The whole cdb directory was shipped to end users and all header files
that cdb*.h included are also need to be shipped to make checkinc.py
pass. However, exposing gp_libpq_fe/*.h will confuse customer because
they are almost the same as libpq/*, as Heikki's suggestion, we should
keep gp_libpq_fe/* unchanged. So to make system work, we include
gp-libpg-fe.h and gp-libpq-int.h directly in c files that need them

Formarly, GPDB are not allowed to expose gpdb-specified headers in
cdb to the end users, so some public headers depend
on those headers need to do some additional work like re-typedef,
adjusting the header order in c files to pass checkinc.py. Now the
limitation is unnecessary, so we expose them too for relief.
cdbicudpfaultinjection.h may include sys/poll.h, so add it to the
hard code standard includes

It was very difficult to verify if interconnect is stucked in resending
phase or if there is udp resending latency within interconnect. To improve
it, this commit record a debug message every Gp_interconnect_debug_retry_interval
times when gp_log_interconnect is set to DEBUG.

This case assumed that 'now' in bellow queries will produce
different results.
-- insert into xxx values (timestamp(2) without time zone 'now');
-- select timestamp(2) without time zone 'now';
However, the results are undetermined and may be equal on powerful
agents, the test is all about as alias so it's ok to add 2 seconds
to the second 'now' to guarantee they are never equal.

In ORCA, we donot process interrupts during planning stage, however
if there are elog/ereport (which further calls errfinish) statements to
print additional messages we prematurely exit out the planning stage
without cleaning up the memory pools leading to inconsistent memory pool
state. This results in crashes for the subsequent queries.

This commit fixes the issue by handling interrupts while
printing messages using elog/ereport in ORCA.
Signed-off-by: NEkta Khanna <ekhanna@pivotal.io>

We are hitting an issue where gpinitsystem sends a gpstop command that
has the data directory in all lower case. This causes an issue in
Linux, but is fine with macOS; linux platforms are case sensitive vs
macOS is not.

Remove the lower casing part of the code as it doesn't seem necessary.
gpinitsystem -c command eventually overwrites the SEG_PREFIX variable in a
different function to use the mixed case, while the gpinitsystem -I does not.

Fix: keep the mixed case segment names

Signed-off-by: NMarbin Tan <mtan@pivotal.io>

gpinitsystem did not check for HEAP_CHECKSUM in the cluster
configuration file with a -c switch. This commit accepts the
HEAP_CHECKSUM setting, and additionally exports it to an
output_configuration_file when specified with the -O switch.

This commit also adds behave tests for the above, and for
reading the input_configuration_file with the -I switch.
Signed-off-by: NLarry Hamel <lhamel@pivotal.io>
Signed-off-by: NXin Zhang <xzhang@pivotal.io>

Signed-off-by: NNadeem Ghani <nghani@pivotal.io>

Bug fix for a scenario with multiple analyzedb processes running
concurrently: the resulting report files were incorrect and/or
overwritten.

This commit adds a lock (a file semaphore) for synchronization between analyzedb processes.
Each process will acquire an exclusive lock, per database, read the most recent report
files (possibly written by concurrently running analyzedb processes) and
incorporate that latest information into its own report.
Signed-off-by: NLarry Hamel <lhamel@pivotal.io>

Signed-off-by: NNadeem Ghani <nghani@pivotal.io>
Signed-off-by: NLarry Hamel <lhamel@pivotal.io>

Will hopefully reduce silly merge conflicts in the future.

They don't exist in the upstream. All but one of the callers actually just
needed list_member_*().

These functions are defined in cdbpartition.c, not tablecmds.c. The
prototypes should be in matching header file.

To match the upstream code.

Seems harmless, in this direction, but let's be tidy.

This gets rid of the GPDB-specific "QE details" message, that was only sent
once at QE backend startup, to notify the QD about the motion listener port
of the QE backend. Use a standard ParameterStatus message instead, pretending
that there is a GUC called "qe_listener_port". This reduces the difference
between the gp_libpq_fe copy of libpq, and libpq proper. I have a dream that
one day we will start using the standard libpq also for QD-QE communication,
and get rid of the special gp_libpq_fe copy altogether, and this is a small
step in that direction.

In the passing, change the type of Gp_listener_port variable from signed to
unsigned. Gp_listener_port actually holds two values: the TCP and UDP
listener ports, and there is bit-shifting code to store those two 16-bit
port numbers in the single 32-bit integer. But the bit-shifting was a bit
iffy, on a signed integer. Making it unsigned makes it more clear what's
happening.

Forgot to toupper() the first character of string to strstr().
Signed-off-by: NXiaoran Wang <xiwang@pivotal.io>

The variable was removed in PostgreSQL 8.3 (commit 7c5e5439), but this
removal of the prototype was missed in the merge.

pcqCtx hasn't existed for some time, and we don't have prosrc as a variable
here either.

toupper() all characters, then strstr() instead.

strcasestr() needs string.h

StringInfo is more appropriate in backend code. (Unless the buffer needs to
be used in a thread.)

In the passing, rename the 'conn' static variable in cdbfilerepconnclient.c.
It seemed overly generic.

* Fix an assertion failure in resgroup.

Suppose a resgroup already has a session assigned but no query running.
In another session (not assigned to this resgroup) drop this resgroup
and recreate it, then an assertion failure would be triggered.

The cause is that the old resgroup's struct in shm is reused for the new
resgroup, so on a QE who only owns a weak ref to this struct may find the
struct's resgroup id different than its expection.

To fix it we add the checks for such cases.

* Changes according to PR reviews.

\copy command reads the file from the client filesystem
but 'on segment'  reads file from the segment system
So \copy command and 'on segment' option are incompatible
Signed-off-by: NXiaoran Wang <xiwang@pivotal.io>

gpcheckmirrorseg.pl assumed GPDB is compiled as 32 bit on macs and hence doesn't
work correctly. Fixing the same as now we compile GPDB in 64 bits on MACs too.

Also, add errlog directory to ignore_list as its not supposed to be same on
primary and mirror. Plus, reflect missing relkinds for visimap, blockdirectory
and sequence tables.

CID 173910 Dereference after null check. In markDirty: Pointer is checked
against null but then dereferenced anyway. Removed the checking against and
added Assert instead as don't see why relation should be NULL in markDirty().

CID 173667 Printf arg type mismatch. A printf format string does not match the
types of one of the arguments. Fixed to make it only one %s and relpath() to get
the path.

Also, fixing the warning
```
guc.c:643:15: warning: incompatible pointer to integer conversion initializing 'bool' (aka 'char') with an expression of type 'void *' [-Wint-conversion]
                NULL, NULL, NULL
```

Clean up after commits e40e78fa, 038aa959, 859d6592 and a21da89e.
Tut tut.

In commit 06f56fe8, we assigned a random segment to server as singleton
QE for load blance. An assertion failure "outputSegIdx[0] == gp_singleton_segindex"
occurs when plan contains initplans because apply_motion_mutator() did
not set segindex of FLOW_SINGLETON to gp_singleton_segindex. Although
segindex field did not affect the creation of singleton gang and also
did not affect QE sending data to the right upstream, we still need
making them consistent to silent the assertion and avoid potential issue.

missing "-y" for yum in commit 80e931e8

This reverts commit 64d150e9.

Error message on Travis was:

```
$ wget http://ftp.jaist.ac.jp/pub/apache/apr/${APR_TAR}
--2017-08-02 06:45:44--  http://ftp.jaist.ac.jp/pub/apache/apr/apr-1.5.2.tar.gz
Resolving ftp.jaist.ac.jp (ftp.jaist.ac.jp)... 150.65.7.130, 2001:df0:2ed:feed::feed
Connecting to ftp.jaist.ac.jp (ftp.jaist.ac.jp)|150.65.7.130|:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2017-08-02 06:45:45 ERROR 404: Not Found.
The command "wget http://ftp.jaist.ac.jp/pub/apache/apr/${APR_TAR}" failed and exited with 8 during .
```

Signed-off-by: NJim Doty <jdoty@pivotal.io>