This is an automated email from the ASF dual-hosted git repository. yjhjstz pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/cloudberry.git
commit c6931b5ff75819219f5fe02d1ca42609c8a4d1b1 Author: Zhenghua Lyu <[email protected]> AuthorDate: Sat Aug 20 13:56:00 2022 +0800 Resolve a GPDB_12_MERGE_FIXME in regress/dsp. The first FIXME is introduced by the commit from merge branch 002f61d8. Remove it and run the full pipelines it passes the tests. The second FIXME is asking if it is necassary to test GUC gp_default_storage_options should keep consistent among master and segments. Although gpconfig's unittest might cover this, it does no harm to also test here for this important GUC. --- src/test/regress/expected/dsp.out | 16 +++------------- src/test/regress/sql/dsp.sql | 17 +++-------------- 2 files changed, 6 insertions(+), 27 deletions(-) diff --git a/src/test/regress/expected/dsp.out b/src/test/regress/expected/dsp.out index 65adc8ffa4..2a5997b7f8 100644 --- a/src/test/regress/expected/dsp.out +++ b/src/test/regress/expected/dsp.out @@ -1197,23 +1197,13 @@ select c.relname, am.amname, c.relkind, c.reloptions dsp_partition1_1_prt_split_p2 | ao_row | r | (6 rows) --- GPDB_12_MERGE_FIXME: gpcheckcat fails for dsp_partition1_1_prt_1 --- with checksum mismatch between master and segments. The reason --- being RESET gp_default_storage_options above somehow seems not --- working properly. So, we should fix RESET to work. Plus, also we --- should enhance code to not use `gp_default_storage_options` guc on --- segments. Instead master should use it and make all the decisions --- and pass the decision to segments. This should avoid any --- mis-matches and eliminates the need for the guc to be in-sync --- between master and segment. -drop table dsp_partition1; RESET gp_default_storage_options; -- cleanup \c postgres -- Test that gp_default_storage_options option must be the same on all nodes --- GPDB_12_MERGE_FIXME: Is this really strictly necessary? There's code in --- gpconfig to verify this specifically for gp_default_storage_options. Can --- we remove it? +-- NOTE: although gpconfig might have unittest to verity this, the GUC +-- gp_default_storage_options is so important that adding tests for it here +-- do no harm. -- start_matchsubs -- m/.*\[ERROR\]*/ -- s/.*\[ERROR\]/[ERROR]/gm diff --git a/src/test/regress/sql/dsp.sql b/src/test/regress/sql/dsp.sql index 849913795e..23c9c5bfb2 100644 --- a/src/test/regress/sql/dsp.sql +++ b/src/test/regress/sql/dsp.sql @@ -477,26 +477,15 @@ select c.relname, am.amname, c.relkind, c.reloptions from pg_class c left join pg_am am on (c.relam = am.oid) where c.relname like 'dsp_partition1%' order by relname; --- GPDB_12_MERGE_FIXME: gpcheckcat fails for dsp_partition1_1_prt_1 --- with checksum mismatch between master and segments. The reason --- being RESET gp_default_storage_options above somehow seems not --- working properly. So, we should fix RESET to work. Plus, also we --- should enhance code to not use `gp_default_storage_options` guc on --- segments. Instead master should use it and make all the decisions --- and pass the decision to segments. This should avoid any --- mis-matches and eliminates the need for the guc to be in-sync --- between master and segment. -drop table dsp_partition1; RESET gp_default_storage_options; -- cleanup \c postgres -- Test that gp_default_storage_options option must be the same on all nodes - --- GPDB_12_MERGE_FIXME: Is this really strictly necessary? There's code in --- gpconfig to verify this specifically for gp_default_storage_options. Can --- we remove it? +-- NOTE: although gpconfig might have unittest to verity this, the GUC +-- gp_default_storage_options is so important that adding tests for it here +-- do no harm. -- start_matchsubs -- m/.*\[ERROR\]*/ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
