聊聊eureka instance的lastDirtyTimestamp

序

本文主要研究一下eureka instance的lastDirtyTimestamp

server端

• lastDirtyTimestamp last timestamp when this instance information was updated.

即該instance在client端最後被修改的時間戳

instance的接口，除了更新meta以及cancelLease操做外，其餘修改的操做都要帶上lastDirtyTimestamp，好比 eureka-core-1.8.8-sources.jar!/com/netflix/eureka/resources/InstanceResource.java

@PUT
    public Response renewLease(
            @HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication,
            @QueryParam("overriddenstatus") String overriddenStatus,
            @QueryParam("status") String status,
            @QueryParam("lastDirtyTimestamp") String lastDirtyTimestamp) {
        //......
    }

    @PUT
    @Path("status")
    public Response statusUpdate(
            @QueryParam("value") String newStatus,
            @HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication,
            @QueryParam("lastDirtyTimestamp") String lastDirtyTimestamp) {
        //......
    }

    @DELETE
    @Path("status")
    public Response deleteStatusUpdate(
            @HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication,
            @QueryParam("value") String newStatusValue,
            @QueryParam("lastDirtyTimestamp") String lastDirtyTimestamp) {
        //......
    }

除renewLease外，其餘兩個的lastDirtyTimestamp僅僅是用來作傳遞用。

renewLease

@PUT
    public Response renewLease(
            @HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication,
            @QueryParam("overriddenstatus") String overriddenStatus,
            @QueryParam("status") String status,
            @QueryParam("lastDirtyTimestamp") String lastDirtyTimestamp) {
        boolean isFromReplicaNode = "true".equals(isReplication);
        boolean isSuccess = registry.renew(app.getName(), id, isFromReplicaNode);

        // Not found in the registry, immediately ask for a register
        if (!isSuccess) {
            logger.warn("Not Found (Renew): {} - {}", app.getName(), id);
            return Response.status(Status.NOT_FOUND).build();
        }
        // Check if we need to sync based on dirty time stamp, the client
        // instance might have changed some value
        Response response = null;
        if (lastDirtyTimestamp != null && serverConfig.shouldSyncWhenTimestampDiffers()) {
            response = this.validateDirtyTimestamp(Long.valueOf(lastDirtyTimestamp), isFromReplicaNode);
            // Store the overridden status since the validation found out the node that replicates wins
            if (response.getStatus() == Response.Status.NOT_FOUND.getStatusCode()
                    && (overriddenStatus != null)
                    && !(InstanceStatus.UNKNOWN.name().equals(overriddenStatus))
                    && isFromReplicaNode) {
                registry.storeOverriddenStatusIfRequired(app.getAppName(), id, InstanceStatus.valueOf(overriddenStatus));
            }
        } else {
            response = Response.ok().build();
        }
        logger.debug("Found (Renew): {} - {}; reply status={}", app.getName(), id, response.getStatus());
        return response;
    }

若是renew不成功，則返回404，client端則會執行register邏輯 若是renew成功，且lastDirtyTimestamp不爲null，則判斷是否須要SyncWhenTimestampDiffers，默認true

validateDirtyTimestamp

private Response validateDirtyTimestamp(Long lastDirtyTimestamp,
                                            boolean isReplication) {
        InstanceInfo appInfo = registry.getInstanceByAppAndId(app.getName(), id, false);
        if (appInfo != null) {
            if ((lastDirtyTimestamp != null) && (!lastDirtyTimestamp.equals(appInfo.getLastDirtyTimestamp()))) {
                Object[] args = {id, appInfo.getLastDirtyTimestamp(), lastDirtyTimestamp, isReplication};

                if (lastDirtyTimestamp > appInfo.getLastDirtyTimestamp()) {
                    logger.debug(
                            "Time to sync, since the last dirty timestamp differs -"
                                    + " ReplicationInstance id : {},Registry : {} Incoming: {} Replication: {}",
                            args);
                    return Response.status(Status.NOT_FOUND).build();
                } else if (appInfo.getLastDirtyTimestamp() > lastDirtyTimestamp) {
                    // In the case of replication, send the current instance info in the registry for the
                    // replicating node to sync itself with this one.
                    if (isReplication) {
                        logger.debug(
                                "Time to sync, since the last dirty timestamp differs -"
                                        + " ReplicationInstance id : {},Registry : {} Incoming: {} Replication: {}",
                                args);
                        return Response.status(Status.CONFLICT).entity(appInfo).build();
                    } else {
                        return Response.ok().build();
                    }
                }
            }

        }
        return Response.ok().build();
    }

若是lastDirtyTimestamp參數大於server本地instance的lastDirtyTimestamp值，則response返回404； 若是是server本地大於lastDirtyTimestamp參數，不是replication模式則返回200，replication模式，則返回409 Conflict，讓調用方同步本身的數據。

PeerEurekaNode

eureka-core-1.8.8-sources.jar!/com/netflix/eureka/cluster/PeerEurekaNode.java

/**
     * Send the heartbeat information of an instance to the node represented by
     * this class. If the instance does not exist the node, the instance
     * registration information is sent again to the peer node.
     *
     * @param appName
     *            the application name of the instance.
     * @param id
     *            the unique identifier of the instance.
     * @param info
     *            the instance info {@link InstanceInfo} of the instance.
     * @param overriddenStatus
     *            the overridden status information if any of the instance.
     * @throws Throwable
     */
    public void heartbeat(final String appName, final String id,
                          final InstanceInfo info, final InstanceStatus overriddenStatus,
                          boolean primeConnection) throws Throwable {
        if (primeConnection) {
            // We do not care about the result for priming request.
            replicationClient.sendHeartBeat(appName, id, info, overriddenStatus);
            return;
        }
        ReplicationTask replicationTask = new InstanceReplicationTask(targetHost, Action.Heartbeat, info, overriddenStatus, false) {
            @Override
            public EurekaHttpResponse<InstanceInfo> execute() throws Throwable {
                return replicationClient.sendHeartBeat(appName, id, info, overriddenStatus);
            }

            @Override
            public void handleFailure(int statusCode, Object responseEntity) throws Throwable {
                super.handleFailure(statusCode, responseEntity);
                if (statusCode == 404) {
                    logger.warn("{}: missing entry.", getTaskName());
                    if (info != null) {
                        logger.warn("{}: cannot find instance id {} and hence replicating the instance with status {}",
                                getTaskName(), info.getId(), info.getStatus());
                        register(info);
                    }
                } else if (config.shouldSyncWhenTimestampDiffers()) {
                    InstanceInfo peerInstanceInfo = (InstanceInfo) responseEntity;
                    if (peerInstanceInfo != null) {
                        syncInstancesIfTimestampDiffers(appName, id, info, peerInstanceInfo);
                    }
                }
            }
        };
        long expiryTime = System.currentTimeMillis() + getLeaseRenewalOf(info);
        batchingDispatcher.process(taskId("heartbeat", info), replicationTask, expiryTime);
    }

這裏的handleFailure，對於404的處理就是從新register；其餘的就是諸如409的狀況，先判斷是否開啓shouldSyncWhenTimestampDiffers配置，這個默認是true，而後將peer返回的最新info信息覆蓋到本地

syncInstancesIfTimestampDiffers

/**
     * Synchronize {@link InstanceInfo} information if the timestamp between
     * this node and the peer eureka nodes vary.
     */
    private void syncInstancesIfTimestampDiffers(String appName, String id, InstanceInfo info, InstanceInfo infoFromPeer) {
        try {
            if (infoFromPeer != null) {
                logger.warn("Peer wants us to take the instance information from it, since the timestamp differs,"
                        + "Id : {} My Timestamp : {}, Peer's timestamp: {}", id, info.getLastDirtyTimestamp(), infoFromPeer.getLastDirtyTimestamp());

                if (infoFromPeer.getOverriddenStatus() != null && !InstanceStatus.UNKNOWN.equals(infoFromPeer.getOverriddenStatus())) {
                    logger.warn("Overridden Status info -id {}, mine {}, peer's {}", id, info.getOverriddenStatus(), infoFromPeer.getOverriddenStatus());
                    registry.storeOverriddenStatusIfRequired(appName, id, infoFromPeer.getOverriddenStatus());
                }
                registry.register(infoFromPeer, true);
            }
        } catch (Throwable e) {
            logger.warn("Exception when trying to set information from peer :", e);
        }
    }

這一段就是peerNode在replicate中遇到409的時候，根據返回的InstanceInfo覆蓋本地的過程。

client端

eureka-client-1.8.8-sources.jar!/com/netflix/appinfo/InstanceInfo.java

/**
     * Sets the dirty flag so that the instance information can be carried to
     * the discovery server on the next heartbeat.
     */
    public synchronized void setIsDirty() {
        isInstanceInfoDirty = true;
        lastDirtyTimestamp = System.currentTimeMillis();
    }

這個是更改lastDirtyTimestamp的基本方法，調用這個方法的其餘方法以下：

@Deprecated
    public void setSID(String sid) {
        this.sid = sid;
        setIsDirty();
    }

    /**
     * Set the status for this instance.
     *
     * @param status status for this instance.
     * @return the prev status if a different status from the current was set, null otherwise
     */
    public synchronized InstanceStatus setStatus(InstanceStatus status) {
        if (this.status != status) {
            InstanceStatus prev = this.status;
            this.status = status;
            setIsDirty();
            return prev;
        }
        return null;
    }

    /**
     * @param isDirty true if dirty, false otherwise.
     * @deprecated use {@link #setIsDirty()} and {@link #unsetIsDirty(long)} to set and unset
     * <p>
     * Sets the dirty flag so that the instance information can be carried to
     * the discovery server on the next heartbeat.
     */
    @Deprecated
    public synchronized void setIsDirty(boolean isDirty) {
        if (isDirty) {
            setIsDirty();
        } else {
            isInstanceInfoDirty = false;
            // else don't update lastDirtyTimestamp as we are setting isDirty to false
        }
    }

    /**
     * Set the dirty flag, and also return the timestamp of the isDirty event
     *
     * @return the timestamp when the isDirty flag is set
     */
    public synchronized long setIsDirtyWithTime() {
        setIsDirty();
        return lastDirtyTimestamp;
    }

    /**
     * Register application specific metadata to be sent to the discovery
     * server.
     *
     * @param runtimeMetadata
     *            Map containing key/value pairs.
     */
    synchronized void registerRuntimeMetadata(
            Map<String, String> runtimeMetadata) {
        metadata.putAll(runtimeMetadata);
        setIsDirty();
    }

isInstanceInfoDirty

@JsonIgnore
    public boolean isDirty() {
        return isInstanceInfoDirty;
    }

    /**
     * @return the lastDirtyTimestamp if is dirty, null otherwise.
     */
    public synchronized Long isDirtyWithTime() {
        if (isInstanceInfoDirty) {
            return lastDirtyTimestamp;
        } else {
            return null;
        }
    }

    /**
     * Unset the dirty flag iff the unsetDirtyTimestamp matches the lastDirtyTimestamp. No-op if
     * lastDirtyTimestamp > unsetDirtyTimestamp
     *
     * @param unsetDirtyTimestamp the expected lastDirtyTimestamp to unset.
     */
    public synchronized void unsetIsDirty(long unsetDirtyTimestamp) {
        if (lastDirtyTimestamp <= unsetDirtyTimestamp) {
            isInstanceInfoDirty = false;
        } else {
        }
    }

這個方法，若是lastDirtyTimestamp <= unsetDirtyTimestamp，則標識isInstanceInfoDirty爲false

InstanceInfoReplicator

eureka-client-1.8.8-sources.jar!/com/netflix/discovery/InstanceInfoReplicator.java

public void run() {
        try {
            discoveryClient.refreshInstanceInfo();

            Long dirtyTimestamp = instanceInfo.isDirtyWithTime();
            if (dirtyTimestamp != null) {
                discoveryClient.register();
                instanceInfo.unsetIsDirty(dirtyTimestamp);
            }
        } catch (Throwable t) {
            logger.warn("There was a problem with the instance info replicator", t);
        } finally {
            Future next = scheduler.schedule(this, replicationIntervalSeconds, TimeUnit.SECONDS);
            scheduledPeriodicRef.set(next);
        }
    }

這裏首先refreshInstanceInfo，以後判斷若是是isInstanceInfoDirty則會執行register操做

refreshInstanceInfo

/**
     * Refresh the current local instanceInfo. Note that after a valid refresh where changes are observed, the
     * isDirty flag on the instanceInfo is set to true
     */
    void refreshInstanceInfo() {
        applicationInfoManager.refreshDataCenterInfoIfRequired();
        applicationInfoManager.refreshLeaseInfoIfRequired();

        InstanceStatus status;
        try {
            status = getHealthCheckHandler().getStatus(instanceInfo.getStatus());
        } catch (Exception e) {
            logger.warn("Exception from healthcheckHandler.getStatus, setting status to DOWN", e);
            status = InstanceStatus.DOWN;
        }

        if (null != status) {
            applicationInfoManager.setInstanceStatus(status);
        }
    }

refreshDataCenterInfoIfRequired

eureka-client-1.8.8-sources.jar!/com/netflix/appinfo/ApplicationInfoManager.java

/**
     * Refetches the hostname to check if it has changed. If it has, the entire
     * <code>DataCenterInfo</code> is refetched and passed on to the eureka
     * server on next heartbeat.
     *
     * see {@link InstanceInfo#getHostName()} for explanation on why the hostname is used as the default address
     */
    public void refreshDataCenterInfoIfRequired() {
        String existingAddress = instanceInfo.getHostName();

        String newAddress;
        if (config instanceof RefreshableInstanceConfig) {
            // Refresh data center info, and return up to date address
            newAddress = ((RefreshableInstanceConfig) config).resolveDefaultAddress(true);
        } else {
            newAddress = config.getHostName(true);
        }
        String newIp = config.getIpAddress();

        if (newAddress != null && !newAddress.equals(existingAddress)) {
            logger.warn("The address changed from : {} => {}", existingAddress, newAddress);

            // :( in the legacy code here the builder is acting as a mutator.
            // This is hard to fix as this same instanceInfo instance is referenced elsewhere.
            // We will most likely re-write the client at sometime so not fixing for now.
            InstanceInfo.Builder builder = new InstanceInfo.Builder(instanceInfo);
            builder.setHostName(newAddress).setIPAddr(newIp).setDataCenterInfo(config.getDataCenterInfo());
            instanceInfo.setIsDirty();
        }
    }

    public void refreshLeaseInfoIfRequired() {
        LeaseInfo leaseInfo = instanceInfo.getLeaseInfo();
        if (leaseInfo == null) {
            return;
        }
        int currentLeaseDuration = config.getLeaseExpirationDurationInSeconds();
        int currentLeaseRenewal = config.getLeaseRenewalIntervalInSeconds();
        if (leaseInfo.getDurationInSecs() != currentLeaseDuration || leaseInfo.getRenewalIntervalInSecs() != currentLeaseRenewal) {
            LeaseInfo newLeaseInfo = LeaseInfo.Builder.newBuilder()
                    .setRenewalIntervalInSecs(currentLeaseRenewal)
                    .setDurationInSecs(currentLeaseDuration)
                    .build();
            instanceInfo.setLeaseInfo(newLeaseInfo);
            instanceInfo.setIsDirty();
        }
    }

重新刷新配置，若是有變動則instanceInfo.setIsDirty()

小結

在server端，處理renewLease的時候，對lastDirtyTimestamp參數進行判斷，若是大於server本地instance的lastDirtyTimestamp值，則response返回404；若是是server本地大於lastDirtyTimestamp參數，不是replication模式則返回200，replication模式，則返回409 Conflict，讓調用方同步本身的數據。

在client端，有兩個屬性，一個是lastDirtyTimestamp，一個是isInstanceInfoDirty。在更新instance的status的時候，會調用setIsDirty，即更新lastDirtyTimestamp以及設置isInstanceInfoDirty爲true；而後client端還有個InstanceInfoReplicator定時任務，會定時讀取配置文件，若是有變動則調用setIsDirty，以後是調用instanceInfo.isDirtyWithTime()，若是是dirty則會從新register並重置isInstanceInfoDirty，更新lastDirtyTimestamp。

doc

• 聊聊eureka client的HeartbeatThread
• 聊聊springcloud的EurekaClientAutoConfiguration
• 聊聊EurekaHealthCheckHandler
• 聊聊eureka server的instance註冊及元數據變動接口