Data Import for Veeam Data Cloud for Salesforce

Import Data Preparation

Supported Storage

The exported CSV files must be uploaded to Amazon S3 or Azure Blob Storage. Veeam Data Cloud for Salesforce connects to that location to read the files; it never writes to it.

File Format

• CSV (comma-separated values) files, one file (or a set of split files) per Salesforce object.
• Files may be uncompressed, gzip-compressed (.gz), or in a password-protected ZIP archive (.zip). The password is entered when the import job is started and is not stored afterward.
• The data inside the files must not be encrypted (e.g., no field-level encryption). A password-protected archive is acceptable, but encrypted field values inside the CSV are not supported.
• The CSV header row must contain the Salesforce field API names for that object (column order doesn't matter, and matching is not case-sensitive).

File Naming

Each file name must identify the object and follow this pattern:

ObjectName[__h][-NNN].csv[.gz|.zip]

• Account.csv
  Current-state records for the Account object.
• Account__h.csv
  History records for the Account object (VBSF exports only).
• Account-001.csv.gz, Account-002.csv.gz
  A split export. The separator (- or _) must appear before the digits so they are read as part of the split and not the object name.

Any file that doesn't match this pattern will cause the whole job to fail before anything is imported, so file names should be verified before uploading.

Folder Structure

• Own exports
  Each daily snapshot must be in its own folder named day_YYYYMMDD_HHMMSS (an optional suffix, e.g. -abc123, is allowed). Seconds precision is required, and every folder in the job must have a unique timestamp. Files placed outside a day_* folder (including the storage root) are not supported and will cause the import to fail.
• Veeam Backup for Salesforce exports
  Files can be placed in the root or in any subfolder, as the restore-point and date information is already embedded in the CSV.

Note: Own and VBSF exports cannot be imported in the same job. A single import job must use only one export format, even across different objects, so each export type requires its own job.

Data Requirements and Known Limitations

• Each object's CSV must include its Salesforce record Id column, plus IsDeleted and the incremental-tracking field where those exist for the object. If a required column is missing from even one folder, the whole object fails to import (no partial import).
• Extra columns not in the Salesforce schema are ignored; the job still completes but is marked Warning, with the ignored columns listed in the job log.
• Missing optional columns and column sets that change between snapshots are both supported, as VDCSF adapts to schema differences over time.
• An empty (header-only) CSV, or a storage location with no matching files, is handled without error (the job completes with nothing imported).
• Objects that no longer exist in Salesforce cannot be imported in this release; fields removed from the org are still imported and kept in the backup as text.
• While a Data Import job is in progress for a tenant, no other backup job can run for that tenant until it finishes.

If a job fails or only partially completes, there is no need to start over. The whole job, a single file, or a filtered subset of files can be re-run, and VDCSF will pick up where it left off without creating duplicate records or gaps in restore-point history.

To limit a (re)run to a subset of files, the File Name Pattern option can be used. It accepts a regular expression matched against the file/folder path. For example:

• day_2026031[9]|day_2026032
  Imports only the daily-snapshot folders for March 19–21, 2026.
• Account
  Imports only files for the Account object.

The pattern is case-sensitive and is applied on every retry, which is important to consider when importing the remaining folders or objects later.

Next Steps

1. Eligibility should be confirmed: the tenant requires a Salesforce subscription of 200+ users and must be an empty VDCSF tenant with no existing backups.
2. The source data should be confirmed as supported: an Own or on-premises VBSF export containing Salesforce object records in CSV format.
3. A Support case should be opened to have Data Import enabled for the tenant, and to request an additional third-party source if one is needed.
4. The export files should be prepared to meet the file format, naming, and folder-structure requirements, then uploaded to Amazon S3 or Azure Blob Storage.
5. A trial import into a Sandbox or UAT tenant is recommended before importing into Production.
6. The Data Import job should be started from the VDCSF UI, and its completion status reviewed, with any affected files re-run as needed.