A macOS Sequoia bug in Apple's fsck_hfs utility falsely reports corruption on large HFS+ volumes. Users with 8GB RAM machines and 24TB+ external drives encounter "Couldn't read node" errors during extended attribute checks. The filesystem remains intact—Apple's disk check tool contains a cache exhaustion problem affecting specific hardware configurations. Data integrity remains secure despite the misleading error messages.

The bug occurs in fsck_hfs's cache management when processing unused nodes in the Attributes B-tree. Apple's open-source code reveals the tool races through thousands of free nodes, allocating cache tags faster than the LRU eviction system can reclaim buffers. On 8GB systems with a 1GB cache, all buffers eventually become held by active tags, causing CacheLookup to return ENOMEM error 12.

Machines with 16GB+ RAM avoid the issue due to larger cache pools. Apple should address the race condition in the LRU eviction mechanism. Users experiencing this problem can verify their filesystem integrity through alternative methods or upgrade to systems with more RAM until Apple releases a patch for the fsck_hfs utility in macOS Sequoia.