本文档基于自建Kubernetes平台使用阿里云logtail日志插件，收集Kubernetes标准输出日志
以下是官方文档链接地址
官方安装logtail：https://help.aliyun.com/document_detail/157317.html
官方通过DaemonSet-控制台方式采集Kubernetes标准输出：https://help.aliyun.com/document_detail/66658.html
如何安装helm：https://helm.sh/docs/intro/install/
以下所有操作都在Kubernetes Master节点上执行
- 下载安装脚本，如下
wget http://logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com/kubernetes/alicloud-log-k8s-custom-install.sh;chmod 744 ./alicloud-log-k8s-custom-install.sh
- 编辑安装脚本，根据环境自定义
<root@PROD-K8S-CP1 ~># cat alicloud-log-k8s-custom-install.sh #!/bin/bash if [ $# -le 4 ] ; then echo "[Invalid Param], use sudo ./install-k8s-log.sh {your-project-suffix} {regionId} {aliuid} {accessKeyId} {accessKeySecret}" exit 1 fi # project环境变量，根据自己需求定义，最终在SLS生成的project名称为kubernetes-$1 project="kubernetes-"$1 regionId=$2 aliuid=$3 accessKeyId=$4 accessKeySecret=$5 helmPackageUrl="http://logtail-release-$regionId.oss-$regionId.aliyuncs.com/kubernetes/alibaba-cloud-log.tgz" wget $helmPackageUrl -O alibaba-cloud-log.tgz if [ $? != 0 ]; then echo "[FAIL] download alibaba-cloud-log.tgz from $helmPackageUrl failed" exit 1 fi helmV3=`helm version | grep v3 | wc -l` if [ $helmV3 -eq 0 ]; then echo "install alibaba-log-controller using helm v2 command." helm install alibaba-cloud-log.tgz --name alibaba-log-controller --set ProjectName=$project --set RegionId=$regionId --set InstallParam=$regionId"_internet" # 注意与上文的project是一样的 --set MachineGroupId="kubernetes-"$1 --set Endpoint=$regionId".log.aliyuncs.com" --set AlibabaCloudUserId=":"$aliuid --set AlibabaCloudK8SCluster="false" --set Privileged="true" --set AccessKeyId=$accessKeyId --set AccessKeySecret=$accessKeySecret --set LogtailImage.Repository="registry.$regionId.aliyuncs.com/log-service/logtail" --set ControllerImage.Repository="registry.$regionId.aliyuncs.com/log-service/alibabacloud-log-controller" else echo "install alibaba-log-controller using helm v3 command." helm install alibaba-log-controller alibaba-cloud-log.tgz --set ProjectName=$project --set RegionId=$regionId --set InstallParam=$regionId"_internet" # 注意与上文的project是一样的 --set MachineGroupId="kubernetes-"$1 --set Endpoint=$regionId".log.aliyuncs.com" --set AlibabaCloudUserId=":"$aliuid --set AlibabaCloudK8SCluster="false" --set Privileged="true" --set AccessKeyId=$accessKeyId --set AccessKeySecret=$accessKeySecret --set LogtailImage.Repository="registry.$regionId.aliyuncs.com/log-service/logtail" --set ControllerImage.Repository="registry.$regionId.aliyuncs.com/log-service/alibabacloud-log-controller" fi installRst=$? if [ $installRst -eq 0 ]; then echo [INFO] your k8s is using project : $project, region : $regionId, aliuid : $aliuid, accessKeyId : $accessKeyId echo "[SUCCESS] install helm package : alibaba-log-controller success." exit 0 else echo "[FAIL] install helm package failed, errno " $installRst exit 0 fi
- 安装helm
- 安装alibaba-log-controller，如下
sh ./alicloud-log-k8s-custom-install.sh {your-project-suffix} {region-id} {aliuid} {access-key-id} {access-key-secret} 具体带入的参数解释请参考官方文档 https://help.aliyun.com/document_detail/157317.html?spm=a2c4g.11186623.6.622.37f27cdfTStC5F#title-7qc-o6x-x8f 具体最终执行命令如下： sh ./alicloud-log-k8s-custom-install.sh non-production cn-hangzhou 1358359200279133 LTAI4FqfAcMq2fUts7kpP7VK SK(根据自己申请的SK填充) <root@PROD-K8S-CP1 ~># sh ./alicloud-log-k8s-custom-install.sh prod cn-hangzhou aliuid LTAI4FqfAcMq2fUts7kpP7VK SK --2021-08-31 17:14:53-- http://logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com/kubernetes/alibaba-cloud-log.tgz Resolving logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com (logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com)... 47.110.177.113 Connecting to logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com (logtail-release-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com)|47.110.177.113|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 3785 (3.7K) [application/octet-stream] Saving to: ‘alibaba-cloud-log.tgz’ 100%[========================================================================================================================================================================>] 3,785 --.-K/s in 0s 2021-08-31 17:14:53 (419 MB/s) - ‘alibaba-cloud-log.tgz’ saved [3785/3785] install alibaba-log-controller using helm v3 command. NAME: alibaba-log-controller LAST DEPLOYED: Tue Aug 31 17:14:53 2021 NAMESPACE: default STATUS: deployed REVISION: 1 TEST SUITE: None [INFO] your k8s is using project : kubernetes-prod, region : cn-hangzhou, aliuid : 1358359200279133, accessKeyId : LTAI4FqfAcMq2fUts7kpP7VK [SUCCESS] install helm package : alibaba-log-controller success.
- 查看helm安装状态
<root@PROD-K8S-CP1 ~># helm list NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION alibaba-log-controller default 1 2021-08-31 17:14:53.415754717 +0800 CST deployed alibaba-cloud-log-0.2.3 1.0 <root@PROD-K8S-CP1 ~># helm status alibaba-log-controller NAME: alibaba-log-controller LAST DEPLOYED: Tue Aug 31 17:14:53 2021 NAMESPACE: default STATUS: deployed REVISION: 1 TEST SUITE: None # 如果在安装过程中遇到问题，可以卸载重新安装 <root@PROD-K8S-CP1 ~># helm uninstall alibaba-log-controller release "alibaba-log-controller" uninstalled、
- 查看logtai启动状态
<root@PROD-K8S-CP1 ~># kubectl get po -n kube-system | grep logtail logtail-ds-6b67m 1/1 Running 0 6m59s logtail-ds-6r8bb 1/1 Running 0 7m13s logtail-ds-88w48 1/1 Running 0 7m28s logtail-ds-bsfkx 1/1 Running 0 9m21s logtail-ds-btlrd 1/1 Running 0 7m44s logtail-ds-dwvds 1/1 Running 0 8m24s logtail-ds-dxt9f 1/1 Running 0 8m16s logtail-ds-hjd9w 1/1 Running 0 8m37s logtail-ds-hqcsh 1/1 Running 0 9m8s logtail-ds-j2jt9 1/1 Running 0 7m7s logtail-ds-kmjmv 1/1 Running 0 8m52s logtail-ds-qcwjv 1/1 Running 0 9m14s logtail-ds-r4p55 1/1 Running 0 8m2s logtail-ds-ss52s 1/1 Running 0 7m56s logtail-ds-sxn7d 1/1 Running 0 8m32s logtail-ds-t4hbd 1/1 Running 0 6m42s logtail-ds-txzvt 1/1 Running 0 9m47s logtail-ds-zcpqs 1/1 Running 0 9m33s logtail-ds-zx5hb 1/1 Running 0 8m8s
- 修改logtail-ds资源限制及优先级
resources: limits: cpu: ‘2‘ memory: 1Gi requests: cpu: 40m memory: 64Mi ... priorityClassName: base-resource
- 调整alibaba-log-controller节点亲和性
# 资源限额 resources: limits: cpu: ‘1‘ memory: 256Mi requests: cpu: 10m memory: 64Mi ------------------------------------------------- # 节点亲和性 affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: kubernetes.io/resource operator: In values: - base ------------------------------------------------------- #污点配置 tolerations: - key: resource value: base effect: NoExecute ---------------------------------------------------------- #资源优先级 priorityClassName: base-resource