[mirror-admin] Breaking the 1TB limit?

Arnaud Gomes-do-Vale Arnaud.Gomes at ircam.fr
Mon Feb 22 18:16:54 EST 2010


Dax Kelson <dkelson at gurulabs.com> writes:

> Are the "big slow" disks really the bottleneck compared to WAN
> bandwidth? Are file requests so random that the memory cache can't
> sufficiently speed up the "big slow" disks?

They are, but the actual performance hit is not so bad. That is, last
time I hit the issue was last year, I think it was for the release of
Fedora 11; bandwith capped around 600-650 Mbps on a 1Gbps link for a
couple of days. At the time the box had 5 1TB SATA drives in a soft
RAID-5 array and 10GB RAM. I have since upgraded to 5*2TB and 14GB so it
*should* get a little faster (yes, bigger drives are actually faster,
though this might be offset by a larger dataset relative to cache size)
but I have not had any chance to test it.

-- 
Arnaud

--


More information about the Mirror-admin mailing list