EDR: Redis service fails to start on clustered environment

Products

• Carbon Black EDR (formerly Cb Response)

Issue/Introduction

• Cluster script fails after starting the Datagrid service on all nodes. Master node (0) appears to be the root cause.
• Datagrid service debug.log shows the following error:

2020-01-06 13:43:46,073 - [INFO] - from com.hazelcast.nio.tcp.TcpIpConnection in hz._hzInstance_1_dev.IO.thread-in-2
[x.x.x.x]:5701 [dev] [3.9.3] Connection[id=243, /127.0.0.1:5701->/127.0.0.1:35220, endpoint=[127.0.0.1]:35220, alive=false, type=PYTHON_CLIENT] closed. Reason: Connection closed by the other side

2020-01-06 13:43:46,074 - [DEBUG] - from com.hazelcast.internal.cluster.ClusterService in hz._hzInstance_1_dev.event-1
[x.x.x.x]:5701 [dev] [3.9.3] Removed connection to [127.0.0.1]:35220

2020-01-06 13:43:46,074 - [INFO] - from com.hazelcast.client.ClientEndpointManager in hz._hzInstance_1_dev.event-1
[x.x.x.x.]:5701 [dev] [3.9.3] Destroying ClientEndpoint{connection=Connection[id=243, 0.0.0.0/0.0.0.0:5701->/127.0.0.1:35220, endpoint=[127.0.0.1]:35220, alive=false, type=PYTHON_CLIENT], principal='ClientPrincipal{uuid='0309394f-e8a4-4cbe-8ded-7182818969ad', ownerUuid='cbfe16ab-4fcb-46b6-aeea-a12c7686eef6'}, firstConnection=true, authenticated=true, clientVersion=3.11, creationTime=1578343424498, latest statistics=null}

Environment

• EDR (formerly CB Response) Server: All Versions
• Clustered

Cause

On all nodes, the /etc/cb/cb.conf files has an inaccurate value for 'RedisHost=<hostname>', which should point to the Master node.

Resolution

1. Edit /etc/cb/cb.conf
2. Change the following value to point to the Master node hostname:

RedisHost=<hostname>

Example:
RedisHost= Master.domain.local

3. Save configuration changes
4. Restart the cluster services