Setting up S3 storage on Kubernetes using Cloudian

0
Hi,We are setting up AKS deployment using the Connected Architecture.Dit anyone already use Cloudian S3 as Mendix File Storage Plan, and if yes which Sorage Type did you use (Minio/Ceph/...)?We tried many things but do not get it working. We keep getting an issue with creating a bucket, the name contains inderscores ("_") but our provider does nog accept bucket names with underscores.Thanx in advance!Ronnie
asked
1 answers
0

Hi Ronnie,

I haven't used Cloudian with Mendix myself, so I can't say which storage type is confirmed to work. Here are a few things that may help.


S3 bucket names can't contain underscores under the standard DNS-compliant naming rules (lowercase letters, numbers, hyphens, dots). Cloudian is correctly rejecting the name, so the fix is to avoid the Operator generating that name.

  1. Create the bucket yourself in Cloudian with a compliant name, for example myapp-prod-files, using hyphens instead of underscores.
  2. Use a storage plan that points to an existing bucket rather than one that creates the bucket for you. Then the Operator uses your bucket name and never has to generate one with underscores. Check the Storage Plans section of the Mendix on Kubernetes documentation for the exact plan type and required fields.
  3. Provide the S3 details in the plan: the Cloudian endpoint URL, the bucket name, and an access key and secret with list, read, write and delete permissions on that bucket.
  4. If it still fails, enable path-style addressing if the option is available, and check the Operator logs for the exact bucket name and error returned by Cloudian.

Kindly mark this as the accepted answer if it helps.

answered