This commit brings following changes:
1) Implementation of tap based testing using perl language for different
scenarios that could not be covered under traditional SQL based diff testing,
or require server start/shutdown. At this point of time, tap testing is only
enabled for Postgres 14 & 13, for rest of back branches it will be done at
laster time as there is substantial change in number of columns and their
names.
2) Changes to github action workflows for Postgres 14 & 13 to accomodate the
requirements for tap testing.
3) Similarly, minor changes to Makefile are also done.
4) Testing of supported GUCs using tap tests for different possible
configuration.
5) pg_stat_monitor_reset_errors testing using the tap testcases.
6) Insufficient shared space and buffer overflow testing via tap testcases.
7) Some sql scripts under 'scripts' folder to generate some work load requried
for tap test cases.
8) Everything under 't' folder is specific to perl based test cases. It houses
perl files and folders for some expected files and result folder.
9) 90%+ code coverage for LOC and functions.
10) PG-339 Fix by diego, change in pgsm_errors.c.
After fixing the problem with utility statements, this whole block:
do $$
declare
n integer:= 1;
begin
loop
PERFORM a,b,c,d FROM t1, t2, t3, t4
WHERE t1.a = t2.b AND t3.c = t4.d ORDER BY a;
exit when n = 1000;
n := n + 1;
end loop;
end $$;
Is only processed once, as those are nested statements, in order to
match the 1000 statements the GUC pg_stat_monitor.track must be set to
'all' and then back to the default of 'top' when done testing it.
This commit adds following three sql based testcases:
1) Test unique application name set by user.
2) Histogram function is working properly as desired.
3) Error on insert is shown with proper message.
We are concerned with finished queries in pg_stat_monitor, so state and
state_code columns were removed from the pg_stat_monitor_view as we only
list finished queries on it.
Removed state regression as it is not necessary anymore.
Also, this allowed us to remove the call to pgss_store on ExecutorStart,
which just updated query state to EXEC, thus saving some CPU.
The issue is that between changing GUC "track" from track='top' to
track='all' the queries are executing using previous state of
track='top', to fix that we sleep 1 second after calling
pg_reload_conf() to ensure that queries will run with new settings.
cmd_type: Added missing DROP TABLE t2;
guc: Adjusted to match the updated settings view, which now display
boolean values as 'yes' and 'no', also added the 'options' column
to the output.
guc_1: Handle PostgreSQL versions <= 12 which don't have the
track_planning feature.
rows.out: Added missing DROP TABLE t2. Also removed the line 'ERROR:
relation "t2" already exists' since we fixed the problem in
cmd_type regression.
top_query: Handling both track = 'top' and track = 'all' cases.
top_query_1: On PostgreSQL >= 14 the sub query from the procedure is
stored as (select $1 + $2), whereas on PG <= 13 it is
stored as SELECT (select $1 + $2).
As the query normalization and query cleaning is always done in the
right place (pgss_store), no more parsed queries have a trailling comma
';' at the end.
Also, on error regression test, after fixing some problems with utility
related queries, we now have two entries for the RAISE WARNING case, the
first entry is the utility query itself, the second entry is the error
message logged by emit_log_hook.
Some queries have the order adjusted due to the fix introduced by the
previous commits.
The regression tests required some adjustmentes as they were based on a
wrong behavior in pg_stat_monitor that was fixed in the last commits.
The problem was that pg_stat_monitor_reset() was not properly clearing
the query buffers, as such, some garbage queries were residing in the
buffers after calling pg_stat_monitor_reset().
One example of a problem, a query such as "SELECT 1 AS num" and the
same query with comments such as:
SELECT $1 AS num /* { "application", psql_app, "real_ip", 192.168.1.3)
*/
Are evaluated to the same query ID, if a test issue the first query, call
pg_stat_monitor_reset() to clear query buffer, then issue the second
query with comments, the result in pg_stat_monitor view would still contain
the first query without comments, this was leading to tests expecting
the wrong output, which is now fixed.
1) Enabled configure and build with proper flags and environment to make sure that
built server is aligned with pg and ppg package distribution in terms of features
and configurations. Earlier build (configure) was not aligned with pg community standard
configuration that are used for community builds and distribution.
2) Enabled installcheck-world regression test suites of the pg server, to verify
the stability and compatibility of pg server after loading pg_stat_monitor in
server.
3) Change in expected files of error.out and error_1.out for error testacase.
4) Disbaled and removed coverage using coveralls.io, as it was not serving the purpose.
(Note: installcheck-world was failing on pg-14 due to some changes that are done in pg14
or upstream PGSS, and that is causing additional line (Query Identifier) is output of some
of the test cases of installcheck-world so it is not enabled in this commit. Problem with
pg14 server installcheck regression is output of an extra line (Query identifier ****)
in some of the test cases after loading extension PGSM and that causes regression to fail
for server after library load.)
Changed the default value of pgsm_bucket_time from 300 to 60. Now neither PMM users will need to adjust the default value, nor
restart of PG server will be required for this purpose.
Naeem Akhter
Ibrar Ahmed
Diego Fronza
Evgeniy Patlan
Naeem Akhter
Mikhail Samoylov