constellation/docs/versioned_docs/version-2.10/workflows/troubleshooting.md

152 lines
6.1 KiB
Markdown
Raw Normal View History

# Troubleshooting
This section aids you in finding problems when working with Constellation.
## Common issues
### Issues with creating new clusters
When you create a new cluster, you should always use the [latest release](https://github.com/edgelesssys/constellation/releases/latest).
If something doesn't work, check out the [known issues](https://github.com/edgelesssys/constellation/issues?q=is%3Aopen+is%3Aissue+label%3A%22known+issue%22).
### Azure: Resource Providers can't be registered
On Azure, you may receive the following error when running `create` or `terminate` with limited IAM permissions:
```shell-session
Error: Error ensuring Resource Providers are registered.
Terraform automatically attempts to register the Resource Providers it supports to
ensure it's able to provision resources.
If you don't have permission to register Resource Providers you may wish to use the
"skip_provider_registration" flag in the Provider block to disable this functionality.
[...]
```
To continue, please ensure that the [required resource providers](../getting-started/install.md#required-permissions) have been registered in your subscription by your administrator.
Afterward, set `ARM_SKIP_PROVIDER_REGISTRATION=true` as an environment variable and either run `create` or `terminate` again.
For example:
```bash
ARM_SKIP_PROVIDER_REGISTRATION=true constellation create
```
Or alternatively, for `terminate`:
```bash
ARM_SKIP_PROVIDER_REGISTRATION=true constellation terminate
```
### Nodes fail to join with error `untrusted PCR value`
This error indicates that a node's [attestation statement](../architecture/attestation.md) contains measurements that don't match the trusted values expected by the [JoinService](../architecture/microservices.md#joinservice).
This may for example happen if the cloud provider updates the VM's firmware such that it influences the [runtime measurements](../architecture/attestation.md#runtime-measurements) in an unforeseen way.
A failed upgrade due to an erroneous attestation config can also cause this error.
You can change the expected measurements to resolve the failure.
:::caution
Attestation and trusted measurements are crucial for the security of your cluster.
Be extra careful when manually changing these settings.
When in doubt, check if the encountered [issue is known](https://github.com/edgelesssys/constellation/issues?q=is%3Aopen+is%3Aissue+label%3A%22known+issue%22) or [contact support](https://github.com/edgelesssys/constellation#support).
:::
:::tip
During an upgrade with modified attestation config, a backup of the current configuration is stored in the `join-config-backup` config map in the `kube-system` namespace. To restore the old attestation config after a failed upgrade, you can copy the attestation config from this resource, put it in your configuration file and retry the upgrade.
:::
You can use the `upgrade apply` command to change measurements of a running cluster:
1. Modify the `measurements` key in your local `constellation-conf.yaml` to the expected values.
2. Run `constellation upgrade apply`.
Keep in mind that running `upgrade apply` also applies any version changes from your config to the cluster.
You can run these commands to learn about the versions currently configured in the cluster:
- Kubernetes API server version: `kubectl get nodeversion constellation-version -o json -n kube-system | jq .spec.kubernetesClusterVersion`
- image version: `kubectl get nodeversion constellation-version -o json -n kube-system | jq .spec.imageVersion`
- microservices versions: `helm list --filter 'constellation-services' -n kube-system`
## Diagnosing issues
### Cloud logging
To provide information during early stages of a node's boot process, Constellation logs messages to the log systems of the cloud providers. Since these offerings **aren't** confidential, only generic information without any sensitive values is stored. This provides administrators with a high-level understanding of the current state of a node.
You can view this information in the following places:
<tabs groupId="csp">
<tabItem value="azure" label="Azure">
1. In your Azure subscription find the Constellation resource group.
2. Inside the resource group find the Application Insights resource called `constellation-insights-*`.
3. On the left-hand side go to `Logs`, which is located in the section `Monitoring`.
- Close the Queries page if it pops up.
5. In the query text field type in `traces`, and click `Run`.
To **find the disk UUIDs** use the following query: `traces | where message contains "Disk UUID"`
</tabItem>
<tabItem value="gcp" label="GCP">
1. Select the project that hosts Constellation.
2. Go to the `Compute Engine` service.
3. On the right-hand side of a VM entry select `More Actions` (a stacked ellipsis)
- Select `View logs`
To **find the disk UUIDs** use the following query: `resource.type="gce_instance" text_payload=~"Disk UUID:.*\n" logName=~".*/constellation-boot-log"`
:::info
Constellation uses the default bucket to store logs. Its [default retention period is 30 days](https://cloud.google.com/logging/quotas#logs_retention_periods).
:::
</tabItem>
<tabItem value="aws" label="AWS">
1. Open [AWS CloudWatch](https://console.aws.amazon.com/cloudwatch/home)
2. Select [Log Groups](https://console.aws.amazon.com/cloudwatch/home#logsV2:log-groups)
3. Select the log group that matches the name of your cluster.
4. Select the log stream for control or worker type nodes.
</tabItem>
</tabs>
### Node shell access
Debugging via a shell on a node is [directly supported by Kubernetes](https://kubernetes.io/docs/tasks/debug/debug-application/debug-running-pod/#node-shell-session).
1. Figure out which node to connect to:
```bash
kubectl get nodes
# or to see more information, such as IPs:
kubectl get nodes -o wide
```
2. Connect to the node:
```bash
kubectl debug node/constell-worker-xksa0-000000 -it --image=busybox
```
You will be presented with a prompt.
The nodes file system is mounted at `/host`.
3. Once finished, clean up the debug pod:
```bash
kubectl delete pod node-debugger-constell-worker-xksa0-000000-bjthj
```