moons

2、还有一个问题是，我下载普通文件，supernode中有该缓存文件的情况下，第一次下载比第二次下载时间耗时要长10s左右（两次下载间隔10分钟）。而且日志不太一样，辛苦谁清楚帮忙解释一下 第一次dfget的日志 ``` 2020-11-04 15:23:24.857 INFO sign:7088-1604474604.857 : get cmd params:["dfget" "--url" "https://glance.bj.bcebos.com/efc2bb16-973a-48f6-ad26-4277f83fb1e3?authorization=bce-auth-v1%2F1a1125029c7349f1b671e81e844587b7%2F2020-08-31T04%3A24%3A06Z%2F-1%2Fhost%2Fac108742b" "--node" "2.71.153.145:8002" "-o" "image.iso" "--console" "--verbose" "-f" "authorization"] 2020-11-04 15:23:24.857 INFO sign:7088-1604474604.857 : get init config:{"url":"https://glance.bj.bcebos.com/efc2bb16-973a-48f6-ad26-4277f83fb1e3?authorization=bce-auth-v1%2F1a1125029c7349f1b671e81e844587b7%2F2020-08-31T04%3A24%3A06Z%2F-1%2Fhost%2Fac1b05033b27422f6f63c9b97c8f9b6a0921f7249875ac1438203cb60f08742b","output":"/root/moons/image.iso","pattern":"p2p","filter":["authorization"],"console":true,"verbose":true,"nodes":["2.71.153.145:8002=1"],"localLimit":"20GB","minRate":"64KB","clientQueueSize":6,"workHome":"/root/.small-dragonfly","logConfig":{"maxSize":0,"maxBackups":0,"path":"/root/.small-dragonfly/logs/dfclient.log"}} --2020-11-04...

> fluentd关于容器log的pos是记录在`pos_file /pilot/pos/{{ $.containerId }}.{{ .Name }}.pos`这里，但是在pod中容器重启之后log所在的目录不会变，容器id发生了变化，所以pos也就发生了变化，fluentd就会对重启之后的log进行重采集，造成log采集重复，有没有遇到这种情况，反正我遇到了 这个log所在的目录应该会变吧！因为容器的ID变了。  

/assign @@repeatedly

> @ZP-AlwaysWin One more thing, just to make it sure. You do have at least one containerd-launched plugin in `/opt/nri/plugins`, right ? Yes, I have it.