fix(pd,store): fg mode exit code propagation in startup scripts

[bitflicker64]

Fix foreground mode in start-hugegraph-pd.sh and start-hugegraph-store.sh (chunks 2–3 of #3043)

Purpose of the PR

• relates to [Bug] Docker containers never restart when Java process crashes — broken cron monitor + tail -f /dev/null supervision #3043

Main Changes

Problem

Both start-hugegraph-pd.sh and start-hugegraph-store.sh had no foreground
branch — the scripts always backgrounded Java with exec ... &, wrote $! to
the pid file, and exited 0, losing Java's exit code entirely. Docker/systemd
supervisors never saw a non-zero exit, so containers were never restarted on
Java crash.

Fix

Add DAEMON="true" default and -d true|false flag to getopts in both scripts.
Daemon branch keeps exec ... & with $! as before. Foreground branch writes
$$ to the pid file before exec (exec replaces the shell with Java, so
$$ == Java's PID after exec), then exec java without & so the process
blocks and Java's exit code propagates out directly.

No trap needed in the foreground branch — exec replaces the shell process with
Java, so signals from Docker/systemd go directly to Java without a wrapper to
forward through (unlike chunk 1 where & + wait required an explicit trap).

Tests

Script	Test file	Tests	Baseline	After fix
start-hugegraph-pd.sh	test-start-hugegraph-pd.sh	4 (daemon regression, foreground blocks, exit code 137, SIGTERM 143)	3 passed, 9 failed	12 passed, 0 failed
start-hugegraph-store.sh	test-start-hugegraph-store.sh	4 (same)	3 passed, 5 failed	11 passed, 0 failed

Both test scripts wired into pd-store-ci.yml (no backend guard needed — pd/store jobs always run).

Note: Store health check is skipped when PD is not running — handled gracefully with a warning, not a failure.

What's NOT in this PR

Docker entrypoints, HEALTHCHECK, and cron monitor removal are in chunks 4–8 (separate PR).

Verifying these changes

• Need tests and can be verified as follows:
  • test-start-hugegraph-pd.sh $PD_DIR — 12 assertions, all pass after fix
  • test-start-hugegraph-store.sh $STORE_DIR — 11 assertions, all pass after fix

Does this PR potentially affect the following parts?

• Dependencies
• Modify configurations
• The public API
• Other affects
• Nope

Documentation Status

• Doc - No Need

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 0.07%. Comparing base (d6e6216) to head (a6c38e8).
⚠️ Report is 1 commits behind head on master.

❗ There is a different number of reports uploaded between BASE (d6e6216) and HEAD (a6c38e8). Click for more details.

HEAD has 2 uploads less than BASE

Flag	BASE (d6e6216)	HEAD (a6c38e8)
	3	1

Additional details and impacted files

@@             Coverage Diff              @@
##             master   #3047       +/-   ##
============================================
- Coverage     36.07%   0.07%   -36.00%     
+ Complexity      338      22      -316     
============================================
  Files           803     781       -22     
  Lines         68095   65712     -2383     
  Branches       8918    8515      -403     
============================================
- Hits          24563      51    -24512     
- Misses        40897   65659    +24762     
+ Partials       2635       2     -2633     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[imbajin]

I re-reviewed the current head a6c38e8 and did not find any blocking issue in the PD/Store foreground-mode changes.

The implementation preserves the existing daemon behavior by default, while -d false writes the current shell PID before exec, so the Java process owns the same PID and exit/signal behavior propagates to the caller. The new CI foreground suites also exercise the important process semantics: blocking behavior, pid-file behavior, SIGKILL exit propagation, and SIGTERM termination for both PD and Store.

A couple of non-blocking follow-ups may still be worth considering:

1. The PD/Store README files do not document the new -d true|false option yet.
2. The Store foreground test mainly validates the process model; it intentionally tolerates health-check timeout when PD is not running, while the later workflow steps still cover the normal PD+Store path.

Overall this looks merge-safe to me as the foreground-mode chunk for #3043. The Docker entrypoints still need to be wired to this foreground mode in a follow-up change before the original Docker restart/lifecycle issue is fully closed.