1

Our Spinnaker is deployed on Ubuntu 18, Spinnaker version is 1.20.3. The only way we can view the logs is to run journalctl -u $microservice there are no logs on /var/log/spinnaker.

Is this normal?

erlchamp
  • 45
  • 4

2 Answers2

1

Yes. The preferred way of installation for Spinnaker is in Kubernetes. A quick and easy way for you to get started and easily migrate is to backup all you config with halyard, export the pipelines as json and run Minnaker in any Ubuntu 18 Compute box

Then import your old spinnaker data and pipelines.

The Ubuntu18 debian deploy flavor that you are running could be useful to debug cloud driver issues or for development purposes.

I suggest that you perform the migration to a Kubernetes cluster.

Andre Leon Rangel
  • 1,659
  • 1
  • 16
  • 28
  • 1
    Unfortunately Kubernetes clusters are far more expensive to run than a VM, and hosting costs are an issue when you work for a small company, but yes, you are correct in that Kubernetes is the preferred method of installation. – Ashley Kleynhans Aug 18 '22 at 20:34
  • 1
    @AshleyKleynhans you can run minnaker in a VM or powerful EC2 instance and you will save some costs – Andre Leon Rangel Aug 21 '22 at 23:46
0

The reason why none of the Spinnaker microservices output any logs to their log file directories in /var/log/spinnaker is because the preferred method of installation for Spinnaker is to use Kubernetes.

If the microservices were to create log files in /var/log/spinnaker, there is a good chance that the Kubernetes pods would die due to running out of storage, hence they all output their logs to STDOUT, and can be retrieved from Kubernetes by running:

kubectl -n spinnaker logs POD_NAME  > my_logfile_name.log

If you prefer to run Spinnaker on a VM rather than in Kubernetes and want to enable the log files so that you can debug a specific issue instead of using journalctl, you can edit the systemd service file for the particular microservice, for example Clouddriver, and add the following line in the [Service] section:

StandardOutput=append:/var/log/spinnaker/clouddriver/clouddriver.log

Then you reload the systemctl daemon and restart the service and it will then output its logs to the specified log file instead of STDOUT, for example:

sudo systemctl daemon-reload
sudo systemctl restart clouddriver.service
Ashley Kleynhans
  • 354
  • 2
  • 13