>I should point out there have been some projects underway in that space,
>including a (I believe defunct) logging filesystem project, and the
The logging filesystem it's not died in my mind. But the more I think at
it, the more I am worried about the amound of extra work the backing
storage has to do when the fs reach the wrap. Sometime in the future I
think I'll sure try it to see how a logfs will scale, maybe a mimimal lazy
implementation just to produce some number during reads and during writes
while the log is wrapping.
Right now I think a journaled fs can blame a logging fs in global
performances (I am not talking about write performances while the log is
not wrapping around).
I planned to have a kernel daemon that will defrag data and move the start
of the log far from the end of the log while the HD is idle but the point
is that it's an huge amount of extra work for the system anyway... and if
the write is faster than the defragger we'll have to wait the
defrag-kernel-daemon to complete I/O to continue writing.
The good point is that a loggingfs will automagically preserve also data
coherency,
Andrea Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/