You may encounter a situation where health check stays in a down state even though it is not failing. Here is an example of what you may see in the sysinfo:
fwd.test Domain name: test DNS status: success Enabled Check failed DOWN IP address: x.x.x.x Enabled OK UP Last status: Success. Successes (total): 16481 (last): Fri, 16 Oct 2015 09:56:40 GMT (consecutive): 3216 Failures (total): 5 (last): Fri, 16 Oct 2015 01:00:33 GMT (consecutive): 0 (external): 0 Last response time: 3 ms Average response time: 6 ms Minimum response time: 1 ms Maximum response time: 132 ms
This all affects health checks created by forwarding hosts, such as those which are testing the TCP connectivity.
Due to a bug, existing IP level health checks aren't deleted during an upgrade, but rather are marked as "disabled and report sick".
The fix is available from 220.127.116.11. However, the issue still remains after upgrading from 6.5.7.(18.104.22.168) to 22.214.171.124 and higher, because the registry content still keeps the old container which should have been removed during a DNS update on an health check.
The workaround is to delete the affected health check and then recreate it.
Once you have created the new health check on 126.96.36.199 SGOS and above, the issue will not return.
Imported Document ID: 000028708
Subscribing will provide email updates when this Article is updated. Login is required.