apache / apache/iceberg-python
feat: Automatic Vended Credential Refresh
- Dominant language
- Python
- Stars
- 1.1k
- Forks
- 581
- Avg merge
- 1d 17h
- Merged PRs (30d)
- 78
Description
### Feature Request / Improvement
Today, long-running writes to PyIceberg risk failing when credentials expire mid-transaction. The Java implementation addressed this by adding a vended-credential refresh mechanism that calls a REST catalog endpoint to obtain new credentials whenever the current ones are within five minutes of expiring.
I propose a similar approach for PyIceberg, delivered in three separate PRs:
1. **VendedCredential class implementation** (largest diff) -> [PR](https://github.com/apache/iceberg-python/pull/3507) ready for review
2. **Credential refresh support in `PyArrowFileIO`** -> PR pending review and merge of 1
3. **Credential refresh support in `FsspecFileIO`** -> PR pending review and merge of 1
PRs 2 and 3 depend on PR 1, which provides the core vended-credential implementation.
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by reviewing PR 3507, which contains the VendedCredential implementation and is the dependency for the remaining work. Then inspect the planned PyArrowFileIO and FsspecFileIO refresh support; the feature is complete when both integrations refresh credentials obtained from the REST catalog before expiry.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- data-engineering
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100