一次docker错误的耗时排查过程记录
由来
客户是深信服的订制系统，基于centos改的，排查半天发现居然是文件损坏，而不是docker的问题。
环境信息
docker信息:
$dockerinfo Containers:0 Running:0 Paused:0 Stopped:0 Images:2 ServerVersion:18.09.3 StorageDriver:overlay2 BackingFilesystem:xfs Supportsd_type:true NativeOverlayDiff:true LoggingDriver:json-file CgroupDriver:cgroupfs Plugins: Volume:local Network:bridgehostmacvlannulloverlay Log:awslogsfluentdgcplogsgelfjournaldjson-filelocallogentriessplunksyslog Swarm:inactive Runtimes:runc DefaultRuntime:runc InitBinary:docker-init containerdversion:e6b3f5632f50dbc4e9cb6288d911bf4f5e95b18e runcversion:6635b4f0c6af3810594d2770f662f34ddc15b40d initversion:fec3683 SecurityOptions: seccomp Profile:default KernelVersion:3.10.0 OperatingSystem:CentOSLinux7(Core) OSType:linux Architecture:x86_64 CPUs:20 TotalMemory:125.3GiB Name:eds-1f21a854 ID:VZLV:26PU:ZUN6:QQEJ:GW3I:YETT:HMEU:CK6J:SIPL:CHKV:6ASN:5NDF DockerRootDir:/data/kube/docker DebugMode(client):false DebugMode(server):false Registry:https://index.docker.io/v1/ Labels: Experimental:false InsecureRegistries: reg.wps.lan:5000 treg.yun.wps.cn 127.0.0.0/8 RegistryMirrors: https://registry.docker-cn.com/ https://docker.mirrors.ustc.edu.cn/ LiveRestoreEnabled:false ProductLicense:CommunityEngine
系统信息
$uname-a Linuxeds-1f21a8543.10.0#1SMPMonSep2812:00:30CST2020x86_64x86_64x86_64GNU/Linux $cat/etc/os-release NAME="CentOSLinux" VERSION="7(Core)" ID="centos" ID_LIKE="rhelfedora" VERSION_ID="7" PRETTY_NAME="CentOSLinux7(Core)" ANSI_COLOR="0;31" CPE_NAME="cpe:/o:centos:centos:7" HOME_URL="https://www.centos.org/" BUG_REPORT_URL="https://bugs.centos.org/" CENTOS_MANTISBT_PROJECT="CentOS-7" CENTOS_MANTISBT_PROJECT_VERSION="7" REDHAT_SUPPORT_PRODUCT="centos" REDHAT_SUPPORT_PRODUCT_VERSION="7"
服务器信息:
$catproduct_name SUGON-60G16 $catsys_vendor SANGFOR $catproduct_version 1.2
排查过程
安装docker服务器挂了
时间202010/2919:51:
实施:客户这边部署的时候安装docker的时候服务器挂了
我:开机后/var/log/message有信息吗
实施:只能恢复快照才能进去，服务器进不了，看不了信息
我:不恢复快照起不来吗
实施:是的
到这里我以为是触发了啥内核bug直接内核panic了服务器起不来。
时间202010/309:07:
我:起不来的时候有进控制台去看啥原因起不来吗
实施:是客户服务器没法查看呢
我:客户没去看下吗
然后实施直接发来一个向日葵远程连接，我上去后发现不是常规的操作系统，是基于centos改过的，没找到/var/log/message，然后手动执行我们的docker安装脚本。
bash-xinstall-docker.sh
然后输出的信息在某一步就没输出了，应该”挂了”，看了下脚本最后一条输出调试信息的后面是启动docker，应该是启动docker触发的。然后很久后还是无法连上和ping通，叫实施问问那边现场看看是硬件服务器的话有没有idrac，ilo之类的看看tty控制台的信息。
现场人员看了下服务器是”正常开机”的，我这边尝试还是连不上，现场问我们的操作是否是改了路由，现场systemctl看了下docker是起来的。现场那边还是ping不通网关。我这边突然想到是不是压根没挂。。。
叫他uptime-s看看上次的启动时间，结果压根没重启。。。
然后现场排查到是iptables的问题，启动docker的时候把他们的规则刷没了。后面他们改了下都放开了。所以前面的启动docker把机器挂了实际上是iptables的影响导致网络断开，机器压根没重启。
启动容器挂掉
然后继续，实施说之前同样的其他机器安装docker的时候没出现上面的问题，而是启动的时候出现上面的问题，我就手动执行下部署，结果报错。脚本开-x调试看是load部署镜像的时候报错了。
errorduringconnect:Posthttp://%2Fvar%2Frun%2Fdocker.sock/v1.39/images/load?quiet=0:readunix@->/var/run/docker.sock:read:EOF
手动执行下:
$dockerload-i./kube/images/deploy.tar errorduringconnect:Posthttp://%2Fvar%2Frun%2Fdocker.sock/v1.39/images/load?quiet=0:readunix@->/var/run/docker.sock:read:connectionresetbypeer
jounalctl看了下dockerdaemon没任何相关日志，这个报错搜了下有的人说是/var/run/docker.sock的docker组不存在，也有人直接chmod777解决的。试了下还是不行。前台debug下docker看看有没有有用的信息：
systemctlstopdocker pkilldockerd dockerd-D
另开一个终端执行load镜像操作:
$dockerload-i./kube/images/deploy.tar ab6425526dab:Loadinglayer[==================================================>]126.3MB/126.3MB c7fe3ea715ef:Loadinglayer[==================================================>]340.3MB/340.3MB 7f7eae7934f7:Loadinglayer[==================================================>]3.584kB/3.584kB e99a66706b50:Loadinglayer[==================================================>]2.105MB/2.105MB 245b79de6bb7:Loadinglayer[==================================================>]283.5MB/283.5MB 2a56a4432cef:Loadinglayer[==================================================>]93.18kB/93.18kB 0c2ea71066ab:Loadinglayer[==================================================>]276.5kB/276.5kB bb3f6df0f87c:Loadinglayer[==================================================>]82.94kB/82.94kB 6f34ead3cef7:Loadinglayer[==================================================>]946.7kB/946.7kB c21422dd15f6:Loadinglayer[==================================================>]1.97MB/1.97MB 940288517f4c:Loadinglayer[==================================================>]458.2kB/458.2kB 0c52f1af61f4:Loadinglayer[==================================================>]5.12kB/5.12kB 049e7edd48bc:Loadinglayer[==================================================>]1.57MB/1.57MB 73307245d702:Loadinglayer[==================================================>]5.632kB/5.632kB f109309d8ffc:Loadinglayer[==================================================>]2.175MB/2.175MB Loadedimage:xxxxxxxxxxxx.cn/platform/deploy-amd64:ubuntu.16.04 $dockerimages REPOSITORYTAGIMAGEIDCREATEDSIZE xxxxxxxxxxxx.cn/platform/deploy-amd64ubuntu.16.043ad94a76af5d3monthsago734MB
调试这边前台日志输出正常
... DEBU[2020-10-30T14:48:39.955963986+08:00]Appliedtarsha256:049e7edd48bc46e3dd5edf89c9caa8f0f7efbb41af403c5a54dd4f1008f604a7tod58edd0d97bb672ef40e82e45c1603ca3ceaad847d9b9fc7c9b0588087019649,size:1518278 DEBU[2020-10-30T14:48:39.960091040+08:00]Applyingtarin/data/kube/docker/overlay2/b044bd592ae800ed071208c6b2f650c5cbdc7452702f56a23b9b4ffe4236ac18/diffstorage-driver=overlay2 DEBU[2020-10-30T14:48:40.059510528+08:00]Appliedtarsha256:73307245d7021f9627ca0b2cbfeab3aac0b65abfd476f6ec26bb92c75892d7e2tob044bd592ae800ed071208c6b2f650c5cbdc7452702f56a23b9b4ffe4236ac18,size:3284 DEBU[2020-10-30T14:48:40.063040538+08:00]Applyingtarin/data/kube/docker/overlay2/03918b1d275aa284532b8b9c59ca158409416f904e13cc7084c598ed343e844f/diffstorage-driver=overlay2 DEBU[2020-10-30T14:48:40.148209852+08:00]Appliedtarsha256:f109309d8ffcb76589ad6389e80335d986b411c80122d990ab00a02a3a916e3eto03918b1d275aa284532b8b9c59ca158409416f904e13cc7084c598ed343e844f,size:2072803 ^CINFO[2020-10-30T14:48:55.593523177+08:00]Processingsignal'interrupt' DEBU[2020-10-30T14:48:55.593617229+08:00]daemonconfiguredwitha15secondsminimumshutdowntimeout DEBU[2020-10-30T14:48:55.593638628+08:00]startcleanshutdownofallcontainerswitha15secondstimeout... DEBU[2020-10-30T14:48:55.594074457+08:00]Unixsocket/run/docker/libnetwork/ebd15186e86385c48c4c5508d5f30eb83d5d74e56f09af5c82b6d6d9d63ec8b8.sockdoesn'texist.cannotacceptclientconnections DEBU[2020-10-30T14:48:55.594106623+08:00]Cleaningupoldmountid:start. INFO[2020-10-30T14:48:55.594157536+08:00]stoppingeventstreamfollowinggracefulshutdownerror=""module=libcontainerdnamespace=moby DEBU[2020-10-30T14:48:55.594343122+08:00]Cleaningupoldmountid:done. DEBU[2020-10-30T14:48:55.594501828+08:00]Cleanshutdownsucceeded INFO[2020-10-30T14:48:55.594520918+08:00]stoppinghealthcheckfollowinggracefulshutdownmodule=libcontainerd INFO[2020-10-30T14:48:55.594531978+08:00]stoppingeventstreamfollowinggracefulshutdownerror="contextcanceled"module=libcontainerdnamespace=plugins.moby DEBU[2020-10-30T14:48:55.594603119+08:00]receivedsignalsignal=terminated INFO[2020-10-30T14:48:55.594739890+08:00]pickfirstBalancer:HandleSubConnStateChange:0xc4201a61b0,TRANSIENT_FAILUREmodule=grpc INFO[2020-10-30T14:48:55.594751465+08:00]pickfirstBalancer:HandleSubConnStateChange:0xc4201a61b0,CONNECTINGmodule=grpc
看了下systemd的配置没啥特殊的，就很迷，不知道为啥前台运行就能导入，后面实在想不到怎么排查，就怀疑可能是socket问题，尝试用socat转发成tcp试试，结果还是不行(此处应该daemon那加tcp监听127试试，不应该通过socket，socat最终也是过的socket)
$socat-d-dTCP-LISTEN:2375,fork,bind=127.0.0.1UNIX:/var/run/docker.sock 2020/10/3017:39:58socat[5201]NlisteningonAF=2127.0.0.1:2375 ^[[C2020/10/3017:42:06socat[5201]NacceptingconnectionfromAF=2127.0.0.1:35370onAF=2127.0.0.1:2375 2020/10/3017:42:06socat[5201]Nforkedoffchildprocess11501 2020/10/3017:42:06socat[5201]NlisteningonAF=2127.0.0.1:2375 2020/10/3017:42:06socat[11501]NopeningconnectiontoAF=1"/var/run/docker.sock" 2020/10/3017:42:06socat[11501]NsuccessfullyconnectedfromlocaladdressAF=1"\0\0\0\0\0\0\x0D\x@7\xE9\xEC\x7E\0\0\0\x01\0\0\0\0" 2020/10/3017:42:06socat[11501]NstartingdatatransferloopwithFDs[6,6]and[5,5] 2020/10/3017:42:12socat[11501]Ewrite(5,0x55f098349920,8192):Brokenpipe 2020/10/3017:42:12socat[11501]Nexit(1) 2020/10/3017:42:12socat[5201]Nchilddied():handlingsignal17 $docker--log-leveldebug-Htcp://127.0.0.1:2375load-ikube/images/deploy.tar c7fe3ea715ef:Loadinglayer[==========================================>]286.9MB/340.3MB unexpectedEOF
最后耗了挺久的，当时忙，去看了下另一个客户的问题，然会回到这边，突发奇想的试试load其他镜像，结果可以。。。
$dockerload-ikube/images/pause_3.1.tar e17133b79956:Loadinglayer[==================================================>]744.4kB/744.4kB Loadedimage:mirrorgooglecontainers/pause-amd64:3.1 $dockerload-ikube/images/tiller_v2.16.1.tar 77cae8ab23bf:Loadinglayer[==================================================>]5.815MB/5.815MB 679105aa33fb:Loadinglayer[==================================================>]6.184MB/6.184MB 639eab5d05b1:Loadinglayer[==================================================>]40.46MB/40.46MB 87e5687e03f2:Loadinglayer[==================================================>]41.13MB/41.13MB Loadedimage:gcr.io/kubernetes-helm/tiller:v2.16.1 $dockerload-ikube/images/calico_v3.1.3.tar cd7100a72410:Loadinglayer[==================================================>]4.403MB/4.403MB ddc4cb8dae60:Loadinglayer[==================================================>]7.84MB/7.84MB 77087b8943a2:Loadinglayer[==================================================>]249.3kB/249.3kB c7227c83afaf:Loadinglayer[==================================================>]4.801MB/4.801MB 2e0e333a66b6:Loadinglayer[==================================================>]231.8MB/231.8MB Loadedimage:calico/node:v3.1.3 2580685bfb60:Loadinglayer[==================================================>]50.84MB/50.84MB Loadedimage:calico/kube-controllers:v3.1.3 0314be9edf00:Loadinglayer[==================================================>]1.36MB/1.36MB 15db169413e5:Loadinglayer[==================================================>]28.05MB/28.05MB 4252efcc5013:Loadinglayer[==================================================>]2.818MB/2.818MB 76cf2496cf36:Loadinglayer[==================================================>]3.03MB/3.03MB 91d3d3a16862:Loadinglayer[==================================================>]2.995MB/2.995MB 18a58488ba3b:Loadinglayer[==================================================>]3.474MB/3.474MB 8d8197f49da2:Loadinglayer[==================================================>]27.34MB/27.34MB 7520364e0845:Loadinglayer[==================================================>]9.216kB/9.216kB b9d064622bd6:Loadinglayer[==================================================>]2.56kB/2.56kB Loadedimage:calico/cni:v3.1.3
只有导入这个的时候才报错
$dockerload-ikube/images/deploy.tar errorduringconnect:Posthttp://%2Fvar%2Frun%2Fdocker.sock/v1.39/images/load?quiet=0:readunix@->/var/run/docker.sock:read:connectionresetbypeer
然后出包的机器上对比了下这个文件的校验值发现不对。。。。
总结
有个疑问就是为啥前台可以，其次文件损坏导入的时候dockerdaemon居然不刷任何日志直接connectionreset，新版本没测试过这种情况。
到此这篇关于docker错误的耗时排查过程记录的文章就介绍到这了,更多相关docker错误耗时排查内容请搜索毛票票以前的文章或继续浏览下面的相关文章希望大家以后多多支持毛票票！