PG-1842 PG-1843 Optimize deletion of leftover relation keys

[jeltz]

Instead of first deleting any leftover key and then writing the new key we do a single pass through the file where we replace any old key that we find. To make this happen on redo too we need to stop generating a separate WAL record for the key deletion for encrypted tables and only generate that record for unencrypted tables where we still need a key deletion record.

We except this optimization to primarily be visible on WAL replay where only a single backend is used to replay everything, but it also speeds up table creation in general on workloads with many tables.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 82.22%. Comparing base (13c1038) to head (d94c696).

❌ Your project status has failed because the head coverage (82.22%) is below the target coverage (90.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@                  Coverage Diff                  @@
##           TDE_REL_17_STABLE     #529      +/-   ##
=====================================================
+ Coverage              82.21%   82.22%   +0.01%     
=====================================================
  Files                     25       25              
  Lines                   3177     3179       +2     
  Branches                 516      517       +1     
=====================================================
+ Hits                    2612     2614       +2     
+ Misses                   456      455       -1     
- Partials                 109      110       +1     

Components	Coverage Δ
access	83.42% <100.00%> (-0.05%)	⬇️
catalog	87.60% <ø> (ø)
common	77.77% <ø> (ø)
encryption	72.97% <ø> (ø)
keyring	73.21% <ø> (ø)
src	94.15% <ø> (ø)
smgr	95.90% <100.00%> (+0.53%)	⬆️
transam	∅ <ø> (∅)

🚀 New features to boost your workflow:

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