fix(server): retry all PD peers while waiting for storage

[bitflicker64]

Kept in sync with hugegraph#175

Purpose of the PR

• close [Bug] wait-storage.sh binds to the first PD answering /v1/health and never fails over to other peers #3123

wait-storage.sh selected the first PD answering /v1/health, then polled
/v1/stores only on that peer. Since the health endpoint can succeed before
the peer has raft-backed store state, a storeless first peer could block Server
startup even when another configured PD already reported an Up store.

Main Changes

• Probe /v1/stores across every configured PD peer on every retry.
• Succeed as soon as any peer reports a store in state Up, without a separate
  /v1/health selection stage.
• Preserve authentication, configured peer order, custom REST endpoints, retry
  timing, and the existing outer timeout failure.
• Add a deterministic shell regression suite and an isolated Server CI job.

Verifying these changes

• Need tests and can be verified as follows:
  • test-wait-storage.sh: 4 scenarios passed, 0 failed.
  • The same suite fails against the unmodified script on the pinned first peer.
  • bash -n passed for the production and test scripts.
  • Server CI workflow YAML parsing and git diff --check passed.
  • Apache RAT and EditorConfig checks passed for the 20-module dependency set.

Does this PR potentially affect the following parts?

• Dependencies
• Modify configurations
• The public API
• Other affects (HStore Server startup readiness)
• Nope

Documentation Status

• Doc - TODO
• Doc - Done
• Doc - No Need

Out of scope

• PD and Store health endpoint semantics
• wait-partition.sh
• Environment-variable and timeout configurability cleanup
• Docker Compose health checks and documentation

[Copilot]

Pull request overview

This pull request fixes wait-storage.sh behavior in multi-PD HStore deployments by avoiding “pinning” to the first PD that answers /v1/health, and instead probing store readiness (/v1/stores) across all configured PD REST peers until any peer reports an Up store. This improves Server startup robustness when some PD peers are listening but not yet raft-ready.

Changes:

• Update wait-storage.sh to retry /v1/stores across every configured PD peer on each retry cycle and succeed as soon as any peer reports "state":"Up".
• Add a deterministic shell regression test suite that mocks curl, sleep, and timeout to validate failover / retry behavior.
• Add a dedicated GitHub Actions job to run the new shell regression suite in CI.

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
hugegraph-server/hugegraph-dist/src/assembly/static/bin/wait-storage.sh	Removes /v1/health gating and probes /v1/stores across all PD peers per retry, succeeding on the first peer that reports an Up store.
hugegraph-server/hugegraph-dist/src/assembly/travis/test-wait-storage.sh	Adds deterministic, mocked shell tests covering peer order, retry/failover, and timeout behavior.
.github/workflows/server-ci.yml	Adds a CI job that runs the new test-wait-storage.sh regression suite.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.