Installation of an FSA Node Was Interrupted During System Deployment, Causing the Further Creation of the Control Cluster Fails

Hi team, here's a new case.

Problem Symptom

During the system deployment on DeviceManager, an installation failure occurred. 

When a user creates a control cluster on DeciveManager later and selects physical disks to store metadata, 

the control cluster fails to be created, and a message is displayed, indicating that an internal system error occurs.

Problem Diagnosis

1. An installation failure occurred during the system deployment, for example, an FSA node was powered off unexpectedly.

2. The installation was then successful upon a retry.

3. When creating a control cluster on DeviceManager later, an internal system error is reported.

4. The RPM package of PlogManager has been installed but the ZooKeeper process has not been installed.

   Run the rpm -qa|grep plog-manager command. If any command output is displayed, PlogManager is installed. Otherwise, PlogManager is not installed.

   Run the ll /opt/fusionstorage/persistence_layer/agent/zk/bin/, command. If a file is displayed in the command output, the installation is successful. Otherwise, the installation fails.

Causes

• After an installation fails and is retried, the installation script cannot be re-executed. However, the installation is reported as successful.

• Others

Solution

1. Uninstall all software from the faulty FSA node and reinstall all the software packages.

   For details about how to uninstall the software, see Uninstalling OceanStor 100D Software in the OceanStor 100D 8.0.3 Product Documentation.

2. Create a control cluster again. If the creation is successful, the problem is solved.

Applicable Versions

OceanStor 100D 8.0.3