背景：早些时候kubernetes集群的cri还使用docker的时候经历过这样的状况：集群运行很久后硬盘跑的快满了......,大文件主要集中在：/var/lib/docker/overlay2下文件有快70G，/var/log/journal/日志也有4-5G。当时的操作是手工的在work节点进行了一下的操作：journalctl--vacuum-size=20M#设置journal日志最大为20M不保留不必要日志。dockerimageprune-a--filter"until=24h"#清除超过创建时间超过24小时的镜像dockercontainerprune--filter"unti

论文信息论文地址：​​https://arxiv.org/pdf/2106.14112.pdf​​源码地址：​​https://github.com/emadeldeen24/TS-TCC​​摘要Learningdecentrepresentationsfromunlabeledtime-seriesdatawithtemporaldynamicsisaverychallengingtask.Inthispaper,wepro-poseanunsupervisedTime-SeriesrepresentationlearningframeworkviaTemporalandContextual

论文信息论文地址：​​https://arxiv.org/pdf/2106.14112.pdf​​源码地址：​​https://github.com/emadeldeen24/TS-TCC​​摘要Learningdecentrepresentationsfromunlabeledtime-seriesdatawithtemporaldynamicsisaverychallengingtask.Inthispaper,wepro-poseanunsupervisedTime-SeriesrepresentationlearningframeworkviaTemporalandContextual