1.1.1 [Alarm] HBase Repeated Reporting Service Available Alarm
Problem phenomena: HBase reporting service is unavailable alarm, the reason for alarm is no active instance, and the alarm is continuously cleared and repeated after reporting.
Cause of the problem: When HBase configures serviceHealthCheck. properties, it should configure this category to
The <role name="HMaster"> configuration item is not configured to the entire service.
Ideas for problem positioning:
1. To check the reason for the alarm is no active instance, first check whether the main and standby state of HMaster on the interface is normal, there is only one main instance.
2. If normal, check the log file and add a line of log printing to the POC version.
LOG. info ("ready raise alarm in service health management: role ={}", managedrole. getName ());
The actual log is ready raise alarm in service health management: region server
3. It can be judged that the current situation is based on the region server to determine whether an alarm needs to be sent, HBase configuration.
ServiceCheckType is active, that is, it detects the primary status of key roles to determine whether to send an alarm.
(The key role is acquired by getting service check whether the role is configured or not). If the service HealthCheck. properties should be configured to the entire service, then both Regionserver instances and HMaster instances will have service check configuration, so both can be key roles. Which one will be used first in the traversal process? One is to determine the service status.
Terms of settlement:
When configuring service HealthCheck. properties with HBase, you should configure this category to
<role name="HMaster"> under configuration
