This reverts commit 9d9798e5.

In commit 90450095 we added some tests
with fault injector, however looks like the output format of
gp_inject_fault() has changed on 6X_STABLE, so although the PR was
verified before merging, the tests ic and external_table are red on
6X_STABLE pipelines.

Fixed by updating the output format of gp_inject_fault().

There was a hang like this: when one QE errors out before 'SetupInterconnect',
QD will keep waiting for the incoming connections to be established and doesn't
check the error message from dispatcher. Other QEs are finished and hang in
function 'waitOnOutbound'.
Co-authored-by: NAsim R P <apraveen@pivotal.io>
Co-authored-by: NGang Xiong <gxiong@pivotal.io>
Co-authored-by: NNing Yu <nyu@pivotal.io>

Commit 6fa3bce8 repurposed the -S
parameter making the existing use instead be --mirror-mode, but
the help text remained. Remove the now incorrect usage entry.

Reported by Github user boazhh in issue #8579
Backported from master commit a398b437

ci pipeline uses two segment nodes for resgroup testing which is different than
ususal three segment node test configuration. Modifying the test to make it
independent of segment numbers.

cherry-picked from 89c55b08

Co-authored-by: NPaul Guo <pguo@pivotal.io>
Co-authored-by: NNing Yu <nyu@pivotal.io>

Cherry-picked from 550658bf

Fix potential shared memory corruption in resource group slot management when query involves entrydb.

We've observed that when we terminate a query that involves entrydb, if QD
detaches from a resource group before the entrydb backend does that in
UnassignResGroup(), data corruption on shared slot pool could happen.  In a
cassert enabled gpdb version, you could see below FATAL message as below when
terminating the query using pg_terminate_backend().

FATAL:  Unexpected internal error (resgroup.c:1545)
DETAIL:  FailedAssertion("!(slot->nProcs == 0)", File: "resgroup.c", Line: 1545)
HINT:  Process 9903 will wait for gp_debug_linger=120 seconds before termination.
Note that its locks and other resources will not be released until then.

While on product release that without cassert enabled some subsequent queries
could lead to panic.

We fix this by let the final process (either QD or entrydb process) do clean up
when the nProcs reference number is zero. This is more robust and is less bug
prone cross future code change and upstream merge.

This patch refactors the related resource code a bit and also moves the fault
inject related negative tests ahead so that we could capture some potential
errors which happen later (typically the case in this patch).

Besides, it fixes another panic caused by potential MySessionState NULL
resetting (see code comment for details). That was revealed when running tests
in my dev machine so it's a potential flaky testing part.
Co-authored-by: NPaul Guo <pguo@pivotal.io>
Co-authored-by: NNing Yu <nyu@pivotal.io>
Co-authored-by: NGang Xiong <gxiong@pivotal.io>

Cherry-picked from 8913827b

This test was added to 5X_STABLE as a part of aee8cac8.

(cherry picked from commit 795b146d)

This was added as a part of a fix for the vacuum segment file drop phase
for AOCS tables on 5X_STABLE: aee8cac8

(cherry picked from commit 8cf66812)

--Removed references to foreign data wrappers install with GPDB
--Modified note. Added information stating PostgrSQL FDWs only access master.

* docs - SERVER is required when accessing an object store

* fix formatting in new note

* Expand Preamble improvement - #168158510

* expand preamble on overview page

* update to reference parameters

* update reference paremeters description

* docs - Add OpenJDK install information

This will be backported to 6X_STABLE

* doc - Revamped Java install based on review comments. Also other review comment updates.

* docs - updates based on review comments.

It should proceed only after the 'forked' connection is really blocking at the
lock. & means blocking but the test framework has no way to really know the sql is blocking at lock waiting so test case should take the responsibility. 
Reviewed-by: NAsim R P <apraveen@pivotal.io>

While processing constraint interval, also consider predicate of type
<cast(ident)> array cmp else we lose the opportunity
to generate implied quals. This is corresponding to the ORCA changes.

* docs - Add Ubuntu install/uninstall R information

This will be backported to 6X_STABLE

* docs - updates to R install based on review comments.

This corresponds to ORCA commit "Fix stats calculation for Dynamic scans
with partition elimination"
Authored-by: NChris Hajas <chajas@pivotal.io>
(cherry picked from commit 9f8e4ff2)

Consider the query below:

```
test=# explain select * from foo, jazz where foo.c = jazz.e and jazz.f = 10 and a in (select b+1 from bar);
						 QUERY PLAN
--------------------------------------------------------------------------------------------------------------------
Gather Motion 3:1  (slice4; segments: 3)  (cost=0.00..1324469.30 rows=1 width=20)
->  Hash Join  (cost=0.00..1324469.30 rows=1 width=20)
     Hash Cond: foo.c = jazz.e
     ->  Dynamic Table Scan on foo (dynamic scan id: 1)  (cost=0.00..1324038.29 rows=34 width=12)
	   Filter: a = (b + 1) AND ((subplan))
	   SubPlan 1
	     ->  Materialize  (cost=0.00..431.00 rows=1 width=4)
		   ->  Broadcast Motion 1:3  (slice2)  (cost=0.00..431.00 rows=3 width=4)
			 ->  Limit  (cost=0.00..431.00 rows=1 width=4)
			       ->  Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..431.00 rows=1 width=4)
				     ->  Limit  (cost=0.00..431.00 rows=1 width=4)
					   ->  Table Scan on bar  (cost=0.00..431.00 rows=34 width=4)
     ->  Hash  (cost=100.00..100.00 rows=34 width=4)
	   ->  Partition Selector for foo (dynamic scan id: 1)  (cost=10.00..100.00 rows=34 width=4)
		 ->  Broadcast Motion 3:3  (slice3; segments: 3)  (cost=0.00..431.00 rows=1 width=8)
		       ->  Table Scan on jazz  (cost=0.00..431.00 rows=1 width=8)
			     Filter: f = 10
Optimizer status: PQO version 3.65.0
(18 rows)
```

Previously, since the subplan was in a qual, we did not populate the
qual properly when executing a dynamic table scan node. Thus the
subplan attribute in PlanState of the dynamic table scan was incorrectly
set to NULL, causing a later crash.

We now populate this similarly to how we do it for dynamic index/bitmap
scans.
Co-authored-by: NSambitesh Dash <sdash@pivotal.io>
Co-authored-by: NChris Hajas <chajas@pivotal.io>
Co-authored-by: NAshuka Xue <axue@pivotal.io>

(cherry picked from commit 81a671f6)

The replication_slots test brings down segments in various ways and
performs incremental as well as full recovery.  After gprecoverseg
returns success, the segments may still be marked as down or not
in-sync.  The test must, therefore, wait until the segments are up and
in sync before moving on.  Especially when the subsequent steps
include rebalance or checking replication slots status, which may fail
if invoked when segments are not in-sync.  CI results indicate that
such failures have happened several times.
Co-authored-by: NPengzhou Tang <ptang@pivotal.io>
Reviewed by: Ashwin Agrawal, Jacob Champion and Jimmy Yih

(cherry picked from commit a7bce5f7)

This will be backported to 6X_STABLE

* docs - pxf: add s3-specific topic, add amazon s3 select topic

* remove DELIMITER specification in LOCATION URI plus upgrade considerations

* update OSS book subnav

* clarify delimiter in formatting options still required

* reword the upgrade step a bit

* edits requested by oliver

* minor edits requested by david

In 6.x, adding a PRIMARY KEY or UNIQUE constraint to a table based on an existing unique index is not supported on a partitioned table.

Also added syntax for table_constraint_using_index

This will be backported to 6X_STABLE

for 6X_STABLE

PR for PL/JAVA for gp6: https://github.com/greenplum-db/pljava/pull/40
PR for PL/R for gp6: https://github.com/greenplum-db/plr/pull/60

* docs - advice for preventing OOM with C UDAs

* Changes from review