问题
通过DataGrip连接的MongoDB节点,之前可以新增或更新数据。某天突然不能新增或更新数据,报错信息如下:
具体的报错信息:
Command failed with error 10107 (NotMaster): 'not master' on server 10.19.21.11:30386. The full response is {"operationTime": {"$timestamp": {"t": 1677831692, "i": 1}}, "ok": 0.0, "errmsg": "not master", "code": 10107, "codeName": "NotMaster", "$clusterTime": {"clusterTime": {"$timestamp": {"t": 1677831692, "i": 1}}, "signature": {"hash": {"$binary": {"base64": "lHLZvy4mhv1/eWxsjPzI4Fo8fwg=", "subType": "00"}}, "keyId": 7191419147941576705}}}
排查
通过排查,得知k8s是1master + 3node节点,通过k8s安装部署的MongoDB副本集,一主两从:
[root@k8s-master01 ~]# kubectl get nodes -o wide
NAME STATUS ROLES AGE VERSION INTERNAL-IP EXTERNAL-IP OS-IMAGE KERNEL-VERSION CONTAINER-RUNTIME
k8s-master01 Ready master 3y153d v1.15.1 10.19.1.123 <none> CentOS Linux 7 (Core) 5.1.0-1.el7.elrepo.x86_64 docker://18.9.6
k8s-node01 Ready <none> 2y66d v1.15.1 10.19.2.12 <none> CentOS Linux 7 (Core) 3.10.0-1062.el7.x86_64 docker://18.9.6
k8s-node02 Ready <none> 3y153d v1.15.1 10.19.2.11 <none> CentOS Linux 7 (Core) 3.10.0-1062.el7.x86_64 docker://18.9.6
k8s-node03 Ready <none> 3y153d v1.15.1 10.19.2.13 <none> CentOS Linux 7 (Core) 3.10.0-1062.el7.x86_64 docker://18.9.6
[root@k8s-master01 ~]# kubectl get pod -n aba-stg-public | grep mongo
mongodb-replicaset-0 1/1 Running 6 353d
mongodb-replicaset-1 1/1 Running 3 263d
mongodb-replicaset-2 1/1 Running 3 262d
进入指定命名空间下的指定容器:
kubectl exec -it -n <ns> <pod> /bin/bash
执行mong
命令,即可连接mongod服务,执行rs.status()
命令查看Replica Set的状态:
报错:there are no users authenticated。
解决方法:
因为是在 admin这个数据库下添加的账号,故而需要在 admin 库下输入正确的用户名和密码进行验证:
use admin
db.auth('root','root')
再次执行命令rs.status();
即可得到正常输出:
{
"set" : "rs0",
"date" : ISODate("2023-03-10T14:08:01.694Z"),
"myState" : 1,
"term" : NumberLong(117),
"syncingTo" : "",
"syncSourceHost" : "",
"syncSourceId" : -1,
"heartbeatIntervalMillis" : NumberLong(2000),
"optimes" : {
"lastCommittedOpTime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"readConcernMajorityOpTime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"appliedOpTime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"durableOpTime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
}
},
"members" : [
{
"_id" : 0,
"name" : "mongodb-replicaset-0.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 1226572,
"optime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"optimeDurable" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"optimeDate" : ISODate("2023-03-10T14:07:57Z"),
"optimeDurableDate" : ISODate("2023-03-10T14:07:57Z"),
"lastHeartbeat" : ISODate("2023-03-10T14:08:00.747Z"),
"lastHeartbeatRecv" : ISODate("2023-03-10T14:08:00.736Z"),
"pingMs" : NumberLong(0),
"lastHeartbeatMessage" : "",
"syncingTo" : "mongodb-replicaset-2.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"syncSourceHost" : "mongodb-replicaset-2.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"syncSourceId" : 2,
"infoMessage" : "",
"configVersion" : 3
},
{
"_id" : 1,
"name" : "mongodb-replicaset-1.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 1226592,
"optime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"optimeDate" : ISODate("2023-03-10T14:07:57Z"),
"syncingTo" : "",
"syncSourceHost" : "",
"syncSourceId" : -1,
"infoMessage" : "",
"electionTime" : Timestamp(1677230706, 1),
"electionDate" : ISODate("2023-02-24T09:25:06Z"),
"configVersion" : 3,
"self" : true,
"lastHeartbeatMessage" : ""
},
{
"_id" : 2,
"name" : "mongodb-replicaset-2.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 1226585,
"optime" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"optimeDurable" : {
"ts" : Timestamp(1678457277, 1),
"t" : NumberLong(117)
},
"optimeDate" : ISODate("2023-03-10T14:07:57Z"),
"optimeDurableDate" : ISODate("2023-03-10T14:07:57Z"),
"lastHeartbeat" : ISODate("2023-03-10T14:08:00.747Z"),
"lastHeartbeatRecv" : ISODate("2023-03-10T14:08:00.736Z"),
"pingMs" : NumberLong(0),
"lastHeartbeatMessage" : "",
"syncingTo" : "mongodb-replicaset-1.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"syncSourceHost" : "mongodb-replicaset-1.mongodb-replicaset.aba-stg-public.svc.cluster.local:27017",
"syncSourceId" : 1,
"infoMessage" : "",
"configVersion" : 3
}
],
"ok" : 1,
"operationTime" : Timestamp(1678457277, 1),
"$clusterTime" : {
"clusterTime" : Timestamp(1678457277, 1),
"signature" : {
"hash" : BinData(0,"Er9WM7g3xmePUgTZwjC4dWIJ++w="),
"keyId" : NumberLong("7191419147941576705")
}
}
}
health为1表明服务器正常,0表明服务器down
state为1表明是Primary,2表明是Secondary,3是Recovering,7是Arbiter,8是Down
另外执行db.isMaster()
可返回副本集节点状态信息。
排查后得知,MongoDB副本集发生Master选举,现在节点1变成Master,之前节点0才是Master。
并且发现,在命令行执行新增或更新命令:
db.getSiblingDB("sms").getCollection("aba_sms_history").updateOne({_id: new ObjectId("605449f67ed85700083fcdec")}, {"$set": {channel: "ALIYUN"}})
在Master节点,是可以执行成功的:
在从节点执行失败:
也就是说副本集状态是正常的?
DataGrip复制数据源,分别对应3个节点,之前11这个节点是Master节点:
现在Master节点变成12,但是也执行失败:
另外13节点不出意料也是失败:
总结
在pod里执行mongo命令连接mongod server成功,然后执行db.isMaster()
,或rs.status()
,打印信息看起来都是正常的。节点1之前是secondary的节点现在变成primary,并且只有primary节点可以写入数据,secondary节点写入数据时报错not master。所以副本集状态是正常的。
但是通过DataGrip连接这3个节点,所有节点都报错not master。
解决
[root@k8s-master01 ~]# kubectl get svc -A -o wide | grep mongo
aba-stg-public mongodb-replicaset ClusterIP None <none> 27017/TCP 2y127d app=mongodb-replicaset,release=mongodb-replicaset
aba-stg-public mongodb-replicaset-client ClusterIP None <none> 27017/TCP 2y127d app=mongodb-replicaset,release=mongodb-replicaset
aba-stg-public mongodb-replicaset-service NodePort 10.104.187.101 <none> 27017:30386/TCP 2y128d statefulset.kubernetes.io/pod-name=mongodb-replicaset-1
可知其中mongodb-replicaset-service
是提供服务的svc,我们查看一下这个svc:
[root@k8s-master01 ~]# kubectl describe svc mongodb-replicaset-service -n aba-stg-public
Name: mongodb-replicaset-service
Namespace: aba-stg-public
Labels: app=mongodb-replicaset
Annotations: field.cattle.io/publicEndpoints:
[{"addresses":["10.19.2.11"],"port":30386,"protocol":"TCP","serviceName":"aba-stg-public:mongodb-replicaset-service","allNodes":true}]
kubectl.kubernetes.io/last-applied-configuration:
{"apiVersion":"v1","kind":"Service","metadata":{"annotations":{},"labels":{"app":"mongodb-replicaset"},"name":"mongodb-replicaset-service"...
Selector: statefulset.kubernetes.io/pod-name=mongodb-replicaset-0
Type: NodePort
IP: 10.104.187.101
Port: mongodb-replicaset 27017/TCP
TargetPort: 27017/TCP
NodePort: mongodb-replicaset 30386/TCP
Endpoints: 10.244.86.172:27017
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
发现其Selector不正确,修改svc:kubectl edit svc mongodb-replicaset-service -n aba-stg-public
。
理论基础:kubectl管控(操作)的所有对象都是一个个yml文件,编辑对象就是编辑一个yml文件,也就是说kubectl edit
命令实际上会调用类似vi
或vim
这些编辑命令,进入编辑模式,然后和操作vi
命令一样,保存退出,kubectl会自动检查yml文件格式是否有语法问题,没有语法问题,会尝试自动重启svc。如果启动失败则说明编辑后的yml文件有误。
如果没有做任何修改,则会提示:Edit cancelled, no changes made.
另外,我们使用Rancher来管理k8s集群资源。也就是说绝对大多数命令行操作都可以通过Rancher来实现。
上述操作步骤对应页面截图:
- 先切换到服务发现Tab页:搜索框搜索mongo:
如上截图所示,会看到3个svc,我们需要操作的是第三个,页面上能看到选择器,即Selector。
注:上面截图的选择器已经是修改后的正确的目标。 - 点击右侧的三个点,然后点击升级,即会触发
kubectl edit
命令:
- 如下图所示,修改标签值,点击保存即可
后续
关于kubect 标签Selector,请扩展阅读学习。