Help Center/ MapReduce Service/ Component Operation Guide (LTS)/ Using Doris/ Common Issues About Doris/ How Do I Handle the FE Startup Failure?
Updated on 2024-10-25 GMT+08:00
View PDF
Share

How Do I Handle the FE Startup Failure?

Symptom

The FE instance failed to be started, and the /var/log/Bigdata/doris/fe/fe.log file kept showing the following error message:

wait catalog to be ready. FE type UNKNOWN

Cause Analysis

  • The FE installation node has multiple network adapter IP addresses. The priority_network parameter is incorrectly set. As a result, an incorrect IP address is matched during FE startup.
  • Most follower FE nodes in the cluster are not started. For example, there are three follower FE nodes and only one follower FE node is started.

Procedure

  1. Log in to FusionInsight Manager and choose Cluster > Services > Doris, and click Configurations.
  2. Search for the priority_network parameter and set it correctly for the FE service. You can view the IP address of the network adapter bound to the FE node by checking the CURRENT_INSTANCE_IP variable in FE installation directory/FusionInsight_Doris_*/1_*_FE/etc/ENV_VARS.

    The priority_network parameter is used to help the system select the correct IP address of the network adapter as the IP address of the FE or BE. You are advised to set this parameter explicitly in any case to prevent incorrect IP address selection after network adapters are added. The value of priority_network is in CIDR format and is used to ensure that all nodes can use the same configuration value. The parameter value consists of two parts. The first part is the IP address in dotted decimal notation, and the second part is the prefix length.

    For example, 10.168.1.0/8 matches all 10.xx.xx.xx IP addresses, and 10.168.1.0/16 matches all 10.168.xx.xx IP addresses. If there are two nodes: 10.168.10.1 and 10.168.10.2, 10.168.10.0/24 can be the value of priority_network.

  3. Click Instance, select the follower FE to be started, and click Start Instance. For example, if there are three followers and only one follower is started, you need to start at least one FE so that a master node can be elected from the FE election group to provide services.
  4. If the FE still fails to be started, contact O&M engineers to rectify the fault.