tzssangglass commented on a change in pull request #5206:
URL: https://github.com/apache/apisix/pull/5206#discussion_r727249565
##########
File path: t/core/config_etcd.t
##########
@@ -57,6 +57,11 @@ apisix:
etcd:
host:
- "https://127.0.0.1:2379";
+--- extra_init_by_lua
+local health_check = require("resty.etcd.health_check")
+health_check.get_target_status = function()
Review comment:
In the lua-resty-etcd v1.5.5 implementation, there is no passive
round-robin health check, so when apisix loads etcd configuration when apisix
starts,
here:https://github.com/apache/apisix/blob/c7523da7d45b674427205c435444133ff1aca258/apisix/init.lua#L84-L86
even if it get `failed to load the configuration: closed`, then sync data
from etcd in ngx.timer will still use `https://127.0.0.1:2379` as the etcd
endpoint and get the `handshake failed`.
but in lua-resty-etcd v1.6.0 implementation, the passive round-robin health
check is default policy, so when apisix loads etcd configuration when apisix
starts, `https://127.0.0.1:2379` is marked as unhealthy endpoint, and then sync
data from etcd in ngx.timer would got `has no healthy etcd endpoint available`
in `choose_endpoint` function, can not get the `handshake failed`.
bypass `get_target_status` to simulate the behaviour of the original test
case.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
