PG17 - Add Regression Test for REINDEX support in event triggers

[m3hm3t]

This PR adds regression tests to verify REINDEX support with event triggers. Tests validates trigger execution, shard placement consistency, and distributed index rebuilding without disruption.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Please upload report for BASE (release-13.0@103125f). Learn more about missing BASE report.

Additional details and impacted files

@@               Coverage Diff               @@
##             release-13.0    #7819   +/-   ##
===============================================
  Coverage                ?   89.63%           
===============================================
  Files                   ?      274           
  Lines                   ?    59748           
  Branches                ?     7452           
===============================================
  Hits                    ?    53558           
  Misses                  ?     4059           
  Partials                ?     2131           

[naisila]

LGTM, @onurctirtir can you give a second look?
Basically, do you see any issue with having start-end event triggers for a REINDEX command on a distributed table? We don't distribute event triggers currently in Citus, but I think it's okay to have unpropagated start-end event trigger for a propagated reindex command?

[onurctirtir]

LGTM, @onurctirtir can you give a second look? Basically, do you see any issue with having start-end event triggers for a REINDEX command on a distributed table? We don't distribute event triggers currently in Citus, but I think it's okay to have unpropagated start-end event trigger for a propagated reindex command?

For reindex commands, the way that Citus acts when "concurrently" option is provided is a bit tricky. So, as long as we also add a test where we provide "concurrently" option to "reindex" command, we should be okay at the high level.

See ExecuteDistributedDDLJob().

[naisila]

@onurctirtir

For reindex commands, the way that Citus acts when "concurrently" option is provided is a bit tricky. So, as long as we also add a test where we provide "concurrently" option to "reindex" command, we should be okay at the high level.

Do you have in mind an isolation test?

[onurctirtir]

@onurctirtir

For reindex commands, the way that Citus acts when "concurrently" option is provided is a bit tricky. So, as long as we also add a test where we provide "concurrently" option to "reindex" command, we should be okay at the high level.

Do you have in mind an isolation test?

I think we don't need to add an isolation test. If we can just use "concurrently" syntax in one of our tests, that should also be fine. This is because, in the past, even simply providing "concurrently" option was causing indefinite self deadlocks or sometimes crashes for create index / reindex commands in Citus.