Pool is 6x striped Stec ZEUSRam as ZIL, 6x OCZ Talos C 230GB drives L2ARC,
and 24x 15k SAS drives striped (no parity, no mirroring) - I know, terrible
for reliability, but I just want to see what kind of IO I can hit.
Checksum is ON - can't recall what default is right now.
Compression is off
Dedupe is off
Trying to figure out vdbench right now, but apparently that's beyond my
abilities at 8:30PM :(
From: Bob Friesenhahn [mailto:bfrie...@simple.dallas.tx.us]
Sent: Tuesday, July 24, 2012 8:13 PM
Subject: Re: [zfs-discuss] IO load questions
On Tue, 24 Jul 2012, matth...@flash.shanje.com wrote:
> ~50,000 IOPS 4k random read. 200MB/sec, 30% CPU utilization on
> Nexenta, ~90% utilization on guest OS. I?m guessing guest OS is
> bottlenecking. Going to try physical hardware next week
> ~25,000 IOPS 4k random write. 100MB/sec, ~70% CPU utilization on
> Nexenta, ~45% CPU utilization on guest OS. Feels like Nexenta CPU is
> bottleneck. Load average of 2.5
> A quick test with 128k recordsizes and 128k IO looked to be 400MB/sec
> performance, can?t remember CPU utilization on either side. Will
> retest and report those numbers.
> It feels like something is adding more overhead here than I would
> expect on the 4k recordsizes/IO workloads. Any thoughts where I should
start on this?
> I?d really like to see closer to 10Gbit performance here, but it seems
> like the hardware isn?t able to cope with it?
All systems have a bottleneck. You are highly unlikely to get close to
10Gbit performance with 4k random synchronous write. 25K IOPS seems pretty
good to me.
The 2.4GHz clock rate of the 4-core Xeon CPU you are using is not terribly
high. Performance is likely better with a higher-clocked more modern design
with more cores.
Verify that the zfs checksum algorithm you are using is a low-cost one and
that you have not enabled compression or deduplication.
You did not tell us how your zfs pool is organized so it is impossible to
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
zfs-discuss mailing list