README
¶
Support
Import repositories to Harness Code Repository with metadata including:
- Pull Requests and comments
- Webhooks
- Branch Protection Rules
Installing
You can install the migrator via github releases or run make build with latest go version present in your system. You can skip this step if you have the migrator installed already.
NOTE
You need to export the repository first using the same migrator tool. Supported SCM providers are Github, Gitlab, Bitbucket, and Bitbucket Server (Stash).
Commands
As a quick start you can run
./harness-migrate git-import <zip-folder-path> --space <target acc/org/project> --endpoint <harness-url> --token <token>
where you have to replace all values enclosed in brackets <>.
You can also provide more advanced options. You can look at those via help:
./harness-migrate git-import --help
Examples
Standard Migration
Importing a repository without mapping users, skip the meta data, and increase the file-size-limit on server to 102MB.
export harness_TOKEN={harness-pat}
./harness-migrate git-import ./harness/harness.zip --space "acc/MyOrg/Myproject" --endpoint "https://app.harness.io/" --skip-users --skip-pr --skip-webhook --skip-rule --file-size-limit 102000000
Incremental Migration
The --no-git flag enables incremental migration for repositories that already exist on Harness Code. This feature allows you to migrate additional pull request metadata from your source SCM without re-importing the git repository itself.
When to Use Incremental Migration
Use incremental migration when:
- You have already imported a repository to Harness Code using the standard migration process
- You want to import additional historical pull requests and metadata from the source SCM
How Incremental Migration Works
- Repository Validation: The tool verifies that the target repository already exists on Harness Code
- Offset Calculation: Automatically determines the current highest PR number on Harness Code to avoid conflicts
- PR Number Adjustment: Source PR numbers are increased by the calculated offset before importing
- Metadata Import: Only pull request data, comments, and metadata are imported - git repository content is skipped
Important Considerations
⚠️ Repository Unavailability: During incremental migration, the target repository will be temporarily unavailable on Harness Code.
⚠️ Backup Recommended: Create a backup of your Harness Code repository before proceeding with incremental migration.
PR Number Mapping Example
If your Harness Code repository has 100 existing PRs and you're migrating 50 historical PRs from GitHub:
- Existing Harness PRs: #1 - #100 (remain unchanged)
- Migrated GitHub PRs: GitHub PR #1 becomes Harness PR #101, GitHub PR #50 becomes Harness PR #150
- Future PRs: Will continue from #151 onwards
This ensures no conflicts between existing and migrated pull request numbers.
Usage
./harness-migrate git-import <zip-folder-path> --space <target-space> --endpoint <harness-url> --token <token> --no-git
Prerequisites
- Target repository must already exist on Harness Code
- You must have write permissions to the target repository
Example
Migrating additional pull requests to an existing repository on Harness Code:
export harness_TOKEN={harness-pat}
./harness-migrate git-import ./additional-prs.zip --space "acc/MyOrg/Myproject" --endpoint "https://app.harness.io/" --no-git
This will:
- Skip git repository import (since it already exists)
- Automatically calculate PR number offset to avoid conflicts
- Import only pull request metadata with adjusted PR numbers
Troubleshooting
Import fails due to
| Error | Resolution |
|---|---|
| Resource not found | Target space must exist on Harness. You can import repositories into account, organization or projects. |
| Unauthorized/Forbidden | Provided token must be able to create and edit a repository. |
| Forbidden | Contact support to enable CODE_IMPORTER_V2_ENABLED FF for your account. |
| Push contains files exceeding the size limit | Increase --file-size-limit, default is 100MB. |
failed to import pull requests and comments for repo 'TARGET-SPACE/REPO' : client error 413: |
The pull requests payload is too large. Use --batch-size with a smaller value (e.g., --batch-size 50). The default is 100. |
Users
When exporting repositories, we collect information about users who have interacted with them. This data is saved in a users.json file within the exported zip. During the import process, if a user exists on the Harness platform with the same email, their activities will be preserved. However, if the user is missing, you have two options: using --skip-users which skips mapping their activities to the migrator (determined by the token you provided) or manually create the user on the target platform first before the import.
Webhooks and Branch Rules names
Imported repositories could have a diff name for webhooks and branch rules. (e.g., names start with webhook_ or rule_)
Documentation
Source Files