RFC: slow-down option

[Marian Marinov]

On 04/03/2014 02:48 PM, Christoph Biedl wrote:
> Marian Marinov wrote...
>
>> I've been using rsync on some backup servers for years. In 2011 we
>> had a situation where the FS of the backup server was behaving
>> strange, even thou there was enough available I/O, the fs(ext4 on
>> 16TB partition with a lot of inodes) was lagging. After much testing
>> we found that rsync was hammering the fs too hard.
>
> I'd like to learn more about that scenario. Mostly, I'm curious
> whether these file transfers involved creation of a *lot* and probably
> rather small files.

The files were *mostly* small files under 10k. We had around 10 concurrent rsyncs running at any given time.
The fs was build on top of RAID6 array. But unfortunately I did not create the journal on separate device and actually 
did not left any space for separate device for the journal, which in my opinion would help in that situation.

When I did the patch, I was also thinking about a more dynamic slowdown, one that would take into account the size of 
the file that was previously transmitted or the size of the file that will be transmitted.

Marian
>
>      Christoph
>