I wrote a simple Spring Cloud Ribbon application, to call a REST service which was registered in Eureka.
But how to override the ribbon.serverListRefreshInterv
@codependent
After put below configuration in application.yml, it looked the configuration didnt take effect.
Compute-Service:
ribbon:
ServerListRefreshInterval: 1000
I observed that the instances list has been updated in Ribbon side (through eureka.client.registry-fetch-interval-seconds parameter), however Ribbon still point out a dead instance:
[tbeatExecutor-0] com.netflix.discovery.DiscoveryClient : DiscoveryClient_RIBBON-CONSUMER/192.168.1.101:Ribbon-Consumer:3333 - Heartbeat status: 200
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : Got delta update with apps hashcode UP_2_
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : Added instance 192.168.1.101:Ribbon-Consumer:3333 to the existing apps in region null
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : Deleted instance 192.168.1.101:Compute-Service:2222 to the existing apps
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : Added instance 192.168.1.101:Compute-Service:1111 to the existing apps in region null
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : The total number of instances fetched by the delta processor : 3
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : The total number of all instances in the client now is 2
[freshExecutor-0] com.netflix.discovery.DiscoveryClient : Completed cache refresh task for discovery. All Apps hash code is Local region apps hashcode: UP_2_, is fetching remote regions? false
[nio-3333-exec-3] o.a.c.c.C.[.[.[/].[dispatcherServlet] : Servlet.service() for servlet [dispatcherServlet] in context with path [] threw exception [Request processing failed; nested exception is org.springframework.web.client.ResourceAccessException: I/O error on GET request for "http://Compute-Service/add": Connection refused; nested exception is java.net.ConnectException: Connection refused] with root cause
192.168.1.101:Compute-Service:1111 was right service instance, while 192.168.1.101:Compute-Service:2222 was dead instance, obviously Ribbon still pointed to the dead instance, which meant Ribbon ServerList cache was not refreshed.
Try with:
myService.ribbon.ServerListRefreshInterval=10000
where myService
is the name of your destination microservice.
UPDATE:
After some source code digging I found out that LoadBalancerBuilder
calls:
@Deprecated
public ZoneAwareLoadBalancer(IClientConfig clientConfig, IRule rule,
IPing ping, ServerList<T> serverList, ServerListFilter<T> filter) {
super(clientConfig, rule, ping, serverList, filter);
}
whose super is:
@Deprecated
public DynamicServerListLoadBalancer(IClientConfig clientConfig, IRule rule, IPing ping,
ServerList<T> serverList, ServerListFilter<T> filter) {
this(
clientConfig,
rule,
ping,
serverList,
filter,
new PollingServerListUpdater()
);
}
Notice the PollingServerListUpdater
constructors:
private static int LISTOFSERVERS_CACHE_REPEAT_INTERVAL = 30 * 1000; // msecs;
public PollingServerListUpdater() {
this(LISTOFSERVERS_CACHE_UPDATE_DELAY, LISTOFSERVERS_CACHE_REPEAT_INTERVAL);
}
public PollingServerListUpdater(IClientConfig clientConfig) {
this(LISTOFSERVERS_CACHE_UPDATE_DELAY, getRefreshIntervalMs(clientConfig));
}
The second one would allow us to override the default refresh interval. However it's the first one that's called, so it ignores de property.
UPDATE 2:
There's an open issue about this: https://github.com/spring-cloud/spring-cloud-netflix/issues/1304