kevinjqliu commented on PR #3580: URL: https://github.com/apache/iceberg-python/pull/3580#issuecomment-4829115624
Root cause looks like gcsfs 2026.6.0, specifically its experimental ExtendedGcsFileSystem control-plane path with the fake GCS server. What changed: - gcsfs 2026.6.0 now passes Iceberg’s custom GCS endpoint (http://0.0.0.0:4443) into the Storage Control gRPC client. The fake server supports the GCS JSON API, but not the Storage Control gRPC API. So every bucket-type probe tries gRPC against fake-gcs-server and burns timeout/retry time. Why it takes so long: - gcsfs probes bucket type for zonal/HNS support on common paths like open, simple_upload, cat_file, rm, and list/info flows. When the probe fails, it returns UNKNOWN, and UNKNOWN is intentionally not cached, so the same slow probe repeats for every operation. A simple write/read/delete test can trigger roughly five probes. 🤷 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
