Zhenghua Lyu

Good catch! But I have some other concerns: the same SQL is also tested with GDD enabled, so GDD's SQL also will hit the fault-inject in theory? Let's first answer...

I have rebased this and open a new one https://github.com/greenplum-db/gpdb/pull/13923 Close this. Thanks a lot for the PR.

Hi, thanks for your contribution @red1452 The CI pipeline fails due to: Planner ``` ====================================================================== DIFF FILE: ../gpdb_src/contrib/auto_explain/regression.diffs ---------------------------------------------------------------------- diff -I HINT: -I CONTEXT: -I GP_IGNORE: -U3 /tmp/build/d62a0504/gpdb_src/contrib/auto_explain/expected/bfv_preload_auto_explain.out /tmp/build/d62a0504/gpdb_src/contrib/auto_explain/results/bfv_preload_auto_explain.out ---...

Looks reasonable at first quick glance. Will look into it this week. Thanks a lot for your contribution!

Hi @red1452 your analysis looks reasonable. But I do not reproduce as your steps: ``` gpadmin@zlyu:~/workspace/gpdb/contrib/auto_explain$ make -j -s install gpadmin@zlyu:~/workspace/gpdb/contrib/auto_explain$ gpconfig -c shared_preload_libraries -v 'auto_explain'; 20220818:09:28:50:158378 gpconfig:zlyu:gpadmin-[INFO]:-completed successfully with...

> I configure and make the current master branch with options `--with-python --disable-gpcloud --enable-cassert CFLAGS="-ggdb -g3 -O0 -fno-omit-frame-pointer" --enable-debug-extensions --enable-orca --enable-depend`. And do the same as description: > > ```...

Hi, the new issue is due to entrydb QE, using planner, the reproduce in this PR will not schedule singleQE to entry, ORCA will. This PR is to gpdb master...

My two cents for this PR: 1. do not change code in `_PG_init` 2. reword the commit message then we can push this. Thanks a lot for your contribution! @red1452

Pushed, thanks a lot for your contribution.

Looks good at first glance. Thanks for your contribution. Please kindly wait for more time and let me and the team look into this.