Re: [tlug] Questions about specifications for some cutting edge video recording via laptop

[Ian Wells <ijw@example.com>]

On 9/12/05, Stephen J. Turnbull <stephen@example.com> wrote:

>>>>> "Ian" == Ian Wells <ijw@example.com> writes:

    Ian> If you write two files to the same disk the FS might choose
    Ian> to put them in different locations on the disk. Thus the head
    Ian> of the drive moves forward and backward between the two
    Ian> locations as the files are written.

True, but with the buffering you should be in pretty good shape.  ISTR
that the Linux VFS is pretty good at scheduling writes, so with fast
hardware Dave might be OK.

Possibly, but there will still be more seeks involved than with writing a single file, surely, as any FS would try and make the files (rather than the writes) as sequential on disk as possible.  Or maybe they block up in memory before writing out, I suppose I can't really jump to conclusions...  Still an area worth checking.

    Ian> Streaming files will behave differently to a cp -r, hence I'm
    Ian> not sure what will happen.

You might want to use dd instead of cp in the experiment, for better
control over blocking.

I'm mixing and matching evidence to suit my theory ;-)  cp -r really does slow things up. dd may do less so, but my Linux machine is (a) dated and (b) not currently booting, so I'm working with what I've seen rather than doing a lot of tests.

--
Ian.