We have deployed an on-premises dremio installation running on kubernetes. According to earlier documentation and values.yaml file, the distributed storage for version >21 does not support local:
The supported distributed storage types are: local (<21.0.0 only), aws, azure, gcp, or azureStorage.
local: (<21.0.0 only) Not recommended for production use. When using local, dist-caching is disabled.
aws: AWS S3, additional parameters required, see “aws” section.
azure: ADLS Gen 1, additional parameters required, see “azure” section.
azureStorage: Azure Storage Gen2, additional paramters required, see “azureStorage” section.
gcp: Google Cloud Storage, additional parameters required, see “gcp” section.
We have just found out that using azureStorage implies some costs, so we are trying to find a solution or setting that does not require cloud storage as distStorage.
I then see in the documentation for the latest version 24.2, that Netapp NAS is supported as distributed storage. So I’m a bit confused about the contradicting statements about distributed storage.
Can you help to clarify which is true? Does Dremio v24.2 support ‘local’ as distributed storage?
… maybe ‘Local’ is not the same as ‘NAS’?
I cannot find any documentation on this.
@humanSupafly Local for distributed storage is no longer supported
But local NAS is supported according to the documentation.
Can you give an example on how to configure this in a config.yaml file, given a (netapp) nfs share?
You mean here? https://github.com/dremio/dremio-cloud-tools/blob/master/charts/dremio_v2/values.yaml
Basically syntax is below
Will check internally on values.yaml and get back to you
thanks for looking into it.
yes, that’s the yaml file i’m trying to figure out how to configure.
As I understand, the documentation, refers to changes in the dremio.config file https://github.com/dremio/dremio-cloud-tools/blob/master/charts/dremio_v2/config/dremio.conf
Upon checking internally, currently the helm charts do not have it. When the K8’s charts were created, NFS/NAS was not considered as we have seen performance issue when it is not a performant one. The charts were intended for cloud deployments and hence only cloud storage dist options where considered
Oke, many thanks for checking it with your dev team.
I have configured it to use local S3 as distributed storage. Hope that is supported (and stabile).