diff --git a/ch_tools/chadmin/cli/wait_group.py b/ch_tools/chadmin/cli/wait_group.py
index 5691446d..cd378248 100644
--- a/ch_tools/chadmin/cli/wait_group.py
+++ b/ch_tools/chadmin/cli/wait_group.py
@@ -1,9 +1,9 @@
 import os
 import sys
+import requests
 import time
 
 from click import FloatRange, group, option, pass_context
-from requests.exceptions import ReadTimeout
 
 from ch_tools.chadmin.cli.chadmin_group import Chadmin
 from ch_tools.chadmin.internal.clickhouse_disks import S3_METADATA_STORE_PATH
@@ -112,7 +112,7 @@ def wait_replication_sync_command(
                 timeout=timeout,
                 settings={"receive_timeout": timeout},
             )
-        except ReadTimeout:
+        except requests.exceptions.ReadTimeout:
             logging.error("Timeout while running SYNC REPLICA on {}.", full_name)
             sys.exit(1)
         except ClickhouseError as e:
@@ -120,6 +120,11 @@ def wait_replication_sync_command(
                 logging.error("Timeout while running SYNC REPLICA on {}.", full_name)
                 sys.exit(1)
             raise
+        except requests.exceptions.ConnectionError:
+            logging.error(
+                "Connection error while running SYNC REPLICA on {}.", full_name
+            )
+            sys.exit(1)
 
     # Replication lag
     while time.time() < deadline:
diff --git a/tests/features/chadmin.feature b/tests/features/chadmin.feature
index be762541..d6a37b33 100644
--- a/tests/features/chadmin.feature
+++ b/tests/features/chadmin.feature
@@ -153,8 +153,28 @@ Feature: chadmin commands.
     """
     When we execute command on clickhouse01
     """
+    supervisorctl stop clickhouse-server
+    """
+    And we sleep for 5 seconds
+    When we execute command on clickhouse01
+    """
+    chadmin wait replication-sync --total-timeout 10 --replica-timeout 3 -p 1 -w 4
+    """
+    Then it fails with response contains
+    """
+    Connection error while running SYNC REPLICA on
+    """
+    Then last command completed with error exit code on clickhouse01
+    When we execute command on clickhouse01
+    """
+    supervisorctl start clickhouse-server
+    """
+    And we sleep for 5 seconds
+    When we execute command on clickhouse01
+    """
     chadmin wait replication-sync --total-timeout 10 --replica-timeout 3 -p 1 -w 4
     """
+    Then last command completed with success exit code on clickhouse01
 
   Scenario Outline: Check replica restore (<replicas_count> replicas, <workers> workers) 
     Given populated clickhouse with <replicas_count> replicated tables on clickhouse01 with db database and table_ prefix
diff --git a/tests/steps/common.py b/tests/steps/common.py
index 4413fe19..d64b212d 100644
--- a/tests/steps/common.py
+++ b/tests/steps/common.py
@@ -115,3 +115,13 @@ def working_http(context):
     host, port = docker.get_exposed_port(container, 8080)
     response = requests.get(f"http://{host}:{port}/")
     assert response.text == "OK", f'expected "OK", got "{response.text}"'
+
+
+@then("last command completed with {result} exit code on {node:w}")
+def step_last_command_completed_with_exit_code(context, result, node):
+    container = docker.get_container(context, node)
+    result = container.exec_run(["echo", "$?"], user="root")
+    context.response = result.output.decode().strip()
+    context.exit_code = result.exit_code
+    exit_code = 0 if result == "success" else 1
+    assert_that(context.exit_code, equal_to(exit_code))