Scale your Charmed MongoDB
This is part of the Charmed MongoDB 5 K8s Tutorial. Please refer to this page for more information and the overview of the content.
Adding and Removing units
Replication is a popular feature of MongoDB; replicas copy data making a database highly available. This means the application can provide self-healing capabilities in case one MongoDB replica fails.
! Disclaimer: this tutorial hosts all replicas are on the same machine, this should not be done in a production environment. To enable high availability in a production environment, replicas should be hosted on different servers to maintain isolation.
Add replicas
You can add two replicas to your deployed MongoDB application with:
juju add-unit mongodb-k8s -n 2
You can now watch the replica set add these replicas with: watch -n 1 -c juju status. It usually takes several minutes for the replicas to be added to the replica set. You’ll know that all three replicas are ready when watch -n 1 -c juju status --color reports:
Model Controller Cloud/Region Version SLA Timestamp
tutorial overlord microk8s/localhost 3.1.6 unsupported 15:58:05Z
App Version Status Scale Charm Channel Rev Address Exposed Message
mongodb-k8s active 3 mongodb-k8s 5/edge 27 10.152.183.20 no
Unit Workload Agent Address Ports Message
mongodb-k8s/0* active idle 10.1.42.137
mongodb-k8s/1 active idle 10.1.42.140
mongodb-k8s/2 active idle 10.1.42.141
You can trust that Charmed MongoDB K8S added these replicas correctly. But if you wanted to verify the replicas got added correctly you could connect to MongoDB via mongo command in the pod. Since your replica set has 2 additional hosts you will need to update the hosts in your URI. You can retrieve these host IPs with:
export HOST_IP_1="mongodb-k8s-1.mongodb-k8s-endpoints"
export HOST_IP_2="mongodb-k8s-2.mongodb-k8s-endpoints"
Then recreate the URI using your new hosts and reuse the username, password, database name, and replica set name that you previously used when you first connected to MongoDB:
export URI=mongodb://$DB_USERNAME:$DB_PASSWORD@$HOST_IP,$HOST_IP_1,$HOST_IP_2:27017/$DB_NAME?replicaSet=$REPL_SET_NAME
Now view and save the output of the URI:
echo $URI
Like earlier we access mongo by sshing into one of the Charmed MongoDB K8s hosts:
juju ssh --container=mongod mongodb-k8s/0
While ssh'd into mongodb-k8s/0 unit , we can access mongo using our new URI that we saved above.
mongo <saved URI>
Now type rs.status() and you should see your replica set configuration. It should look something like this:
{
"set" : "mongodb-k8s",
"date" : ISODate("2023-05-13T16:01:53.297Z"),
"myState" : 1,
"term" : NumberLong(1),
"syncSourceHost" : "",
"syncSourceId" : -1,
"heartbeatIntervalMillis" : NumberLong(2000),
"majorityVoteCount" : 2,
"writeMajorityCount" : 2,
"votingMembersCount" : 3,
"writableVotingMembersCount" : 3,
"optimes" : {
"lastCommittedOpTime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"lastCommittedWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"readConcernMajorityOpTime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"appliedOpTime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"durableOpTime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"lastAppliedWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastDurableWallTime" : ISODate("2023-05-13T16:01:51.536Z")
},
"lastStableRecoveryTimestamp" : Timestamp(1683993671, 1),
"electionCandidateMetrics" : {
"lastElectionReason" : "electionTimeout",
"lastElectionDate" : ISODate("2023-05-13T15:32:11.406Z"),
"electionTerm" : NumberLong(1),
"lastCommittedOpTimeAtElection" : {
"ts" : Timestamp(1683991931, 1),
"t" : NumberLong(-1)
},
"lastSeenOpTimeAtElection" : {
"ts" : Timestamp(1683991931, 1),
"t" : NumberLong(-1)
},
"numVotesNeeded" : 1,
"priorityAtElection" : 1,
"electionTimeoutMillis" : NumberLong(10000),
"newTermStartDate" : ISODate("2023-05-13T15:32:11.448Z"),
"wMajorityWriteAvailabilityDate" : ISODate("2023-05-13T15:32:11.469Z")
},
"members" : [
{
"_id" : 0,
"name" : "mongodb-k8s-0.mongodb-k8s-endpoints:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 1862,
"optime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"optimeDate" : ISODate("2023-05-13T16:01:51Z"),
"lastAppliedWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastDurableWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"syncSourceHost" : "",
"syncSourceId" : -1,
"infoMessage" : "",
"electionTime" : Timestamp(1683991931, 2),
"electionDate" : ISODate("2023-05-13T15:32:11Z"),
"configVersion" : 5,
"configTerm" : 1,
"self" : true,
"lastHeartbeatMessage" : ""
},
{
"_id" : 1,
"name" : "mongodb-k8s-1.mongodb-k8s-endpoints:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 269,
"optime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"optimeDurable" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"optimeDate" : ISODate("2023-05-13T16:01:51Z"),
"optimeDurableDate" : ISODate("2023-05-13T16:01:51Z"),
"lastAppliedWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastDurableWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastHeartbeat" : ISODate("2023-05-13T16:01:52.993Z"),
"lastHeartbeatRecv" : ISODate("2023-05-13T16:01:52.993Z"),
"pingMs" : NumberLong(0),
"lastHeartbeatMessage" : "",
"syncSourceHost" : "mongodb-k8s-0.mongodb-k8s-endpoints:27017",
"syncSourceId" : 0,
"infoMessage" : "",
"configVersion" : 5,
"configTerm" : 1
},
{
"_id" : 2,
"name" : "mongodb-k8s-2.mongodb-k8s-endpoints:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 86,
"optime" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"optimeDurable" : {
"ts" : Timestamp(1683993711, 1),
"t" : NumberLong(1)
},
"optimeDate" : ISODate("2023-05-13T16:01:51Z"),
"optimeDurableDate" : ISODate("2023-05-13T16:01:51Z"),
"lastAppliedWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastDurableWallTime" : ISODate("2023-05-13T16:01:51.536Z"),
"lastHeartbeat" : ISODate("2023-05-13T16:01:52.992Z"),
"lastHeartbeatRecv" : ISODate("2023-05-13T16:01:51.978Z"),
"pingMs" : NumberLong(0),
"lastHeartbeatMessage" : "",
"syncSourceHost" : "mongodb-k8s-1.mongodb-k8s-endpoints:27017",
"syncSourceId" : 1,
"infoMessage" : "",
"configVersion" : 5,
"configTerm" : 1
}
],
"ok" : 1,
"$clusterTime" : {
"clusterTime" : Timestamp(1683993711, 1),
"signature" : {
"hash" : BinData(0,"DhUrEN4lw0ixgUSIb76By1yU2BA="),
"keyId" : NumberLong("7232690270372888581")
}
},
"operationTime" : Timestamp(1683993711, 1)
}
Now exit the MongoDB shell by typing:
exit
Now you should be back in the host of the Charmed MongoDB K8s’ unit mongodb-k8s/0. To exit this host type:
exit
You should now be in the shell you started where you can interact with Juju and Kubernetes (MicroK8s).
Remove replicas
Removing a unit from the application, scales the replicas down. Before we scale down the replicas, list all the units with juju status, here you will see three units mongodb-k8s/0, mongodb-k8s/1, and mongodb-k8s/2. Each of these units hosts a MongoDB replica.
Because Kubernetes models does not support removing named units, we will specify number of units for application using --num-units
and we will specify the number of units to be removed.
juju remove-unit mongodb-k8s --num-units 1
You’ll know that the replica was successfully removed when juju status --watch 1s reports:
Model Controller Cloud/Region Version SLA Timestamp
tutorial overlord microk8s/localhost 3.1.6 unsupported 16:12:55Z
App Version Status Scale Charm Channel Rev Address Exposed Message
mongodb-k8s active 2 mongodb-k8s 5/edge 27 10.152.183.20 no
Unit Workload Agent Address Ports Message
mongodb-k8s/0* active idle 10.1.42.137
mongodb-k8s/1 active idle 10.1.42.140
mongodb-k8s/2 terminated lost 10.1.42.141 unit stopped by the cloud
You can check the status of pod in K8s to see that a pod was removed by using command microk8s.kubectl get pods --namespace=tutorial
NAME READY STATUS RESTARTS AGE
modeloperator-69977f8c44-zh7pc 1/1 Running 0 52m
mongodb-k8s-0 2/2 Running 0 49m
mongodb-k8s-1 2/2 Running 0 19m
As previously mentioned you can trust that Charmed MongoDB K8s removed this replica correctly. This can be checked by verifying that the new URI where the removed host has been excluded works properly.