Configuring Object Direct in WEKA

Prev Next

WEKA's Object Direct mount mode is used to land bulk or raw data directly on Wasabi Hot Cloud Storage, ahead of high-performance processing on WEKA NeuralMesh. Data writes to Wasabi with only brief staging on NeuralMesh's SSD tier. Reads for migration are served directly from Wasabi without promoting data to the performance tier.

Requirements

Configuration

Attach a Wasabi bucket using the following command.

weka fs tier s3 add wasabi-bucket --site=local --hostname=s3.wasabisys.com --auth-method=AWSSignature4 --port 443 --protocol HTTPS_UNVERIFIED --region=us-east-1 --bucket=<bucket name> --access-key-id=<access key> --secret-key=<secret key>

This configuration example uses Wasabi’s us-east-1 storage region. To use another Wasabi storage region, use the appropriate URL as described in Service URLs for Wasabi's Storage Regions.

Create a filesystem group using the command. (This example creates a filesystem group “archive” with 1 week retention and 1 hour demote schedule.)

weka fs group add archive --target-ssd-retention=1w --start-demote=1h

Create a filesystem using the following command. (This example create a filesystem named “wasabi-fs”.)

weka fs add wasabi-fs 1PB --fs-group archive --ssd-capacity 100TB --obs-name wasabi-bucket

To set up an SMB protocol share with Object Direct, use the following command. (This example creates a SMB share named “wasabi-smb”.)

weka smb share add wasabi-smb wasabi-fs --obs-direct

To set up a NFS protocol share with Object Direct, use the following command. (This example creates a NFS share named “wasabi-fs”.)

weka nfs permission add wasabi-fs archive --permission-type rw --obs-direct on

When data is copied into the shares, the data will be written into the SSD tier and immediately transferred to Wasabi. When data is retrieved, it is served directly from Wasabi and not cached on the SSD tier.