Configuring the Crunchy Data PostgreSQL operator in a Digital Ocean managed kubernetes cluster

11/5/2020

I am having trouble configuring the Crunchy Data PostgreSQL operator in my Digital Ocean managed kubernetes cluster. Per their official installation/troubleshooting guide, I changed the default storage classes in the provided manifest file to do-block-storage and I've tried toggling the disable_fsgroup value, all to no avail. I'm getting the following output from running kubectl describe... on the operator pod:

Events:
  Type     Reason     Age                From                              Message
  ----     ------     ----               ----                              -------
  Normal   Scheduled  <unknown>                                            Successfully assigned pgo/postgres-operator-697fd6dbb6-n764r to test-dev-pool-35jcv
  Normal   Started    69s                kubelet, test-dev-pool-35jcv  Started container event
  Normal   Created    69s                kubelet, test-dev-pool-35jcv  Created container event
  Normal   Pulled     69s                kubelet, test-dev-pool-35jcv  Container image "registry.developers.crunchydata.com/crunchydata/pgo-event:centos7-4.5.0" already present on machine
  Normal   Started    68s (x2 over 69s)  kubelet, test-dev-pool-35jcv  Started container scheduler
  Normal   Created    68s (x2 over 69s)  kubelet, test-dev-pool-35jcv  Created container scheduler
  Normal   Pulled     68s (x2 over 69s)  kubelet, test-dev-pool-35jcv  Container image "registry.developers.crunchydata.com/crunchydata/pgo-scheduler:centos7-4.5.0" already present on machine
  Normal   Started    64s (x2 over 69s)  kubelet, test-dev-pool-35jcv  Started container operator
  Normal   Created    64s (x2 over 70s)  kubelet, test-dev-pool-35jcv  Created container operator
  Normal   Pulled     64s (x2 over 70s)  kubelet, test-dev-pool-35jcv  Container image "registry.developers.crunchydata.com/crunchydata/postgres-operator:centos7-4.5.0" already present on machine
  Normal   Started    64s (x2 over 70s)  kubelet, test-dev-pool-35jcv  Started container apiserver
  Normal   Created    64s (x2 over 70s)  kubelet, test-dev-pool-35jcv  Created container apiserver
  Normal   Pulled     64s (x2 over 70s)  kubelet, test-dev-pool-35jcv  Container image "registry.developers.crunchydata.com/crunchydata/pgo-apiserver:centos7-4.5.0" already present on machine
  Warning  BackOff    63s (x4 over 67s)  kubelet, test-dev-pool-35jcv  Back-off restarting failed container

Any ideas?

Edit: Solved! I was specifying the default storage incorrectly. The proper edits are

- name: BACKREST_STORAGE
value: "digitalocean"
- name: BACKUP_STORAGE
value: "digitalocean"
- name: PRIMARY_STORAGE
value: "digitalocean"
- name: REPLICA_STORAGE
value: "digitalocean"

- name: STORAGE5_NAME
value: "digitalocean"
- name: STORAGE5_ACCESS_MODE
value: "ReadWriteOnce"
- name: STORAGE5_SIZE
value: "1Gi"
- name: STORAGE5_TYPE
value: "dynamic"
- name: STORAGE5_CLASS
value: "do-block-storage"

See this GitHub issue for how to correctly format the file for DO.

-- Ender
digital-ocean
kubernetes
postgresql

0 Answers