Smyatkin-Maxim opened a new pull request, #1821: URL: https://github.com/apache/cloudberry/pull/1821
Here are quite a few commints from gp7 to support gpupgrade from 6 to 7. Most of these commits are itself borrowed from PG. We also need them to support gpupgrade from GP6 to Cloudberry. At the moment I'm targeting Cloudberry 2, completely igniring main. My goal is to upgrade production clusters from GP6 to Cloudberry 2. As soon as this goal is reached, we might start thinking about either GP6->CB3 support or at least CB2->CB3 support. These commits should't break anything but aren't enough for an upgrade. For gpupgrade to work we need this PR to land + a bunch of additional patches to core(15-20 smallish commits) and a bunch of patches to gpupgrade itself. I already have a prototype which completely restores `regression` dataset (the one created during installcheck), it consists of 18 patches: ``` ls -lt ./gpupgrade-patches/ total 108 -rw-rw-r-- 1 smiatkin smiatkin 1788 Jun 14 21:22 0018-pg_dump-drop-unused-tgispartition-alias-in-getTrigge.patch -rw-rw-r-- 1 smiatkin smiatkin 1496 Jun 14 21:22 0017-pg_dump-restrict-default-type-storage-options-to-bas.patch -rw-rw-r-- 1 smiatkin smiatkin 2134 Jun 14 21:22 0016-pg_dump-simplify-attribute-encoding-attrelid-query.patch -rw-rw-r-- 1 smiatkin smiatkin 1943 Jun 14 21:22 0015-pg_dump-fix-getPartitionDefs-GPDB6-version-guard-and.patch -rw-rw-r-- 1 smiatkin smiatkin 6675 Jun 14 21:22 0014-pg_dump-address-external-partition-EXCHANGE-without-.patch -rw-rw-r-- 1 smiatkin smiatkin 1801 Jun 14 21:22 0013-pg_dump-skip-subpartition-TEMPLATE-during-binary-upg.patch -rw-rw-r-- 1 smiatkin smiatkin 3630 Jun 14 21:22 0012-pg_dump-only-emit-column-ENCODING-for-AOCO-tables.patch -rw-rw-r-- 1 smiatkin smiatkin 4228 Jun 14 21:22 0011-pg_dump-synthesize-table-AM-from-relstorage-and-stri.patch -rw-rw-r-- 1 smiatkin smiatkin 6360 Jun 14 21:22 0010-pg_dump-translate-gp_default_storage_options-to-defa.patch -rw-rw-r-- 1 smiatkin smiatkin 1867 Jun 14 21:22 0009-pg_dump-add-toast-table-OID-to-binary-upgrade-preass.patch -rw-rw-r-- 1 smiatkin smiatkin 3740 Jun 14 21:22 0008-pg_dump-force-array-type-OID-preassignment-for-AO-AO.patch -rw-rw-r-- 1 smiatkin smiatkin 8490 Jun 14 21:22 0007-pg_dump-skip-toast-AO-aux-OID-preassignment-for-lega.patch -rw-rw-r-- 1 smiatkin smiatkin 1157 Jun 14 21:22 0006-pg_dump-fix-GPDB7-CBDB2-major-version-constants.patch -rw-rw-r-- 1 smiatkin smiatkin 4153 Jun 14 21:22 0005-pg_dumpall-dump-role-PASSWORD-and-gate-GPDB-profile-.patch -rw-rw-r-- 1 smiatkin smiatkin 13568 Jun 14 21:22 0004-pg_upgrade-map-GP5-6-partition-parents-and-external-.patch -rw-rw-r-- 1 smiatkin smiatkin 1189 Jun 14 21:22 0003-pg_upgrade-default-file_encryption_method-for-cluste.patch -rw-rw-r-- 1 smiatkin smiatkin 1317 Jun 14 21:22 0002-Map-AO-compresslevel-5-and-6-to-zstd-in-datumstream.patch -rw-rw-r-- 1 smiatkin smiatkin 2776 Jun 14 21:22 0001-Allow-RLE-compresslevel-up-to-6-when-zstd-is-availab.patch ``` Those are partially my own, partially LLM-generated. So I'll have to invest some time to sanity-check it. But it indeed allows a very complex gpupgrade for `regression` databases to pass. So, more PRs are coming, but probably very small and simple ones. -- 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]
