OkHttpClient client = new OkHttpClient.Builder()
.connectionPool(new ConnectionPool(maxIdleConnections,
keepAliveDuration, TimeUnit))
.build();
I can set the connection pool parameters with the code above，How can I set The Keepalive timeout duration is different for different addresses.
Related
I am building a distributed workflow orchestrator, grpc is used to communicate with the server cluster by workers.If a new server is added to the server grpc client is not able to detect this change. However i have done a workaround by adding a max connection age to the server options
grpc.KeepaliveParams(keepalive.ServerParameters{
MaxConnectionAge: time.Minute * 1,
})
We have two implementation of workers, one in golang and other in java this workaround works perfectly in golang client. Every minute the client makes new connection and is able to detect new servers in cluster. But this is not working with java client.
public CustomNameResolverFactory(String host, int port) {
ManagedChannel managedChannel = NettyChannelBuilder
.forAddress(host, port)
.withOption( ChannelOption.CONNECT_TIMEOUT_MILLIS, 10000 )
.usePlaintext().build();
GetServersRequest request = GetServersRequest.newBuilder().build();
GetServersResponse servers = TaskServiceGrpc.newBlockingStub(managedChannel).getServers(request);
List<Server> serversList = servers.getServersList();
System.out.println(servers);
LOGGER.info("found servers {}", servers);
for (Server server : serversList) {
String rpcAddr = server.getRpcAddr();
String[] split = rpcAddr.split(":");
String hostName = split[0];
int portN = Integer.parseInt(split[1]);
addresses.add(new EquivalentAddressGroup(new InetSocketAddress(hostName, portN)));
}
}
Java client code- https://github.com/Mohitkumar/orchy-worker-java/blob/master/src/main/java/com/orchy/client/CustomNameResolverFactory.java
Golang client code- https://github.com/Mohitkumar/orchy/blob/main/worker/lb/resolver.go
We have 2 different instances of spring boot Webclients. We want one of them to log all requests/responses and the other to log nothing. Settiing:
logging.level.reactor.netty.http.client=debug
in the application.properties file causes both instances to log requests/responses. Is there a way to programmatically set the log level on one of the instances to log debug levels and the other to not?
We create the first webclient like this:
WebClient.create();
For this one we don't want debug logging.
The second one, for which we want logging, we create like this:
return WebClient.builder()
.clientConnector(createWebClientWithTimeout())
// Add base url to all requests (callers only need to add the path and query params)
.baseUrl(baseUrl)
// Increase the buffer size from 256K to 1M
.codecs(configurer -> configurer
.defaultCodecs()
.maxInMemorySize(1024 * 1024))
// Filter to add bearer token to auth header before sending request
.filter((request, next) -> getToken(custId).map(setAuthHeader(request)).flatMap(next::exchange))
// Filter to send the request, and try again if it has an auth error
.filter((request, next) -> next.exchange(request).flatMap(clientResponse -> {
:::
return Mono.just(clientResponse);
})).build();
and
private ClientHttpConnector createWebClientWithTimeout() {
// create reactor netty HTTP client
HttpClient httpClient = HttpClient.newConnection()
.wiretap("reactor.netty.http.client.HttpClient",
LogLevel.DEBUG, AdvancedByteBufFormat.TEXTUAL)
.compress(true)
.tcpConfiguration(tcpClient -> {
tcpClient = tcpClient.option(ChannelOption.CONNECT_TIMEOUT_MILLIS, linkedinConnectTimeoutInSeconds * 1000)
.doOnConnected(conn -> conn
.addHandlerLast(new ReadTimeoutHandler(linkedinReadTimeoutInSeconds, TimeUnit.SECONDS))
.addHandlerLast(new WriteTimeoutHandler(linkedinWriteTimeoutInSeconds, TimeUnit.SECONDS)));
return tcpClient;
});
// create a client http connector using above http client
return new ReactorClientHttpConnector(httpClient);
}
If we don't set logging.level.reactor.netty.http.client=debug, neither of these webclients log netty debug statements. If we set the debug parameter, both log the requests, the second one also logs request/response headers and bodies.
I have the following program through which I can detect the connection failure i.e doBeforeRetry.
Can someone tell me how to detect the successful connection or reconnection. I want to integrate a Health Check program that monitors this connection, but I am unable to capture the event that informs the connections is successfull.
Thanks
requester = RSocketRequester.builder()
.rsocketConnector(connector -> {
connector.reconnect(Retry
.fixedDelay(Integer.MAX_VALUE,Duration.ofSeconds(1))
.doBeforeRetry(e-> System.out.println("doBeforeRetry===>"+e))
.doAfterRetry(e-> System.out.println("doAfterRetry===>"+e))
);
connector.payloadDecoder(PayloadDecoder.ZERO_COPY);
}
).dataMimeType(MediaType.APPLICATION_CBOR)
.rsocketStrategies(strategies)
.tcp("localhost", 7999);
I achieved the detection of successful connection or reconnection with the following approach.
Client Side (Connection initialization)
Mono<RSocketRequester> requester = Mono.just(RSocketRequester.builder()
.rsocketConnector(
// connector configuration goes here
)
.dataMimeType(MediaType.APPLICATION_CBOR)
.setupRoute("client-handshake")
.setupData("caller-name")
.tcp("localhost", 7999)));
One the server side
#ConnectMapping("client-handshake")
public void connect(RSocketRequester requester, #Payload String callerName) {
LOG.info("Client Connection Handshake: [{}]", callerName);
requester
.route("server-handshake")
.data("I am server")
.retrieveMono(Void.class)
.subscribe();
}
On the client side, when I receive the callback on the below method, I detect the connection is successfull.
#MessageMapping("server-handshake")
public Mono<ConsumerPreference> handshake(final String response){
LOG.info("Server Connection Handshake received : Server message [{}]", response.getCallerName());
connectionSuccess.set(true);
return Mono.empty();
}else{
throw new InitializationException("Invalid response message received from Server");
}
}
Additionally, created a application level heartbeat to ensure, the liveliness of the connection.
If you want to know if it's actually healthy, you should probably have a side task that is polling the health of the RSocket, by sending something like a custom ping protocol to your backend. You could time that and confirm that you have a healthy connection, record latencies and success/failures.
Timeout Without Using Proxy
I start netcat in my local as follows, which basically listens to connections on port 9090:
netcat -l -p 9090
And using Apache HttpComponents, I create a connection to it with a timeout of 4 seconds..
RequestConfig requestConfig = RequestConfig.custom()
.setSocketTimeout(4000)
.setConnectTimeout(4000)
.setConnectionRequestTimeout(4000)
.build();
HttpGet httpget = new HttpGet("http://127.0.0.1:9090");
httpget.setConfig(requestConfig);
try (CloseableHttpResponse response = HttpClients.createDefault().execute(httpget)) {}
In terminal (where I have netcat running) I see:
??]?D???;#???9?Mۡ?NR?w?{)?V?$?(=?&?*kj?
?5??98?#?'<?%?)g#? ?/??32?,?+?0??.?2???/??-?1???D
<!-- 4 seconds later -->
read(net): Connection reset by peer
In client side what I see is:
Exception in thread "main" org.apache.http.conn.ConnectTimeoutException:
Connect to 127.0.0.1:9090 [/127.0.0.1] failed: Read timed out
This is all expected.
Timeout Using Proxy
I change the client code slightly and configure a proxy, following the docs here.
RequestConfig requestConfig = RequestConfig.custom()
.setSocketTimeout(4000)
.setConnectTimeout(4000)
.setConnectionRequestTimeout(4000)
.build();
HttpHost proxy = new HttpHost("127.0.0.1", 9090);
DefaultProxyRoutePlanner routePlanner = new DefaultProxyRoutePlanner(proxy);
CloseableHttpClient httpclient = HttpClients.custom()
.setRoutePlanner(routePlanner)
.build();
HttpGet httpget = new HttpGet("https://127.0.0.1:9090");
httpget.setConfig(requestConfig);
try (CloseableHttpResponse response = httpclient.execute(httpget)) {}
And again start netcat, and this time on serverside
CONNECT 127.0.0.1:9090 HTTP/1.1
Host: 127.0.0.1:9090
User-Agent: Apache-HttpClient/4.4.1 (Java/1.8.0_212)
But timeout is not working for CONNECT. I just wait forever..
How can I configure the httpclient to timeout for 4 seconds just like in the first case I described?
RequestConfig only take effect once a connection to the target via the specific route has been fully established . They do not apply to the SSL handshake or any CONNECT requests that take place prior to the main message exchange.
Configure socket timeout at the ConnectionManager level to ensure connection level operations time out after a certain period of inactivity.
One possibility:
// This part is the same..
httpget.setConfig(requestConfig);
ExecutorService executorService = Executors.newSingleThreadExecutor();
Callable<CloseableHttpResponse> callable = () -> {
try (CloseableHttpResponse response = httpclient.execute(httpget)) {
return response;
}
};
Future<CloseableHttpResponse> future = executorService.submit(callable);
try {
future.get(4, TimeUnit.SECONDS);
} catch (InterruptedException | ExecutionException | TimeoutException e) {
httpget.abort();
executorService.shutdownNow();
}
But I am open to other suggestions..
Settings settings = Settings.builder()
.put("client.transport.ignore_cluster_name", false)
.put("client.transport.sniff", true)
.put("cluster.name", "TESTCULSTER").build();
TransportClient client = new PreBuiltTransportClient(settings)
.addTransportAddress(new TransportAddress(InetAddress.getByName("127.0.0.1"), 9300));
// ClusterAdminClient clusterAdminClient = client.admin().cluster();
ClusterHealthResponse healths = client.admin().cluster().prepareHealth().get();
String clusterName = healths.getClusterName();
System.out.println(clusterName);
I am geeting this error
Exception in thread "main" NoNodeAvailableException[None of the configured nodes are available: [{#transport#-1}{zJ52yLDcR82UUQ7j-oxm6w}{127.0.0.1}{127.0.0.1:9300}]]
You want to connect to an elasticsearch via Java, right?
I suggest to use the HTTP transport (port 9200) rather than 9300
You enabled "sniffing", that means Java client will try to connect to each nodes directly, be sure communication is possible
Be sure you can curl your ES node from where you run the Java client