Vitaly Zaitsev via devel wrote on Sun, Jul 12, 2020:
> On 11.07.2020 14:20, Dominique Martinet wrote:
> > BTW, given the size gains ws. time difference for compression I would
> > advocate for default zstd compression instead of :1 -- I'd think another
> > 12% compression improvement[1] for almost no time difference isn't to be
> > sneezed at?
> Now please open this file again and check Data Operations table.

The passage you quoted has nothing to do with the paper, but about
another part of the thread.

Note I never said this paper wasn't about data, nor that btrfs is good
wrt write amplification in the access pattern they do in this paper ; I
said you might want to think whether a pattern with small writes and
fsyncs in a cow filesystem is a good idea and is going to be
representative of a typical user's workload.

That is precisely why previously in the thread things like databases
have been discussed along with the nocow chattr flag.

(Yes, that means applications need to start being concious of what fs
they are being run on, or at least the fedora configuration needs to do
that check for them)

devel mailing list --
To unsubscribe send an email to
Fedora Code of Conduct:
List Guidelines:
List Archives:

Reply via email to