test: flaky election_basic.test.lua test

Found that on slow test hosts like FreeBSD VMware test flaky failed, like:

  [001] @@ -153,7 +153,7 @@
  [001]   | ...
  [001]  assert(leader_count == 1)
  [001]   | ---
  [001] - | - true
  [001] + | - error: assertion failed!
  [001]   | ...
  [001]  -- All nodes have the same leader.
  [001]  r1_leader = test_run:eval('election_replica1', leader_id_cmd)[1]

It happened because there was not enough time right after wait_fullmesh()
call and before check of the current leader to make any of the replicas
leader. Later in the code was not correct logic when checked who is the
leader and choosed 'election_replica3' if not the others, it caused the
later fails in the test. Decided to wait after wait_fullmesh() when any
of the replicas became leader using test_cond() routine.

Closes #5368