How to replace a blade on a FortiAnalyzer-BigData appliance
This section contains instructions on how to gracefully remove and replace a malfunctioning hardware blade running one of the hosts in an active system. In order to allow the high availability mechanism to take effect, only one blade can be decommissioned at a time.
Finding a blade's location A blade’s host name uses the following naming convention:
For example, in a blade named blade-198-18-1-3, the number 1 represents the |
To replace a blade:
Before proceeding with the steps below, your version of FortiAnalyzer-BigData bootloader must match your current version of FortiAnalyzer-BigData. Check the version of your bootloader and upgrade it to match your FortiAnalyzer-BigData as needed. To check the bootloader version, run the following command from the Security Event Manager Controller: To upgrade the bootloader, see Upgrade Bootloader. |
- Power off the malfunctioned blade, and then remove the blade from the chassis.
- From Security Event Manager Controller run the following command to enable installation of system software on the new blade:
fazbdctl decommission <host name or blade IP of the blade to replace>
- Insert the replacement blade, and power it on.
- Monitor the status of the blade from CMM. See Remotely control blades via CMM in the FortiAnalyzer-BigData Getting Started Guide for your appliance: 4500F or 4500G.
Wait until the host OS completes booting. This may take 5-10 minutes.
- After the host boots up and joins the cluster, it will appear in Cluster Manager > Hosts web GUI.
- Go to Cluster Manager > Hosts and click Assign Role to recover the host. The new host will have a pending status.
- Wait for the Assign job to complete for all services to become healthy.
- Follow the Found New recommended configurations notification in the Cluster Manager > Hosts view (if it appears) to apply the optimized configuration for the newly added hosts.
- (Recommended) After the Assign Role job is complete, rebalance the data across the cluster. See How to rebalance the data.