stripe-alloc: STRIPE-META-CLAIMS.md, evidence for dropping stripe_meta
Verdicts on Zygo's four claims for dropping stripe_meta from the series,
from existing benchmarks, the series history and the mailing list (no
new tests): not faster than raid1 (holds in practice), trapping eats
the space saving (holds on aged and full filesystems), high complexity
and stall/deadlock risk (holds), overlap with Johannes Thumshirn's
zoned ENOSPC work (same problems, different mechanisms, no code
collision). Ends with the cost of dropping it: 14 patches, about
+1211/-130 lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B6wxsJn1aY7PwnhgXdz6Ec