Backup to S3 Failed “Error interacting with filesystem”

Problem: Backup to AWS S3 fails due to Error interacting with filesystem.

If you have a backup that is failing with an “Error interacting with filesystem” message, this can be caused by the S3 client buffering files.

Resolution:

You can change the setting of the TAMR_BACKUP_FS_EXTRA_CONFIG configuration variable to skip local file buffering.

Step 1. Run a get command to find the current setting for TAMR_BACKUP_FS_EXTRA_CONFIG:

<tamr-home-directory>/tamr/utils/unify-admin.sh config:get TAMR_BACKUP_FS_EXTRA_CONFIG

The result can be an empty {} value or an existing setting within the braces.

Step 2: Run a set command to add 'fs.s3a.fast.upload':'true' to the current setting. If configuring more than one value, use commas to separate them.

<tamr-home-directory>/tamr/utils/unify-admin.sh config:set TAMR_BACKUP_FS_EXTRA_CONFIG="{'fs.s3a.fast.upload':'true'}"

This setting prevents the S3 client from staging files locally before copying them to the remote S3 bucket. This avoids any disk space issues during staging.

Step 3: Restart Tamr Core and its dependencies.

For additional detail, see Setting Configuration Variables and Restarting Tamr Core.