On Mon, Dec 28, 2015 at 08:54:29PM +0100, Arnt Karlsen wrote:
> On Mon, 28 Dec 2015 19:37:08 +0100, Didier wrote in message
> <56818154.2070004@???>:
>
> > Le 28/12/2015 19:22, Simon Hobson a écrit :
> > > The cache gets written out when the background system processes
> > > clean up and write the dirty pages out to disk. How long this takes
> > > depends on tuneable kernel parameters and how busy the system is.
> > > If the system, and in particular the storage, is otherwise idle
> > > then IIRC your small file will get written almost instantly. If the
> > > system is really busy, with a large dirty cache, then it'll take a
> > > lot longer.
> >

> >      That's the logic one would naively expect but I'm not sure of
> > it. I'm afraid the data remains in the cache and  not backed-up to
> > disk until some process needs room in the cache. You can do the
> > experiment of writing data to a usb memory stick and then wait long
> > after the light has stopped blinking. Then you can either sync or
> > umount the device and it will blink again for  some time before the
> > command returns.

> >
>
> ..probably the best way is keep those flashy blinky lights blinking
> merrily away to try scare the idiots from yanking the removables,
> while we try finish _all_ the writing to try prepare the media for
> the inevitable yank, and only then stop the blinky lights and show
> the "Finished writing to disk, you may now safely remove the disk.
> Ok." etc eye candy.

(1)

On the Commodore Amiga, users were warned not to remove the floppy disks
while the disk light was on.

That seemed to work. But the disks were written almost immediately, and
the OS made sure the disk light stayed on ontil the disk cache was
cleared.

(2)

Isn't there some kind of mount option saying that buffers should be
flushed immediately when written to?

-- hendrik