Junlong
Junlong
https://github.com/mosn/mosn/blob/647bb0fe16de924eba8797a0ccb5e4d3b1daf93c/pkg/stagemanager/stage_manager.go#L431 没太清楚怎么通知的，能给下路径吗，tks
@doujiang24 @nejisama
> 目前的reload是靠向旧进程发送HUP信号触发热升级流程执行reload，但是MOSN是可以通过配置关闭热升级的，有没有测试过这种情况下，reload是什么结果？ 如果打开热升级执行 reload的话，新进程会向老mosn进程向发热升级信号，老进程会报错 "[server] SIGHUP received: current state expected running while got 13"，最终导致无法执行reload，这个问题我在最新的commit内修复了，执行reload的时候不发送热升级信号。
> 补充对应的文档到MOSN官网，需要说明各种细节。 比如mosn reload -c config.json 并不是用config.json去reload，还是用原来start的配置进行reload的等 好的，等PR合进好，我再提个PR到mosn.io
``` Error: -01 07:41:19,967 [ERROR] [config] network filter create failed, type:tcp_proxy, error: unsupported network filter type: tcp_proxy Error: -01 07:41:19,967 [ERROR] [config] network filter factories len is 0, listener: &{ListenerConfig:{Name:0.0.0.0_80...
可以返回多值的。只不过他是返回在一个`ResultsMap`中，你可以通过判断`ResultsMap`内的所有`value`是否都为`false`来实现。 ``` err,result := svc.Pool.ExecuteConcurrent(reqData) if err != nil{ logger.LogInfof("Execute err:%s ", err) return err } for ruleName, v := range result{ if v.(bool){ continue } return fmt.Errorf("ruleName:[%v] result is...
不会几秒dump一次，最极端场景下是每次 [cooldown](https://github.com/mosn/holmes/#:~:text=WithCPUDump(10%2C%2025%2C%2080%2Ctime.Minute)%20means%20dump%20will%20happen%20when%20cpu%20usage%20%3E%2010%25%20%26%26%20cpu%20usage%20%3E%20125%25%20*%20previous%20cpu%20usage%20recorded%20or%20cpu%20usage%20%3E%2080%25.%20time.Minute%20means%20once%20a%20dump%20happened%2C%20the%20next%20dump%20will%20not%20happen%20before%20cooldown%20finish%2D1%20minute.) 一结束，立马dump一次。
Hi，@Stebalien . My idea is using a _service_ struct to control the buses. If you approve, I can try to achieve it. Thanks.
Hi Curl, thanks for your PR, this is a large PR for holmes, I need more time to figure out them, and these changes seems very like [automaxprocs/tree/master/internal/cgroups](https://github.com/uber-go/automaxprocs/tree/master/internal/cgroups). It seems...