The Elk cluster on the FusionInsight platform is abnormal. An error log is displayed, indicating that the Elk fails to be started.
Perform the following operations to check the cluster status:
Step 1: Log in to any server on which the Elk service is running properly as user omm.
Step 2: Run the following command to enable environment variables:
Step 3: Run the following command to check the status of the Elk cluster.
cm_ctl query -v -C
The preceding figure shows the nodes 15353 and 15356 are abnormal. Collect Elk logs on the two nodes.
Lots of messages shown in the following is printed in the /var/log/Bigdata/omm/cm/cm_agent/system_call-2017-10-09_082951-current.log file:
2017-10-12 13:57:12.479 CST 140208237779488 0 LOG: could not translate host name "localhost", service "25108" to address: Name or service not known
2017-10-12 13:57:12.479 CST 140208237779488 0 FATAL: could not create listen socket for "localhost"
Thu Oct 12 13:57:17 CST 2017
2017-10-12 13:57:17.591 CST 140047661976096 0 LOG: [Alarm Module]Host IP: 10.xx.xx.53
2017-10-12 13:57:17.629 CST 140047661976096 0 LOG: could not translate host name "localhost", service "25108" to address: Name or service not known
2017-10-12 13:57:17.629 CST 140047661976096 0 FATAL: could not create listen socket for "localhost"
Thu Oct 12 13:57:22 CST 2017
2017-10-12 13:57:22.879 CST 140133872257568 0 LOG: [Alarm Module]Host IP: 10.xx.xx.53
This error is caused by failing to parse the host name of the local host. It is found that the original IP address 127.0.0.1 of the local host is commented out manually in the hosts file. As a result, the parsing fails.
# 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
The localhost item in the /etc/hosts file is commented out. As a result, the host name of the local host cannot be parsed and the Elk service is abnormal.
Clear the localhost item that has been commented out in the cat /etc/hosts file of the abnormal node, the Elk service automatically becomes normal.