Skip to main content
Version: 2.0

Recover your cluster

Recovery of a Constellation cluster means getting it back into a healthy state after too many concurrent node failures in the control plane. Reasons for an unhealthy cluster can vary from a power outage, or planned reboot, to migration of nodes and regions. Recovery events are rare, because Constellation is built for high availability and automatically and securely replaces failed nodes. When a node is replaced, Constellation's control plane first verifies the new node before it sends the node the cryptographic keys required to decrypt its state disk.

Constellation provides a recovery mechanism for cases where the control plane has failed and is unable to replace nodes. The constellation recover command connects to a node, establishes a secure connection using attested TLS, and provides that node with the key to decrypt its stateful disk and continue booting. This process has to be repeated until enough nodes are back running for establishing a member quorum for etcd and the Kubernetes state can be recovered.

Identify unhealthy clusters​

The first step to recovery is identifying when a cluster becomes unhealthy. Usually, this can be first observed when the Kubernetes API server becomes unresponsive.

You can check the health status of the nodes via the cloud service provider (CSP). Constellation provides logging information on the boot process and status via cloud logging. In the following, you'll find detailed descriptions for identifying clusters stuck in recovery for each CSP.

In the Azure portal, find the cluster's resource group. Inside the resource group, open the control plane Virtual machine scale set constellation-scale-set-controlplanes-<suffix>. On the left, go to Settings > Instances and check that enough members are in a Running state.

Second, check the boot logs of these Instances. In the scale set's Instances view, open the details page of the desired instance. On the left, go to Support + troubleshooting > Serial console.

In the serial console output, search for Waiting for decryption key. Similar output to the following means your node was restarted and needs to decrypt the state disk:

{"level":"INFO","ts":"2022-09-08T09:56:41Z","caller":"cmd/main.go:55","msg":"Starting disk-mapper","version":"2.0.0","cloudProvider":"azure"}
{"level":"INFO","ts":"2022-09-08T09:56:43Z","logger":"setupManager","caller":"setup/setup.go:72","msg":"Preparing existing state disk"}
{"level":"INFO","ts":"2022-09-08T09:56:43Z","logger":"recoveryServer","caller":"recoveryserver/server.go:59","msg":"Starting RecoveryServer"}
{"level":"INFO","ts":"2022-09-08T09:56:43Z","logger":"rejoinClient","caller":"rejoinclient/client.go:65","msg":"Starting RejoinClient"}

The node will then try to connect to the JoinService and obtain the decryption key. If this fails due to an unhealthy control plane, you will see log messages similar to the following:

{"level":"INFO","ts":"2022-09-08T09:56:43Z","logger":"rejoinClient","caller":"rejoinclient/client.go:77","msg":"Received list with JoinService endpoints","endpoints":["10.9.0.5:30090","10.9.0.6:30090"]}
{"level":"INFO","ts":"2022-09-08T09:56:43Z","logger":"rejoinClient","caller":"rejoinclient/client.go:96","msg":"Requesting rejoin ticket","endpoint":"10.9.0.5:30090"}
{"level":"WARN","ts":"2022-09-08T09:57:03Z","logger":"rejoinClient","caller":"rejoinclient/client.go:101","msg":"Failed to rejoin on endpoint","error":"rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing dial tcp 10.9.0.5:30090: i/o timeout\"","endpoint":"10.9.0.5:30090"}
{"level":"INFO","ts":"2022-09-08T09:57:03Z","logger":"rejoinClient","caller":"rejoinclient/client.go:96","msg":"Requesting rejoin ticket","endpoint":"10.9.0.6:30090"}
{"level":"WARN","ts":"2022-09-08T09:57:23Z","logger":"rejoinClient","caller":"rejoinclient/client.go:101","msg":"Failed to rejoin on endpoint","error":"rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing dial tcp 10.9.0.6:30090: i/o timeout\"","endpoint":"10.9.0.6:30090"}
{"level":"ERROR","ts":"2022-09-08T09:57:23Z","logger":"rejoinClient","caller":"rejoinclient/client.go:110","msg":"Failed to rejoin on all endpoints"}

This means that you have to recover the node manually. For this, you need its IP address, which you can obtain from the Overview page under Private IP address.

Recover a cluster​

The following process needs to be repeated until a member quorum for etcd is established. For example, assume you have 5 control-plane nodes in your cluster and 4 of them have been rebooted due to a maintenance downtime in the cloud environment. You have to run through the following process for 2 of these nodes and recover them manually to recover the quorum. From there, your cluster will auto heal the remaining 2 control-plane nodes and the rest of your cluster.

Recovering a node requires the following parameters:

Note that the recovery command needs to connect to the recovering nodes. Nodes only have private IP addresses in the VPC of the cluster, hence, the command needs to be issued from within the VPC network of the cluster. The easiest approach is to set up a jump host connected to the VPC network and perform the recovery from there.

Given these prerequisites a node can be recovered like this:

$ constellation recover -e 34.107.89.208 --master-secret constellation-mastersecret.json
Pushed recovery key.

In the serial console output of the node you'll see a similar output to the following:

{"level":"INFO","ts":"2022-09-08T10:26:59Z","logger":"recoveryServer","caller":"recoveryserver/server.go:93","msg":"Received recover call"}
{"level":"INFO","ts":"2022-09-08T10:26:59Z","logger":"recoveryServer","caller":"recoveryserver/server.go:125","msg":"Received state disk key and measurement secret, shutting down server"}
{"level":"INFO","ts":"2022-09-08T10:26:59Z","logger":"recoveryServer.gRPC","caller":"zap/server_interceptors.go:61","msg":"finished streaming call with code OK","grpc.start_time":"2022-09-08T10:26:59Z","system":"grpc","span.kind":"server","grpc.service":"recoverproto.API","grpc.method":"Recover","peer.address":"192.0.2.3:41752","grpc.code":"OK","grpc.time_ms":15.701}
{"level":"INFO","ts":"2022-09-08T10:27:13Z","logger":"rejoinClient","caller":"rejoinclient/client.go:87","msg":"RejoinClient stopped"}