区块链教程open-ethereum-pool矿池源码分析API分析
ApiServer相关定义
type ApiConfig struct {
Enabled bool `json:"enabled"`
Listen string `json:"listen"`
StatsCollectInterval string `json:"statsCollectInterval"`
HashrateWindow string `json:"hashrateWindow"`
HashrateLargeWindow string `json:"hashrateLargeWindow"`
LuckWindow []int `json:"luckWindow"`
Payments int64 `json:"payments"`
Blocks int64 `json:"blocks"`
PurgeOnly bool `json:"purgeOnly"`
PurgeInterval string `json:"purgeInterval"`
}
type ApiServer struct {
config *ApiConfig
backend *storage.RedisClient
hashrateWindow time.Duration
hashrateLargeWindow time.Duration
stats atomic.Value
miners map[string]*Entry
minersMu sync.RWMutex
statsIntv time.Duration
}
type Entry struct {
stats map[string]interface{}
updatedAt int64
}
//代码位置api/server.go
CollectStats原理
//config.StatsCollectInterval为CollectStats定时器，时间为5秒
//调取：stats, err := s.backend.CollectStats(s.hashrateWindow, s.config.Blocks, s.config.Payments)
//s.hashrateWindow即cfg.HashrateWindow，即：为每个矿工估计算力的快速时间间隔，30分钟
//s.config.Blocks即：前端显示的最大块数，50个
//s.config.Payments即：前端显示的最大付款数量，50个
func (r *RedisClient) CollectStats(smallWindow time.Duration, maxBlocks, maxPayments int64) (map[string]interface{}, error) {
//换算成秒
window := int64(smallWindow / time.Second)
//创建map
stats := make(map[string]interface{})
//Redis事务块
tx := r.client.Multi()
defer tx.Close()
//换算成秒
now := util.MakeTimestamp() / 1000
cmds, err := tx.Exec(func() error {
//针对min和max参数需要额外说明的是，-inf和+inf分别表示Sorted-Sets中分数的最高值和最低值
//缺省情况下，min和max表示的范围是闭区间范围，即min <= score <= max内的成员将被返回
//然而我们可以通过在min和max的前面添加"("字符来表示开区间，如(min max表示min < score <= max，而(min (max表示min < score < max
//-inf <= score < now-window
//Zremrangebyscore 命令用于移除有序集中，指定分数（score）区间内的所有成员
//ZREMRANGEBYSCORE eth:hashrate -inf (now-window
//慎重使用
tx.ZRemRangeByScore(r.formatKey("hashrate"), "-inf", fmt.Sprint("(", now-window))
//显示整个有序集成员
//ZRANGE eth:hashrate 0 -1 WITHSCORES
tx.ZRangeWithScores(r.formatKey("hashrate"), 0, -1)
//Hgetall 命令用于返回哈希表中，所有的字段和值
//HGETALL eth:stats
tx.HGetAllMap(r.formatKey("stats"))
//Zrevrange 命令返回有序集中，指定区间内的成员
//ZREVRANGE eth:blocks:candidates 0 -1 WITHSCORES
//candidates为候选者
tx.ZRevRangeWithScores(r.formatKey("blocks", "candidates"), 0, -1)
//同上
//ZREVRANGE eth:blocks:immature 0 -1 WITHSCORES
//immature为未成年
tx.ZRevRangeWithScores(r.formatKey("blocks", "immature"), 0, -1)
//同上
//ZREVRANGE eth:blocks:matured 0 49 WITHSCORES
//matured为成熟
tx.ZRevRangeWithScores(r.formatKey("blocks", "matured"), 0, maxBlocks-1)
//Zcard 命令用于计算集合中元素的数量
//ZCARD eth:blocks:candidates
tx.ZCard(r.formatKey("blocks", "candidates"))
//同上
//ZCARD eth:blocks:immature
tx.ZCard(r.formatKey("blocks", "immature"))
//同上
//ZCARD eth:blocks:matured
tx.ZCard(r.formatKey("blocks", "matured"))
//同上
//ZCARD eth:payments:all
tx.ZCard(r.formatKey("payments", "all"))
//同上
//ZREVRANGE eth:payments:all 0 49 WITHSCORES
tx.ZRevRangeWithScores(r.formatKey("payments", "all"), 0, maxPayments-1)
return nil
})
if err != nil {
return nil, err
}
//Hgetall 命令用于返回哈希表中，所有的字段和值
//HGETALL eth:stats
result, _ := cmds[2].(*redis.StringStringMapCmd).Result()
stats["stats"] = convertStringMap(result)
//Zrevrange 命令返回有序集中，指定区间内的成员
//ZREVRANGE eth:blocks:candidates 0 -1 WITHSCORES
//Zcard 命令用于计算集合中元素的数量
//ZCARD eth:blocks:candidates
candidates := convertCandidateResults(cmds[3].(*redis.ZSliceCmd))
stats["candidates"] = candidates
stats["candidatesTotal"] = cmds[6].(*redis.IntCmd).Val()
//ZREVRANGE eth:blocks:immature 0 -1 WITHSCORES
//ZCARD eth:blocks:immature
immature := convertBlockResults(cmds[4].(*redis.ZSliceCmd))
stats["immature"] = immature
stats["immatureTotal"] = cmds[7].(*redis.IntCmd).Val()
//ZREVRANGE eth:blocks:matured 0 49 WITHSCORES
//ZCARD eth:blocks:matured
matured := convertBlockResults(cmds[5].(*redis.ZSliceCmd))
stats["matured"] = matured
stats["maturedTotal"] = cmds[8].(*redis.IntCmd).Val()
//ZREVRANGE eth:payments:all 0 49 WITHSCORES
//ZCARD eth:payments:all
payments := convertPaymentsResults(cmds[10].(*redis.ZSliceCmd))
stats["payments"] = payments
stats["paymentsTotal"] = cmds[9].(*redis.IntCmd).Val()
//显示整个有序集成员
//ZRANGE eth:hashrate 0 -1 WITHSCORES
totalHashrate, miners := convertMinersStats(window, cmds[1].(*redis.ZSliceCmd))
stats["miners"] = miners
stats["minersTotal"] = len(miners)
stats["hashrate"] = totalHashrate
return stats, nil
}
CollectLuckStats原理
//调取：stats["luck"], err = s.backend.CollectLuckStats(s.config.LuckWindow)
//"luckWindow": [64, 128, 256],
//Collect stats for shares/diff ratio for this number of blocks
func (r *RedisClient) CollectLuckStats(windows []int) (map[string]interface{}, error) {
//创建stats map
stats := make(map[string]interface{})
tx := r.client.Multi()
defer tx.Close()
//max即256
max := int64(windows[len(windows)-1])
cmds, err := tx.Exec(func() error {
//Zrevrange 命令返回有序集中，指定区间内的成员
//ZREVRANGE eth:blocks:immature 0 -1 WITHSCORES
tx.ZRevRangeWithScores(r.formatKey("blocks", "immature"), 0, -1)
//ZREVRANGE eth:blocks:matured 0 max-1 WITHSCORES
tx.ZRevRangeWithScores(r.formatKey("blocks", "matured"), 0, max-1)
return nil
})
if err != nil {
return stats, err
}
//获取blocks
blocks := convertBlockResults(cmds[0].(*redis.ZSliceCmd), cmds[1].(*redis.ZSliceCmd))
calcLuck := func(max int) (int, float64, float64, float64) {
var total int
var sharesDiff, uncles, orphans float64
for i, block := range blocks {
if i > (max - 1) {
break
}
//叔块
if block.Uncle {
uncles++
}
//孤块
if block.Orphan {
orphans++
}
//shares/Diff
sharesDiff += float64(block.TotalShares) / float64(block.Difficulty)
//total计数
total++
}
if total > 0 {
//单块平均shares/Diff
sharesDiff /= float64(total)
//uncles率
uncles /= float64(total)
//孤块率
orphans /= float64(total)
}
//返回total计数，平均shares/Diff，uncles率，孤块率
return total, sharesDiff, uncles, orphans
}
//遍历windows，逐一计算calcLuck，即最近64块、128块、256块的数据统计
for _, max := range windows {
total, sharesDiff, uncleRate, orphanRate := calcLuck(max)
row := map[string]float64{
"luck": sharesDiff, "uncleRate": uncleRate, "orphanRate": orphanRate,
}
//写入stats map
stats[strconv.Itoa(total)] = row
//计数不对
if total < max {
break
}
}
return stats, nil
}
func convertBlockResults(rows ...*redis.ZSliceCmd) []*BlockData {
var result []*BlockData
//遍历rows
for _, row := range rows {
//遍历blocks
for _, v := range row.Val() {
// "uncleHeight:orphan:nonce:blockHash:timestamp:diff:totalShares:rewardInWei"
block := BlockData{}
block.Height = int64(v.Score)
block.RoundHeight = block.Height
fields := strings.Split(v.Member.(string), ":")
block.UncleHeight, _ = strconv.ParseInt(fields[0], 10, 64)
block.Uncle = block.UncleHeight > 0
block.Orphan, _ = strconv.ParseBool(fields[1])
block.Nonce = fields[2]
block.Hash = fields[3]
block.Timestamp, _ = strconv.ParseInt(fields[4], 10, 64)
block.Difficulty, _ = strconv.ParseInt(fields[5], 10, 64)
block.TotalShares, _ = strconv.ParseInt(fields[6], 10, 64)
block.RewardString = fields[7]
block.ImmatureReward = fields[7]
block.immatureKey = v.Member.(string)
result = append(result, &block)
}
}
return result
}
purgeStale原理
//config.PurgeInterval为FlushStaleStats定时器，时间为10分钟
//调取：total, err := s.backend.FlushStaleStats(s.hashrateWindow, s.hashrateLargeWindow)
//s.hashrateWindow即cfg.HashrateWindow，即：为每个矿工估计算力的快速时间间隔，30分钟
//s.hashrateLargeWindow即cfg.HashrateLargeWindow，即：长期和精确的hashrate时间间隔，3小时
func (r *RedisClient) FlushStaleStats(window, largeWindow time.Duration) (int64, error) {
//换算成秒
now := util.MakeTimestamp() / 1000
//max即(now-window，即<(now-window
max := fmt.Sprint("(", now-int64(window/time.Second))
//Zremrangebyscore 命令用于移除有序集中，指定分数（score）区间内的所有成员
//ZREMRANGEBYSCORE eth:hashrate -inf (now-window
//慎重使用
total, err := r.client.ZRemRangeByScore(r.formatKey("hashrate"), "-inf", max).Result()
if err != nil {
return total, err
}
var c int64
//创建map
miners := make(map[string]struct{})
//即(now-largeWindow，即
来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/31557831/viewspace-2216407/，如需转载，请注明出处，否则将追究法律责任。
转载于:http://blog.itpub.net/31557831/viewspace-2216407/