DE 

//Cloudogu EcoSystem Docs

  1. docs.cloudogu.com
  2. User manual
  3. Automatic Migration
  4. How-To Guides
  5. Troubleshooting
Getting Started
User manual
Developer documentation

Troubleshooting

Importer

Volumes are created too small

  • Increase the VolumeIncreaseFactor (etcdctl set /config/ces-exporter/volumeIncreaseFactor <Factor>) to > 1 in the exporter

Change the migration time

  • To change the migration time, either the Helm values in the Component-CR or the ConfigMap ces-importer-config must be adjusted directly.

    • The ces-importer pod must always be restarted afterward.

      kubectl rollout restart deployment/ces-importer -n ecosystem

Dogus hang in the starting state and prevent a new migration

  • Scale all dogu deployments to 0 and restart the migration
  • As of v3.7.0 of the Dogu operator, this problem has been fixed and can no longer occur

Changed configuration is not adopted

After changing the configuration, the importer must always be restarted so that the new configuration is adopted.

kubectl rollout restart deployment/ces-importer -n ecosystem

Error in the final migration

If an error occurs in the final migration, the following steps must be carried out:

  1. Configure a new time for the final migration

    • config/migration/finalSchedule

    • The importer must then be restarted!

      kubectl rollout restart deployment/ces-importer -n ecosystem
  2. ensure that the source instance is still accessible

Crash of the importer

After a crash of the importer, all already stopped Dogus must be restarted manually (via the Dogu-CR). Otherwise, they can no longer be stopped by the Dogu operator and the migration will not be successful.

Dogus hang in pending state after migration

In CES-VM only the memory limit could be set. If a memory limit is set in Kubernetes but no memory request, the memory request is automatically set to the limit. This can lead to Dogus requesting too much RAM and remaining in the pending state. If this is the case, either the memory request should be adjusted or the performance should be increased.

Dogus no longer start after a failed migration

After a failed migration, some Dogus may not be able to start anymore. This can happen if the configuration of a Dogu has already been transferred but the data has not. The next successful migration run should resolve this problem.

Exporter

Show logs

CES-MN
kubectl logs deployments/ces-exporter -n ecosystem
CES-VM
docker logs -f ces-exporter

Check API

To check the exporter's API for functionality, curl can be used, for example.

curl -l -H "X-CES-EXPORTER-API-KEY: <myApiKey>" <exporterFqdn>/ces-exporter/export-mode

Check SSH configuration

The following command can be used to check the SSH connection to the exporter independently of the importer:

ssh -p 7022 -l root -i ./<path/to/private/key> -o StrictHostKeyChecking=no -o BatchMode=yes <host-ip>

Check rsync

The migration is carried out by the rsync program. To check whether rsync is working correctly, the following command can be used:

rsync -avhz --delete --sparse --stats -e ssh -p 7022 -l root -i /.ssh/privateKey -o StrictHostKeyChecking=no -o BatchMode=yes <host-ip>:</path/on/exporting/machine/> </path/on/importing/machine/>

Attention: Rsync migrates data from the source system to the specified folder and deletes data in the target folder. It is best to use an empty folder for the test.