K8S集群灾备环境部署

etcd是kubernetes集群极为重要的一块服务，存储了kubernetes集群所有的数据信息，如Namespace、Pod、Service、路由等状态信息。如果etcd集群发生灾难或者 etcd 集群数据丢失，都会影响k8s集群数据的恢复。因此，通过备份etcd数据来实现kubernetes集群的灾备环境十分重要。

 

一、etcd集群备份

etcd不同版本的 etcdctl 命令不一样，但大致差不多，这里备份使用 napshot save进行快照备份。

需要注意几点：

• 备份操作在etcd集群的其中一个节点执行就可以。
• 这里使用的是etcd v3的api，因为从 k8s 1.13 开始，k8s不再支持 v2 版本的 etcd，即k8s的集群数据都存在了v3版本的etcd中。故备份的数据也只备份了使用v3添加的etcd数据，v2添加的etcd数据是没有做备份的。
• 本案例使用的是二进制部署的k8s v1.18.6 + Calico 容器环境（下面命令中的"ETCDCTL_API=3 etcdctl" 等同于 "etcdctl"）

 

1）开始备份之前，先来查看下etcd数据

etcd 数据目录
[root@k8s-master01 ~]# cat /opt/k8s/bin/environment.sh |grep "ETCD_DATA_DIR="
export ETCD_DATA_DIR="/data/k8s/etcd/data"
     
etcd WAL 目录
[root@k8s-master01 ~]# cat /opt/k8s/bin/environment.sh |grep "ETCD_WAL_DIR="
export ETCD_WAL_DIR="/data/k8s/etcd/wal"

[root@k8s-master01 ~]# ls /data/k8s/etcd/data/
member
[root@k8s-master01 ~]# ls /data/k8s/etcd/data/member/
snap
[root@k8s-master01 ~]# ls /data/k8s/etcd/wal/
0000000000000000-0000000000000000.wal  0.tmp

　　

2）执行etcd集群数据备份

在etcd集群的其中一个节点执行备份操作，然后将备份文件拷贝到其他节点上。

 

先在etcd集群的每个节点上创建备份目录

# mkdir -p /data/etcd_backup_dir

在etcd集群其中个一个节点（这里在k8s-master01）上执行备份：

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl --cacert=/etc/kubernetes/cert/ca.pem --cert=/etc/etcd/cert/etcd.pem --key=/etc/etcd/cert/etcd-key.pem --endpoints=https://172.16.60.231:2379 snapshot save /data/etcd_backup_dir/etcd-snapshot-`date +%Y%m%d`.db

将备份文件拷贝到其他的etcd节点

[root@k8s-master01 ~]# rsync -e "ssh -p22" -avpgolr /data/etcd_backup_dir/etcd-snapshot-20200820.db root@k8s-master02:/data/etcd_backup_dir/
[root@k8s-master01 ~]# rsync -e "ssh -p22" -avpgolr /data/etcd_backup_dir/etcd-snapshot-20200820.db root@k8s-master03:/data/etcd_backup_dir/

　　

可以将上面k8s-master01节点的etcd备份命令放在脚本里，结合crontab进行定时备份：

[root@k8s-master01 ~]# cat /data/etcd_backup_dir/etcd_backup.sh
#!/usr/bin/bash

date;
CACERT="/etc/kubernetes/cert/ca.pem"
CERT="/etc/etcd/cert/etcd.pem"
EKY="/etc/etcd/cert/etcd-key.pem"
ENDPOINTS="172.16.60.231:2379"

ETCDCTL_API=3 /opt/k8s/bin/etcdctl 
--cacert="${CACERT}" --cert="${CERT}" --key="${EKY}" 
--endpoints=${ENDPOINTS} 
snapshot save /data/etcd_backup_dir/etcd-snapshot-`date +%Y%m%d`.db

# 备份保留30天
find /data/etcd_backup_dir/ -name "*.db" -mtime +30 -exec rm -f {} ;

# 同步到其他两个etcd节点
/bin/rsync -e "ssh -p5522" -avpgolr --delete /data/etcd_backup_dir/ root@k8s-master02:/data/etcd_backup_dir/
/bin/rsync -e "ssh -p5522" -avpgolr --delete /data/etcd_backup_dir/ root@k8s-master03:/data/etcd_backup_dir/

　　

设置crontab定时备份任务，每天凌晨5点执行备份：

[root@k8s-master01 ~]# chmod 755 /data/etcd_backup_dir/etcd_backup.sh
[root@k8s-master01 ~]# crontab -l
#etcd集群数据备份
0 5 * * * /bin/bash -x /data/etcd_backup_dir/etcd_backup.sh > /dev/null 2>&1

　　

二、etcd集群恢复
etcd集群备份操作只需要在其中的一个etcd节点上完成，然后将备份文件拷贝到其他节点。
但etcd集群恢复操作必须要所有的etcd节点上完成！

1）模拟etcd集群数据丢失
删除三个etcd集群节点的data数据 (或者直接删除data目录)

# rm -rf /data/k8s/etcd/data/*

查看k8s集群状态：

[root@k8s-master01 ~]# kubectl get cs
NAME                 STATUS      MESSAGE                                                                                           ERROR
etcd-2               Unhealthy   Get https://172.16.60.233:2379/health: dial tcp 172.16.60.233:2379: connect: connection refused
etcd-1               Unhealthy   Get https://172.16.60.232:2379/health: dial tcp 172.16.60.232:2379: connect: connection refused
etcd-0               Unhealthy   Get https://172.16.60.231:2379/health: dial tcp 172.16.60.231:2379: connect: connection refused
scheduler            Healthy     ok
controller-manager   Healthy     ok

　　

由于此时etcd集群的三个节点服务还在，过一会儿查看集群状态恢复正常：

[root@k8s-master01 ~]# kubectl get cs
NAME                 STATUS    MESSAGE             ERROR
controller-manager   Healthy   ok
scheduler            Healthy   ok
etcd-0               Healthy   {"health":"true"}
etcd-2               Healthy   {"health":"true"}
etcd-1               Healthy   {"health":"true"}

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl --endpoints="https://172.16.60.231:2379,https://172.16.60.232:2379,https://172.16.60.233:2379" --cert=/etc/etcd/cert/etcd.pem --key=/etc/etcd/cert/etcd-key.pem --cacert=/etc/kubernetes/cert/ca.pem endpoint health
https://172.16.60.231:2379 is healthy: successfully committed proposal: took = 9.918673ms
https://172.16.60.233:2379 is healthy: successfully committed proposal: took = 10.985279ms
https://172.16.60.232:2379 is healthy: successfully committed proposal: took = 13.422545ms

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl --endpoints="https://172.16.60.231:2379,https://172.16.60.232:2379,https://172.16.60.233:2379" --cert=/etc/etcd/cert/etcd.pem --key=/etc/etcd/cert/etcd-key.pem --cacert=/etc/kubernetes/cert/ca.pem member list --write-out=table
+------------------+---------+------------+----------------------------+----------------------------+------------+
|        ID        | STATUS  |    NAME    |         PEER ADDRS         |        CLIENT ADDRS        | IS LEARNER |
+------------------+---------+------------+----------------------------+----------------------------+------------+
| 1d1d7edbba38c293 | started | k8s-etcd03 | https://172.16.60.233:2380 | https://172.16.60.233:2379 |      false |
| 4c0cfad24e92e45f | started | k8s-etcd02 | https://172.16.60.232:2380 | https://172.16.60.232:2379 |      false |
| 79cf4f0a8c3da54b | started | k8s-etcd01 | https://172.16.60.231:2380 | https://172.16.60.231:2379 |      false |
+------------------+---------+------------+----------------------------+----------------------------+------------+

　　

如上发现，etcd集群三个节点的leader都是false，即没有选主。此时需要重启三个节点的etcd服务：

# systemctl restart etcd

　　

重启后，再次查看发现etcd集群已经选主成功，集群状态正常！

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl -w table --cacert=/etc/kubernetes/cert/ca.pem   --cert=/etc/etcd/cert/etcd.pem   --key=/etc/etcd/cert/etcd-key.pem   --endpoints="https://172.16.60.231:2379,https://172.16.60.232:2379,https://172.16.60.233:2379" endpoint status
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
|          ENDPOINT          |        ID        | VERSION | DB SIZE | IS LEADER | IS LEARNER | RAFT TERM | RAFT INDEX | RAFT APPLIED INDEX | ERRORS |
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
| https://172.16.60.231:2379 | 79cf4f0a8c3da54b |   3.4.9 |  1.6 MB |      true |      false |         5 |      24658 |              24658 |        |
| https://172.16.60.232:2379 | 4c0cfad24e92e45f |   3.4.9 |  1.6 MB |     false |      false |         5 |      24658 |              24658 |        |
| https://172.16.60.233:2379 | 1d1d7edbba38c293 |   3.4.9 |  1.7 MB |     false |      false |         5 |      24658 |              24658 |        |
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+

　　

但是，k8s集群数据其实已经丢失了。namespace命名空间下的pod等资源都没有了。此时就需要通过etcd集群备份文件来恢复，即通过上面的etcd集群快照文件恢复。

[root@k8s-master01 ~]# kubectl get ns
NAME              STATUS   AGE
default           Active   9m47s
kube-node-lease   Active   9m39s
kube-public       Active   9m39s
kube-system       Active   9m47s
[root@k8s-master01 ~]# kubectl get pods -n kube-system
No resources found in kube-system namespace.
[root@k8s-master01 ~]# kubectl get pods --all-namespaces
No resources found

　　

2）etcd集群数据恢复，即kubernetes集群数据恢复

在etcd数据恢复之前，先依次关闭所有master节点的kube-aposerver服务，所有etcd节点的etcd服务：

# systemctl stop kube-apiserver
# systemctl stop etcd

　　

特别注意：在进行etcd集群数据恢复之前，一定要先将所有etcd节点的data和wal旧工作目录删掉，这里指的是/data/k8s/etcd/data文件夹跟/data/k8s/etcd/wal文件夹，可能会导致恢复失败（恢复命令执行时报错数据目录已存在）。

# rm -rf /data/k8s/etcd/data && rm -rf /data/k8s/etcd/wal

　　

在每个etcd节点执行恢复操作：

172.16.60.231节点
-------------------------------------------------------
ETCDCTL_API=3 etcdctl 
--name=k8s-etcd01 
--endpoints="https://172.16.60.231:2379" 
--cert=/etc/etcd/cert/etcd.pem 
--key=/etc/etcd/cert/etcd-key.pem 
--cacert=/etc/kubernetes/cert/ca.pem 
--initial-cluster-token=etcd-cluster-0 
--initial-advertise-peer-urls=https://172.16.60.231:2380 
--initial-cluster=k8s-etcd01=https://172.16.60.231:2380,k8s-etcd02=https://172.16.60.232:2380,k8s-etcd03=https://192.168.137.233:2380 
--data-dir=/data/k8s/etcd/data 
--wal-dir=/data/k8s/etcd/wal 
snapshot restore /data/etcd_backup_dir/etcd-snapshot-20200820.db


172.16.60.232节点
-------------------------------------------------------
ETCDCTL_API=3 etcdctl 
--name=k8s-etcd02 
--endpoints="https://172.16.60.232:2379" 
--cert=/etc/etcd/cert/etcd.pem 
--key=/etc/etcd/cert/etcd-key.pem 
--cacert=/etc/kubernetes/cert/ca.pem 
--initial-cluster-token=etcd-cluster-0 
--initial-advertise-peer-urls=https://172.16.60.232:2380 
--initial-cluster=k8s-etcd01=https://172.16.60.231:2380,k8s-etcd02=https://172.16.60.232:2380,k8s-etcd03=https://192.168.137.233:2380 
--data-dir=/data/k8s/etcd/data 
--wal-dir=/data/k8s/etcd/wal 
snapshot restore /data/etcd_backup_dir/etcd-snapshot-20200820.db


192.168.137.233节点
-------------------------------------------------------
ETCDCTL_API=3 etcdctl 
--name=k8s-etcd03 
--endpoints="https://192.168.137.233:2379" 
--cert=/etc/etcd/cert/etcd.pem 
--key=/etc/etcd/cert/etcd-key.pem 
--cacert=/etc/kubernetes/cert/ca.pem 
--initial-cluster-token=etcd-cluster-0 
--initial-advertise-peer-urls=https://192.168.137.233:2380 
--initial-cluster=k8s-etcd01=https://172.16.60.231:2380,k8s-etcd02=https://172.16.60.232:2380,k8s-etcd03=https://192.168.137.233:2380 
--data-dir=/data/k8s/etcd/data 
--wal-dir=/data/k8s/etcd/wal 
snapshot restore /data/etcd_backup_dir/etcd-snapshot-20200820.db

　　

依次启动所有etcd节点的etcd服务：

# systemctl start etcd
# systemctl status etcd

　　

检查 ETCD 集群状态（如下，发现etcd集群里已经成功选主了）

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl --endpoints="https://172.16.60.231:2379,https://172.16.60.232:2379,https://172.16.60.233:2379" --cert=/etc/etcd/cert/etcd.pem --key=/etc/etcd/cert/etcd-key.pem --cacert=/etc/kubernetes/cert/ca.pem endpoint health
https://172.16.60.232:2379 is healthy: successfully committed proposal: took = 12.837393ms
https://172.16.60.233:2379 is healthy: successfully committed proposal: took = 13.306671ms
https://172.16.60.231:2379 is healthy: successfully committed proposal: took = 13.602805ms

[root@k8s-master01 ~]# ETCDCTL_API=3 etcdctl -w table --cacert=/etc/kubernetes/cert/ca.pem   --cert=/etc/etcd/cert/etcd.pem   --key=/etc/etcd/cert/etcd-key.pem   --endpoints="https://172.16.60.231:2379,https://172.16.60.232:2379,https://172.16.60.233:2379" endpoint status
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
|          ENDPOINT          |        ID        | VERSION | DB SIZE | IS LEADER | IS LEARNER | RAFT TERM | RAFT INDEX | RAFT APPLIED INDEX | ERRORS |
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
| https://172.16.60.231:2379 | 79cf4f0a8c3da54b |   3.4.9 |  9.0 MB |     false |      false |         2 |         13 |                 13 |        |
| https://172.16.60.232:2379 | 4c0cfad24e92e45f |   3.4.9 |  9.0 MB |      true |      false |         2 |         13 |                 13 |        |
| https://172.16.60.233:2379 | 5f70664d346a6ebd |   3.4.9 |  9.0 MB |     false |      false |         2 |         13 |                 13 |        |
+----------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+

　　

再依次启动所有master节点的kube-apiserver服务：

# systemctl start kube-apiserver
# systemctl status kube-apiserver

　　

查看kubernetes集群状态：

[root@k8s-master01 ~]# kubectl get cs
NAME                 STATUS      MESSAGE                                  ERROR
controller-manager   Healthy     ok
scheduler            Healthy     ok
etcd-2               Unhealthy   HTTP probe failed with statuscode: 503
etcd-1               Unhealthy   HTTP probe failed with statuscode: 503
etcd-0               Unhealthy   HTTP probe failed with statuscode: 503

由于etcd服务刚重启，需要多刷几次状态就会正常：
[root@k8s-master01 ~]# kubectl get cs
NAME                 STATUS    MESSAGE             ERROR
controller-manager   Healthy   ok
scheduler            Healthy   ok
etcd-2               Healthy   {"health":"true"}
etcd-0               Healthy   {"health":"true"}
etcd-1               Healthy   {"health":"true"}

　　

查看kubernetes的资源情况：

[root@k8s-master01 ~]# kubectl get ns
NAME              STATUS   AGE
default           Active   7d4h
kevin             Active   5d18h
kube-node-lease   Active   7d4h
kube-public       Active   7d4h
kube-system       Active   7d4h

[root@k8s-master01 ~]# kubectl get pods --all-namespaces
NAMESPACE     NAME                                       READY   STATUS              RESTARTS   AGE
default       dnsutils-ds-22q87                          0/1     ContainerCreating   171        7d3h
default       dnsutils-ds-bp8tm                          0/1     ContainerCreating   138        5d18h
default       dnsutils-ds-bzzqg                          0/1     ContainerCreating   138        5d18h
default       dnsutils-ds-jcvng                          1/1     Running             171        7d3h
default       dnsutils-ds-xrl2x                          0/1     ContainerCreating   138        5d18h
default       dnsutils-ds-zjg5l                          1/1     Running             0          7d3h
default       kevin-t-84cdd49d65-ck47f                   0/1     ContainerCreating   0          2d2h
default       nginx-ds-98rm2                             1/1     Running             2          7d3h
default       nginx-ds-bbx68                             1/1     Running             0          7d3h
default       nginx-ds-kfctv                             0/1     ContainerCreating   1          5d18h
default       nginx-ds-mdcd9                             0/1     ContainerCreating   1          5d18h
default       nginx-ds-ngqcm                             1/1     Running             0          7d3h
default       nginx-ds-tpcxs                             0/1     ContainerCreating   1          5d18h
kevin         nginx-ingress-controller-797ffb479-vrq6w   0/1     ContainerCreating   0          5d18h
kevin         test-nginx-7d4f96b486-qd4fl                0/1     ContainerCreating   0          2d1h
kevin         test-nginx-7d4f96b486-qfddd                0/1     Running             0          2d1h
kube-system   calico-kube-controllers-578894d4cd-9rp4c   1/1     Running             1          7d3h
kube-system   calico-node-d7wq8                          0/1     PodInitializing     1          7d3h

在etcd集群数据恢复后，pod容器也会慢慢恢复到running状态。至此，kubernetes整个集群已经通过etcd备份数据恢复了。

 

三、最后总结

Kubernetes 集群备份主要是备份 ETCD 集群。而恢复时，主要考虑恢复整个顺序：

停止kube-apiserver --> 停止ETCD --> 恢复数据 --> 启动ETCD --> 启动kube-apiserve

 

特别注意：

• 备份ETCD集群时，只需要备份一个ETCD数据，然后同步到其他节点上。
• 恢复ETCD数据时，拿其中一个节点的备份数据恢复即可。