How to reset a single host

This section contains instructions on how to gracefully reset a software malfunctioned Security Event Manager host in a running system. In order to allow the high availability mechanism to take effect, only one host can be reset at a time.

Finding a blade's location A blade’s host name is follows a naming convention: blade-10-0-{chass ID}-{blade ID}. A blade named "blade-10-0-1-3" means that "1" represents the chassis ID and the "3" represents the blade ID. Therefore, the blade is the third blade to the left on the first chassis. The internal IP of the blade is 10.0.1.3.

There are three types of host roles: Master Node, MetaStore Node and Data Node (see Roles). You can find the role type of a host in from Cluster Manager > Hosts. Some role types requires a different method for resetting.

• Reset a blade with a the Data Node role
• Reset a blade with the Master Node or MetaStore Node role

To reset a host with Data Node Role:

1. Access the Security Event Manager Controller (see To connect to the Security Event Manager Controller:) and run the following command to decommission the host by its IP address:
   fazbdctl -c delete -h {member_ip_addr}
2. Access the bootloader of the malfunctioned host (see Bootloader), enter the Reset OS option and wait until it finishes rebooting.
3. Reconnect to the Security Event Manager Controller and run the following command to ensure that the host has joined the cluster:
   fazbdctl -c show -t members
4. After the host status changes to "Joined", the host will show up in Cluster Manager > Hosts.
5. From the Hosts page, click Assign Role to add the host.
   The newly added host should have a "new" label.
6. Once the “Assign Role” job finishes running, the host soft reset is done.

To reset a host with Master Node or MetaStore Node role:

1. Access the bootloader of the malfunctioned host (see Bootloader), enter the Reset OS option and wait until it finishes rebooting.
2. Access the Security Event Manager Controller (see To connect to the Security Event Manager Controller:) and run the following command to ensure that the host has joined the cluster.

   fazbdctl -c show -t members

3. After the host status changes to "Joined", run follow command to soft reset FortiAnalyzer-BigData (see Soft reset FortiAnalyzer-BigData):
   fazbdctl -c reset -h cluster
4. Once the reset is finishes running, the host soft reset is done.