* docs - add pxf server cfg topic, rework other cfg topics, misc edits

* some of the edits requested by francisco

* add Hive back in to list of JDBC SQL dbs

* sneaking in a misc unrelated formatting fix

* edits requested by david

The curl slist API properly handle NULLs so we can be less verbose
and skip the check before passing to the slist cleanup function.
Reviewed-by: NFrancisco Guerrero <aguerrero@pivotal.io>

The header callback was storing the response in the context, but as
it was never used we might as well save the memory and just return
the required return of the number of bytes we would've saved should
we have allocated.
Reviewed-by: NFrancisco Guerrero <aguerrero@pivotal.io>

The only time the internal buffer cleanup code was called was just
before freeing the entire context, so individually zeroing out the
members is pointless. Remove the function entirely and inline the
buffer freeing into the context cleanup codepath.

For zeroing the error buffer, it's only called right after allocating
the error buffer with palloc0() in the first place so the memory will
always be zeroed out when reaching here.
Reviewed-by: NFrancisco Guerrero <aguerrero@pivotal.io>

If strlen(addr) is zero then based on how get_dest_address() works
addr will be NULL, and pfree() on NULL is not permitted.  Also, we
know that addr will either be a non-empty string or NULL, so we can
just as well test for addr being NULL and avoid a strlen() call.

Fix by only pfreeing when addr is set. (this is in an elog(ERROR..)
context so freeing isn't terribly interesting but it also doesn't
hurt so I'm keeping the current codepath.)
Reviewed-by: NFrancisco Guerrero <aguerrero@pivotal.io>

The libchurl abstraction layer has many internal helper functions
which weren't marked static and thus exported. Fix by marking all
as static.
Reviewed-by: NFrancisco Guerrero <aguerrero@pivotal.io>

The default libcurl on centos7 uses NSS SSL library instead of OPENSSL.
GPDB with that libcurl will report "Unknown cipher in list: AES128-SHA"
on connecting gpfdist SSL server.
Actually, we don't have to set cipher list in GPDB. It may choose stronger
cipher on negotiation. So remove setting cipher list.

Patch back to 6X_STABLE

GPDB uses libcurl-gnutls.so.4 as default libcurl on ubuntu. gpfdist SSL
connection reports error (56 - Failure when receiving data from the peer)
on handling POST message. We find it shutdown socket directly without sending
close_notify to client. So we call SSL_shutdown() before socket shutdown.

But SSL_accept() will return error with above patch on Centos, because
GPDB curl reuses SSL session ID in the second client hello, but server considers
that session is shutdown, so it won't accept that session ID.
The solution is disabling SSL session ID cache by setting curl option.
Then both Centos and ubuntu work well.

Enable gpfdist SSL test case.

Patch back to 6X_STABLE

When determine waiting relation in global dead lock detector,
we should also check whether the waiter's lockmode conflict with
holder's holdmask.
Co-authored-by: NNing Yu <nyu@pivotal.io>
Reviewed-by: NZhenghua Lyu <zlv@pivotal.io>

`create rule <...> do instead select * from t for update`
will dispatch a query with lockingclause node. Add
deserialization method for it to make things correct.

(cherry-picked from 256d8380)
Co-authored-by: NGoutam Tadi <gtadi@pivotal.io>
Co-authored-by: NAmil Khanzada <akhanzada@pivotal.io>

* Source for the segment recovery flowchart

* Rename gprecovery-flow-chart.graffle to recovermatrix.graffle to match png filename

* docs - add info about pxf jdbc statement properties

* misc edits requested by david

The utilities reference page contained (unlinked) mentions of long
since deprecated utilities, and since they aren't shipping it's time
to remove them:

	gpsizecalc - removed in June 2010
	gpskew - removed in October 2009
	gprebuildsystem - removed in January 2010
	gpchecknet - removed in April 2010
	gpcheckos - removed in July 2016

Reviewed-by: Lisa Owen
Reviewed-by: David Yozie

gpdetective was removed in 4.3.5, so it seems about time to also
remove the documentation for it (which was unreachable due to the
app being marked deprecated).

Reviewed-by: Lisa Owen
Reviewed-by: David Yozie

* docs - add pageinspect module landing page

* remove peer scope

* cannot inspect ao or external relations

The "combine" function for the int4 sum/avg aggregate functions is
backported to Greenplum 6.0 in commit 313cef6e (from
postgres/postgres@11c8669c0cc) but we made the inevitable omission of
setting the "strictness" of int4_avg_combine to false. This by itself is
harmless as the actual function body *does* guard against NULL input,
but it also prohibits a whole host of optimizations when the executor
and planner can detect NULL input early on and short-circuit the
execution. Oops.

This patch flips the `proisstrict` flag back to true for
int4_avg_combine.

Backpatch to 6X_STABLE.

(cherry picked from commit c334bedc)

In pg_get_expr(), after getting the relname, if the table that the relid tells
is dropped, an error will raise later when opening the relation to get column
names.

pg_get_expr() is used by GPDB add-on view 'pg_partitions' which is widely used
by regression tests for partition tables. Lots of parallel test cases issue view
pg_partitions and drop partition tables concurrently, so those cases are very
flaky. Serialize test cases will cost more testing time and be fragile, so GPDB
holds a AccessShareLock here to make tests stable.

optimizer_enable_dml is set to true by default. When set to false, ORCA will
fall back to planner for all DML queries.

* docs - add to pxf upgrade procedure (v5.0.1 to v5.3.2)

* remove extraneous to

When referring to the product name and not a database instance running
Greenplum, the capitalization should be "Greenplum Database".

In the past several weeks, gpload test case 22 failed two times.
It may report: Fatal Python error: GC object already tracked.
We have reproduced this issue locally and in dev pipeline for 3
days, but it can't be reproduced.
So we decide to disable it now, in order to not blocking other PRs.

Assume user1 has the privilege to database db1 and user2 has not, when
user1 try to create a schema in db1 and authorize it to user2, a
permission denied error is reported in QE. The RCA is QD set current
user to user2 before dispatching the query to QEs, so QE will also
set current user to user2, however, user2 has no provilege to create
schema in database db1.

To fix this, we delay setting the current user to user2 until the query
is dispatched to QEs.

GPDB used to allow command like "START_REPLICATION %X/%X [SYNC]" to
start a replication, user can specify SYNC option to skip waiting
synchronous in replications. Now start replication command is made
similar to upstream, the SYNC option is not supported, however, the
internal flag "synchronous" is still used and always be false which
make master and standby never synchronized.

* docs - enhance pxf jdbc partitioning content

* add missing comma

* simplify some content

* docs - clarify pxf filter partitioning support for hive

* clarify the hadoop cfg update content

* remove cluster start

* edits requested by david

* remove disable statement per shivram

* docs - add pxf jdbc.connection.transactionIsolation server cfg property

* use read uncommitted in example; add external

* docs - jdbc cfg supports connection- and session-level properties

* some edits requested by david

* reword jdbc server cfg opening paragraph

* clarify rejected session prop/value chars as requested by ivan

* docs - discuss pxf fragment metadata caching

* a large number of

* docs - pxf init/sync support to master standby

* edits requested by david

* edits requested by francisco and oliver

* pxf sync from master TO standby or seg host

* identify sync run on master in pxf sync option description

Some of the task files were used by jobs in the orca pipeline, but those
jobs have been removed so the files are not being used anymore.
Co-authored-by: NKaren Huddleston <khuddleston@pivotal.io>
Co-authored-by: NDavid Sharp <dsharp@pivotal.io>

A motion hazard is a deadlock between motions, a classic motion hazard
in a join executor is formed by its inner and outer motions, it can be
prevented by prefetching the inner plan, refer to motion_sanity_check()
for details.

A similar motion hazard can be formed by the outer motion and the join
qual motion.  A join executor fetches a outer tuple, filters it with the
join qual, then repeat the process on all the outer tuples.  When there
are motions in both outer plan and the join qual then below state is
possible:

0. processes A and B belong to the join slice, process C belongs to the
   outer slice, process D belongs to the JoinQual slice;
1. A has read the first outer tuple and is fetching tuples from D;
2. D is waiting for ACK from B;
3. B is fetching the first outer tuple from C;
4. C is waiting for ACK from A;

So a deadlock is formed A->D->B->C->A.  We can prevent it also by
prefetching the join qual.
Reviewed-by: NJesse Zhang <jzhang@pivotal.io>
Reviewed-by: NGang Xiong <gxiong@pivotal.io>
Reviewed-by: NZhenghua Lyu <zlv@pivotal.io>

(cherry picked from commit fa762b69)