupStoreClient unhealthy on CONNECTION_LOST (#2317)

[jisaac]

This is an automated email from the ASF dual-hosted git repository.

jisaac pushed a commit to branch 5.3
in repository https://gitbox.apache.org/repos/asf/phoenix.git


The following commit(s) were added to refs/heads/5.3 by this push:
     new 39b6471969 PHOENIX-7493 Addendum Only making HAGroupStoreClient 
unhealthy on CONNECTION_LOST (#2317)
39b6471969 is described below

commit 39b64719697f46a119e0ea816e67d446f8625e18
Author: ritegarg <58840065+riteg...@users.noreply.github.com>
AuthorDate: Thu Nov 13 16:29:10 2025 -0800

    PHOENIX-7493 Addendum Only making HAGroupStoreClient unhealthy on 
CONNECTION_LOST (#2317)
    
    Co-authored-by: Ritesh Garg 
<ritesh.g...@riteshg-ltmd34g.internal.salesforce.com>
---
 .../main/java/org/apache/phoenix/jdbc/HAGroupStoreClient.java |  7 ++++++-
 .../it/java/org/apache/phoenix/jdbc/HAGroupStoreClientIT.java | 11 ++++++++++-
 2 files changed, 16 insertions(+), 2 deletions(-)

diff --git 
a/phoenix-core-client/src/main/java/org/apache/phoenix/jdbc/HAGroupStoreClient.java
 
b/phoenix-core-client/src/main/java/org/apache/phoenix/jdbc/HAGroupStoreClient.java
index 43bd24048b..93102d79ac 100644
--- 
a/phoenix-core-client/src/main/java/org/apache/phoenix/jdbc/HAGroupStoreClient.java
+++ 
b/phoenix-core-client/src/main/java/org/apache/phoenix/jdbc/HAGroupStoreClient.java
@@ -124,11 +124,16 @@ public class HAGroupStoreClient implements Closeable {
             case INITIALIZED:
               latch.countDown();
               break;
-            case CONNECTION_LOST:
             case CONNECTION_SUSPENDED:
+              LOGGER.warn("Received CONNECTION_SUSPENDED event, will wait 
until either "
+                + "ZK connection is restored or session timeout occurs");
+              break;
+            case CONNECTION_LOST:
+              LOGGER.error("Received CONNECTION_LOST event, marking cache as 
unhealthy");
               isHealthy = false;
               break;
             case CONNECTION_RECONNECTED:
+              LOGGER.info("Received CONNECTION_RECONNECTED event, marking 
cache as healthy");
               isHealthy = true;
               break;
             default:
diff --git 
a/phoenix-core/src/it/java/org/apache/phoenix/jdbc/HAGroupStoreClientIT.java 
b/phoenix-core/src/it/java/org/apache/phoenix/jdbc/HAGroupStoreClientIT.java
index d591a0769a..d5b1406c1a 100644
--- a/phoenix-core/src/it/java/org/apache/phoenix/jdbc/HAGroupStoreClientIT.java
+++ b/phoenix-core/src/it/java/org/apache/phoenix/jdbc/HAGroupStoreClientIT.java
@@ -17,6 +17,8 @@
  */
 package org.apache.phoenix.jdbc;
 
+import static 
org.apache.phoenix.jdbc.HighAvailabilityGroup.PHOENIX_HA_ZK_SESSION_TIMEOUT_MS_DEFAULT;
+import static 
org.apache.phoenix.jdbc.HighAvailabilityGroup.PHOENIX_HA_ZK_SESSION_TIMEOUT_MS_KEY;
 import static org.apache.phoenix.jdbc.PhoenixHAAdmin.toPath;
 import static org.junit.Assert.assertThrows;
 
@@ -336,7 +338,14 @@ public class HAGroupStoreClientIT extends BaseTest {
     // Shutdown the ZK Cluster to simulate CONNECTION_SUSPENDED event
     utility.shutdownMiniZKCluster();
 
-    Thread.sleep(ZK_CURATOR_EVENT_PROPAGATION_TIMEOUT_MS);
+    // Check that immediately after ZK is down, the connection state
+    // should be SUSPENDED (and not LOST), so no exception should be thrown
+    assert 
haGroupStoreClient.getCRRsByClusterRole(ClusterRoleRecord.ClusterRole.ACTIVE).size()
+        == 2;
+
+    long sessionTimeout = config.getLong(PHOENIX_HA_ZK_SESSION_TIMEOUT_MS_KEY,
+      PHOENIX_HA_ZK_SESSION_TIMEOUT_MS_DEFAULT);
+    Thread.sleep(sessionTimeout + ZK_CURATOR_EVENT_PROPAGATION_TIMEOUT_MS);
     // Check that HAGroupStoreClient instance is not healthy and throws 
IOException
     assertThrows(IOException.class,
       () -> 
haGroupStoreClient.getCRRsByClusterRole(ClusterRoleRecord.ClusterRole.ACTIVE));