Hey Anton, I think that's a great idea. Would be great to get Spark 3.4 and Flink 1.7 support out. I saw some questions on the Apache Slack today and also noticed that Spark supports timestamps without timezone from Spark 3.4 onward and went ahead and created an issue for that <https://github.com/apache/iceberg/issues/7456>.
We might also want to add the Hadoop and Parquet upgrades as a part of this. Do we have a consensus on upgrading Hadoop? Hadoop in Iceberg is still at 2.7.x, Parquet 1.13.0 requires Hadoop 3.x+ <https://issues.apache.org/jira/browse/PARQUET-2276>, I've backported support for Hadoop 2.9.x <https://github.com/apache/parquet-mr/pull/1083> which will be part of the patch release 1.13.1. Today at the Parquet sync I volunteered as a release manager to do the patch release. This would also come with running Flink without Hadoop, which would be awesome. I'm all in for doing another release in May, so we can get this in as well. WDYT? Kind regards, Fokko Driesprong Op vr 28 apr 2023 om 19:54 schreef Jack Ye <yezhao...@gmail.com>: > Looking at the milestone https://github.com/apache/iceberg/milestone/26, > most of the PRs are in good progress except for > https://github.com/apache/iceberg/issues/7449. Given the fact that many > people are looking forward to using Spark 3.4 and Flink 1.17, it's probably > worth having a quick release. > > We might also want to add the Hadoop and Parquet upgrades as a part of > this. > > -Jack > > > > On Thu, Apr 27, 2023 at 6:35 PM Anton Okolnychyi > <aokolnyc...@apple.com.invalid> wrote: > >> We briefly discussed the idea of releasing 1.3.0 around mid May during >> the sync. The primary goal is to support Spark 3.4 and Flink 1.17. There >> are also a few other notable changes that went in or are very close. Do we >> see any blockers that we should track? >> >> - Anton > >
