[Bug]: Local storage adaptor silently fails to upload when out of space #7715
Labels
area/API
Improvements or additions to the API
area/block-adapter
area/gateway
Changes to the gateway
bug/critical
bug
Something isn't working
contributor
P3
What happened?
Lakefs opensource version was deployed in a Kubernetes cluster. The storage space was smaller than the size of files we wanted to upload. Using Aws cli tool to upload the files to the lakefs s3 interface , everything seems to have went fine. But storage fills up in the server and there is a mismatch on file sizes in the meta store vs the actual file size on disk. (Note this is using local storage for lakefs storage) .
Steps to Reproduce:
Expected behavior
The aws client or any other uploading tool should be notified and be aware of error if lakefs doesn't have enough storage.
lakeFS version
treeverse/lakefs:1.18.0
How lakeFS is installed
Kubernetes
Affected clients
Aws-cli s3 , python sdk
Relevant log output
probably this chat on lakefs slack help might be releavant than the logs... https://lakefs.slack.com/archives/C016726JLJW/p1713891191846399 https://lakefs.slack.com/archives/C016726JLJW/p1714576349984149
Contact details
kebedey@renci.org
The text was updated successfully, but these errors were encountered: