https://bugzilla.kernel.org/show_bug.cgi?id=107301 --- Comment #15 from Jan Kara <jack@xxxxxxx> --- So I guess there are two separate issues: 1) mbcache scales poorly - that is worth addressing regardless of whether ceph / lustre really need it or not since as you mention there are cases where mbcache helps and scalability is an issue. 2) some usecases do not benefit from mbache much (or at all) and so we could possibly have a heuristic to disable mbcache altogether. My current feeling is that if mbcache is implemented properly, then the overhead of it is a hash-table insertion / deletion when creating / removing xattr and that should be pretty cheap compared to all the other work we do to create external xattr (although cache line contention could be an issue here to some degree). -- You are receiving this mail because: You are watching the assignee of the bug. -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
