Health check
How does the health check function of Global Server Load Balance (GSLB) work?
GSLB checks the virtual server availability of backend servers by performing health checks. It sends out various probes according to the configuration and checks the response to determine the status of the virtual server. Virtual servers that respond successfully for the configured number of times are considered healthy and its IP address will be included in the DNS response. Virtual servers that fail to respond successfully for a certain number of times are determined to be unhealthy and its IP address will be not in the DNS query results.
How long does it take to get the virtual server status after initiating a health check on a virtual server?
Normally it takes no more than 1 minute for a newly configured health check to activate. After that, the time needed to get the result of the virtual server status is dependant on the health check configuration details including “Interval”, “Timeout”, “Up Retry” and “Down Retry”. For example, if the “Down Retry” value is 2 and the “Interval” value is 5 and the virtual server fails to respond, it will take about 10 seconds for the virtual server status to update to the down status. It will take about 0 to 5 seconds for the system to process the status so after the configuration is active, it will take about 15 seconds to get the virtual server status back.
Why are there default health checks?
The default health checks are configured for general usage and best practice. In the majority of cases, the default health checks can satisfy the user's need and additional health check are not needed.
Why is the health check result healthy even though the server is down?
There are several possibilities for this. If the health check is newly applied to the virtual server, it needs some time to activate and wait for the probe response before it can change the virtual server's status to down. If the number of active health checks exceeds the license limit, some of the health checks will stop probing.
Why is the health check result unhealthy even though the server is running normally?
The best way to find out why the health check is down is by capturing packages. In most of cases, the health check is down because there is no response to the probe packages. If there are return packages for DNS, HTTP and HTTPS health checks, you will need to check whether the content of the response is consistent with the configuration. For example, for the DNS health check, you will need to check whether the returned host IP address is same with the configuration.
How can I shorten the health check time and show the latest status more quickly?
There are four default health checks with parameter: interval 30, timeout 10, 1 time up retry, and 3 times down retry. These defaults cannot be changed. However, users can self-define health checks with shorter interval/timeout and retry times to decrease the health check time and see the status change quickly.
Does GSLB support IPv6 for health check?
No. GSLB does not currently support IPv6 type health check.