Feb 2022

1 commit
Update "don't truncate with failsafe" rationale.
15 Feb 2022Peter Geogheganmaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% surewritten from the discussion, no matching patch2 of 2 reviewers saw the final version

The commit only expands the comment above should_attempt_truncation() in src/backend/access/heap/vacuumlazy.c to explain that truncation can require an unavailable XID before relfrozenxid advances; no attached patch contains that edit. John's investigation in #67 and explanation in #69 supply the rationale, and Peter explicitly announces committing the comment in #72.

  • John Naylorsaw the final versionLLM analysis, 99% sureJohn tested VACUUM under a GetNewTransactionId() breakpoint in #67 and explained the truncation failure in #69, supplying the committed rationale and receiving Reported-By credit.
  • Robert Haassaw the final versionLLM analysis, 88% sureRobert identified truncation's XID allocation as the obstacle to recovery in #52 and endorsed John's resulting analysis in #70, establishing the explanation recorded by the commit.

Credits across 1 commit

Reported by
John Naylor1
Committer
Peter Geoghegan1