> Read-ahead still seems to have a big leak somewhere. With 10GB dd reads,
> it caused my glusterfsd (which was configured for read-ahead) to grow
> slowly for a little while, until suddenly the growth became very rapid
> (probably about the same as the data it was passing) and glusterfsd died
> (just like before).
did you try with read-ahead on srever or client side??
avati
--
Shaw's Principle:
Build a system that even a fool can use,
and only a fool will want to use it.
