pgpool: Test: stablize 006.memqcache regression test.

Test: stablize 006.memqcache regression test.

The test sometimes fails on slower environment like build farm.
I suspect one of the causes was replication delay. From the regression log:

2026-09-17 20:21:19.129: child pid 18495: LOG: pool_read_kind: error message from 1 th backend:role "foo" does not exist
2026-09-17 20:21:19.129: child pid 18495: ERROR: unable to read message kind
2026-09-17 20:21:19.129: child pid 18495: DETAIL: kind does not match between main(53) slot[1] (45)

It seems when a new role was created, standby did not catch up
it. This commit sets up synchronous replication in streaming
replication mode to avoid the delay.

Author: Tatsuo Ishii <ishii(at)postgresql(dot)org>
Backpatch-through: v4.3

Branch
------
V4_4_STABLE

Details
-------
https://git.postgresql.org/gitweb?p=pgpool2.git;a=commitdiff;h=2d9f72de1221670c421302f576d110db44a40d28

Modified Files
--------------
src/test/regression/tests/006.memqcache/test.sh | 10 ++++++++++
1 file changed, 10 insertions(+)