PBTree: Fix Dead Lock and Refactor write/update inteface

[bigreybear]

Description

The recent deadlock issues are primarily attributed to the double-bucket index design. Each thread maintains a local collection that organizes pages into buckets based on available space, while the PageManager maintains another gloabl set. When a write thread requires a page for writing/updating, it first checks its local buckets and retrieves pages without locking(as they are already locked). If no suitable page is found locally, the thread then accesses the global buckets, where it must retrieve pages with a lock.

The issue arises when a page, which is appropriate for the expected size and referred by the thread previously, is recorded in the global buckets but not in the local ones. This results in the page being locked twice but unlocked only once, leading to a deadlock. Specifically, if a page increases its available space but is not checked and registered in the local buckets, a deadlock becomes highly probable.

This revision made a comprehensive review of any modifications made to the SchemaPage.spareSize field, which determines the page's suitability for the expected space. All methods that modify this attribute are listed below. Additionally, these methods are now accompanied by statements for recording them in the appropriate buckets

SegmentedPage spare size write-value occurence

1. deleteSegment (handled by previous PR)
2. purgeSegment (handled by previous PR)
3. relocateSegment
   called by:
   3.1 SegmentedPage.write // covered when write within PageManager
   3.2 SegmentedPage.update // covered when update within PageManager
4. reaarangeSegment
   called by:
   4.1 relocateSegment, trace to 3.
   4.2 compactSegment // covered
   compactSegment called by
   4.2.1 allocNewSegment, only called by pre-allocate, now covered with bucket sort
   4.2.2 transplantSegment // covered previously
5. extendSegmentInPlace // only extend last segment so only decrease spare size
6. registerNewSegment // only decrease

Furthermore, this PR also refactore SegmentedPage.write/update inteface so that overflow will return a negative value rather than an exception.

Content1 ...

Content2 ...

Content3 ...

---

This PR has:

• been self-reviewed.
  • concurrent read
  • concurrent write
  • concurrent read and write
• added documentation for new or modified features or behaviors.
• added Javadocs for most classes and all non-trivial methods.
• added or updated version, license, or notice information
• added comments explaining the "why" and the intent of the code wherever would not be obvious
  for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold
  for code coverage.
• added integration tests.
• been tested in a test IoTDB cluster.

---

Key changed/added classes (or packages if there are too many classes) in this PR