etcdctl defrag & 剔除、添加etcd节点

零、准备工作

find / -name etcdctl
cp /var/lib/containerd/io.containerd.snapshotter.v1.overlayfs/snapshots/12/fs/usr/local/bin/etcdctl /usr/local/bin/etcdctl

alias ec="etcdctl --endpoints=https://127.0.0.1:2379 --cacert /etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key"

ec  endpoint status  member list --cluster

一、etcdctl defrag

ec defrag

ks exec -it $(ks get pod | grep etcd | awk 'NR==1 {print $1}') sh

alias ec="etcdctl --endpoints=https://127.0.0.1:2379 --cacert /etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key"
ec endpoint status --cluster -w table

ec defrag

ec endpoint status --cluster -w table

如果报错。。。

1、关闭etcd

mv /etc/kubernetes/manifests/etcd.yaml .

2、etcdctl defrag

etcdctl defrag --data-dir /opt/qfusion/etcd

3、启动etcd

mv etcd.yaml /etc/kubernetes/manifests/etcd.yaml

4、查看etcd服务

ec  endpoint status  member list --cluster

nerdctl ps -a |grep etcd

二、剔除etcd

1、member remove

etcdctl member remove <memberID>

2、移除配置

mv /etc/kubernetes/manifests/etcd.yaml .

3、移除etcd数据目录

mv /opt/qfusion/etcd /opt/qfusion/etcd_bak

三、添加etcd

1、在正常的节点上操作，假如etcd节点

ec member add 10-10-180-15 --peer-urls=https://10.10.180.15:2380

2、启动etcd

    - --initial-cluster=10-10-180-15=https://10.10.180.15:2380,10-10-180-8=https://10.10.180.8:2380,10-10-180-9=https://10.10.180.9:2380 #加入etcd集群的时候需要补全集群信息，否则etcd会报 member count is unequal 的错误而无法成功启动
    - --initial-cluster-state=existing # 以加入etcd集群的方式启动etcd

initial-cluster-state=new # 以新建etcd集群的方式启动etcd