Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 10 Sep 2010 20:50:55 -0400 (EDT)
From:      Rick Macklem <rmacklem@uoguelph.ca>
To:        gil@vidals.net
Cc:        freebsd-fs@freebsd.org
Subject:   Re: ZFS write speed performance tuning (NFS)
Message-ID:  <1639122954.761765.1284166255552.JavaMail.root@erie.cs.uoguelph.ca>
In-Reply-To: <AANLkTikv8H6uj12fXqXrcvMWeD-LhsUHwbQRjh9fS-zM@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
> I'm building my first ZFS NAS and I'm struggling with write speeds
> over NFS.
> I want to know if it is possible to somehow set NFS to async AND keep
> slogging active with my SSD write cache?
> 
> Here is a brief run down of my write speeds on a 1 gigabit LAN and
> FreeBSD
> 8.1 using ZFS:
> 
> WRITE SPEEDS of 1 GB file:
> 55 MB/sec: scp copy test
> 28 MB/sec: NFS with async (zil_disable="1")
> 12 MB/sec: NFS with sync (default settings)
> 
> 3 x 1TB 7200 RPM SATA HDs
> CPU Intel i7
> Controller: LSI 1068e
> Controller: ICH10R (for read and write SSDs)
> 
> This NAS will eventually go into my VMware environment and
> unfortunately ESX
> uses o_sync in their NFS client. AFAIK, there is no way to change the
> VMware
> NFS client, so I thought of turning off async on the server side and
> yet
> somehow still use the ZIL SSD Cache for fast writes?
> 
> Looking forward to learning more about ZFS.
> 
> --Gil Vidals
> 
It sounds like you may already be doing this, but the following email
suggests that the log be on a separate SSD device.
>From freebsd-fs archive:http://docs.FreeBSD.org/cgi/mid.cgi?4ABD56D6.50301




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1639122954.761765.1284166255552.JavaMail.root>