Re: [PATCH - RFC] new "nosharetransport" option for NFS mounts.

[NeilBrown <neilb@suse.de>]

[-- Attachment #1: Type: text/plain, Size: 1182 bytes --]

On Tue, 9 Jul 2013 13:22:53 +1000 NeilBrown <neilb@suse.de> wrote:


> A customer upgraded from SLES10 (2.6.16 based) to SLES11 (3.0 based) and saw
> a slowdown on some large DB jobs of between 1.5 and 2 times (i.e. total time
> 150% to 200% of what is was before).
> After some analysis they created multiple virtual IPs on the server and
> mounted the several filesystem each from different IPs and got the
> performance back (they see this as a work-around rather than a genuine
> solution).
> Numbers are like "500MB/s on a single connection, 850MB/sec peaking to
> 1000MB/sec on multiple connections".
> 
> If I can get something more concrete I'll let you know.

Slightly more concrete:

 4 mounts from the one server, 10 threads of fio on each mount.
All over a 10G Ethernet.

1 IP address without "nosharetransport":  ~700MB/s
4 IP addresses without "nosharetransport": ~1100MB/s
1 IP address with "nosharetransport": ~1100MB/s

This is all NFSv3.  NFSv4 is much slower with nosharecache (32MB/s!), but I
might have botched the backport to 3.0 (or didn't address the v4 specific
issues you raised).  I haven't looked into that yet.


NeilBrown

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]