rangareddy commented on issue #7991: URL: https://github.com/apache/hudi/issues/7991#issuecomment-4834816958
The HEAD/LIST surge is the 0.12.x meta-sync regression (HUDI-5477) — the Glue sync client reloaded the full archived timeline from S3 on every commit without caching, so cost scaled with the size of `.hoodie/archived`. Fixed in 0.12.3 / 0.13.0 via #7561 (active-timeline-only read + archived-timeline caching); on Glue's capped 0.12.1, request the patched jar from AWS Support, and enable the metadata table (`hoodie.metadata.enable=true`) to cut the recursive LIST calls. The : partition error is just config: keep the typed `datecreated:TIMESTAMP,tenant:SIMPLE` on the keygen side only, and set `hive_sync.partition_fields=datecreated,tenant` (untyped) with MultiPartKeysValueExtractor. -- 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]
