十年网站开发经验 + 多家企业客户 + 靠谱的建站团队
量身定制 + 运营维护+专业推广+无忧售后,网站问题一站解决
这篇文章主要介绍“dubbo-go中leastActiveLoadBalance的用法”,在日常操作中,相信很多人在dubbo-go中leastActiveLoadBalance的用法问题上存在疑惑,小编查阅了各式资料,整理出简单好用的操作方法,希望对大家解答”dubbo-go中leastActiveLoadBalance的用法”的疑惑有所帮助!接下来,请跟着小编一起来学习吧!
成都创新互联成立与2013年,是专业互联网技术服务公司,拥有项目做网站、成都网站制作网站策划,项目实施与项目整合能力。我们以让每一个梦想脱颖而出为使命,1280元中宁做网站,已为上家服务,为中宁各地企业和个人服务,联系电话:028-86922220
本文主要研究一下dubbo-go的leastActiveLoadBalance
dubbo-go-v1.4.2/cluster/loadbalance/least_active.go
const ( // LeastActive ... LeastActive = "leastactive" ) func init() { extension.SetLoadbalance(LeastActive, NewLeastActiveLoadBalance) } type leastActiveLoadBalance struct { } // NewLeastActiveLoadBalance ... func NewLeastActiveLoadBalance() cluster.LoadBalance { return &leastActiveLoadBalance{} }
leastActiveLoadBalance的NewLeastActiveLoadBalance方法创建leastActiveLoadBalance
dubbo-go-v1.4.2/cluster/loadbalance/least_active.go
func (lb *leastActiveLoadBalance) Select(invokers []protocol.Invoker, invocation protocol.Invocation) protocol.Invoker { count := len(invokers) if count == 0 { return nil } if count == 1 { return invokers[0] } var ( leastActive int32 = -1 // The least active value of all invokers totalWeight int64 // The number of invokers having the same least active value (LEAST_ACTIVE) firstWeight int64 // Initial value, used for comparison leastCount int // The number of invokers having the same least active value (LEAST_ACTIVE) leastIndexes = make([]int, count) // The index of invokers having the same least active value (LEAST_ACTIVE) sameWeight = true // Every invoker has the same weight value? ) for i := 0; i < count; i++ { invoker := invokers[i] // Active number active := protocol.GetMethodStatus(invoker.GetUrl(), invocation.MethodName()).GetActive() // current weight (maybe in warmUp) weight := GetWeight(invoker, invocation) // There are smaller active services if leastActive == -1 || active < leastActive { leastActive = active leastIndexes[0] = i leastCount = 1 // next available leastIndex offset totalWeight = weight firstWeight = weight sameWeight = true } else if active == leastActive { leastIndexes[leastCount] = i totalWeight += weight leastCount++ if sameWeight && (i > 0) && weight != firstWeight { sameWeight = false } } } if leastCount == 1 { return invokers[0] } if !sameWeight && totalWeight > 0 { offsetWeight := rand.Int63n(totalWeight) + 1 for i := 0; i < leastCount; i++ { leastIndex := leastIndexes[i] offsetWeight -= GetWeight(invokers[i], invocation) if offsetWeight <= 0 { return invokers[leastIndex] } } } index := leastIndexes[rand.Intn(leastCount)] return invokers[index] }
Select方法遍历invokers,挨个通过protocol.GetMethodStatus(invoker.GetUrl(), invocation.MethodName()).GetActive()获取active信息,并通过GetWeight(invoker, invocation)获取weight,然后计算leastCount、totalWeight、sameWeight;对于leastCount为1的返回invokers[0],对于sameWeight为false且totalWeight大于0的,遍历leastIndexes,计算offsetWeight,若offsetWeight小于等于0,则返回invokers[leastIndex],否则通过leastIndexes[rand.Intn(leastCount)]计算index,返回invokers[index]
dubbo-go-v1.4.2/cluster/loadbalance/util.go
// GetWeight ... func GetWeight(invoker protocol.Invoker, invocation protocol.Invocation) int64 { url := invoker.GetUrl() weight := url.GetMethodParamInt64(invocation.MethodName(), constant.WEIGHT_KEY, constant.DEFAULT_WEIGHT) if weight > 0 { //get service register time an do warm up time now := time.Now().Unix() timestamp := url.GetParamInt(constant.REMOTE_TIMESTAMP_KEY, now) if uptime := now - timestamp; uptime > 0 { warmup := url.GetParamInt(constant.WARMUP_KEY, constant.DEFAULT_WARMUP) if uptime < warmup { if ww := float64(uptime) / float64(warmup) / float64(weight); ww < 1 { weight = 1 } else if int64(ww) <= weight { weight = int64(ww) } } } } return weight }
GetWeight方法通过url.GetMethodParamInt64(invocation.MethodName(), constant.WEIGHT_KEY, constant.DEFAULT_WEIGHT)获取weight,若weight大于0,则计算warmup,重新计算weight值
leastActiveLoadBalance的NewLeastActiveLoadBalance方法创建leastActiveLoadBalance;Select方法遍历invokers,挨个通过protocol.GetMethodStatus(invoker.GetUrl(), invocation.MethodName()).GetActive()获取active信息,并通过GetWeight(invoker, invocation)获取weight,然后计算leastCount、totalWeight、sameWeight,最后计算index
到此,关于“dubbo-go中leastActiveLoadBalance的用法”的学习就结束了,希望能够解决大家的疑惑。理论与实践的搭配能更好的帮助大家学习,快去试试吧!若想继续学习更多相关知识,请继续关注创新互联网站,小编会继续努力为大家带来更多实用的文章!