Ken Hirsch wrote:
>
> Daniel Phillips wrote:
>
> > Um, this isn't the only place where support for dumb filesystems messes
> > things up for smart filesystems, it's just the most painful. Dumb
> > filesystems should have the right to remain dumb, but smart filesystems
> > should pay a price for that. It would be so easy to fix this right now,
> > why don't we?
> >
> > Of course, "right now" == 2.5, not 2.4.
>
> On a slightly longer-term topic:
> Has there every been any discussion of a general write-ahead enforcement
> mechanism for Linux?
>
> I'm thinking along these lines--
> A program (resource manager) would attach an open file to a log manager.
>
> Any write to the file would also cause a log record write, which returns a
> strictly increasing log sequence number. That number is associated with
> each
> page in the cache.
>
> Before writing any page to disc, the file system would check to see
> whether the
> page's LSN is <= secure LSN for the log manager of that page's file.
>
> For file systems that support this, it would require
> (1) A way to attach a file to a log manager (ioctl or ...)
> (2) A log manager handle associated with each open file
> (3) Some way to associate an LSN with a write()
> (4) A place to store the LSN for each page in memory
> (5) A way to check the highest secure LSN for the log manager
>
> The log manager might well be writing the log data across a network, rather
> than or in addition to writing it to disk (and it may be on a different file
> system from the data file).
>
> The advantages I see are that database systems and other resource managers
> would have much less complicated buffering logic, that the data would not be
> double-buffered, that the LRU algorithm would be more accurate (there was a
> discussion of this a few months ago with regard to multiple file systems).
> Hopefully we could better performance without having to tune so much.
>
> It would also be a step toward a common open-source transaction manager.
>
> There are issues with regard to checkpointing, but other than that, what do
> you think of this idea?
>
> A good/bad idea in general?
> A good idea but the implementation ... ?
> Yeah, right, try getting these FS implementers to agree. ?
You don't have to get anyone to agree. Code it up and try it, and
welcome to the filesystem party.
--
Daniel
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [EMAIL PROTECTED]
