Date: Fri, 25 Apr 2014 12:46:10 +0100 From: "Steven Hartland" <killing@multiplay.co.uk> To: "Robert Blacquiere" <freebsd-net@blacquiere.nl>, <freebsd-net@freebsd.org> Subject: Re: NFS over LAGG / lacp poor performance Message-ID: <6312D8095E3B480DAEDDAB9C39A4B581@multiplay.co.uk> References: <535A1354.2040309@wp.pl> <20140425111316.GJ9177@calendar.blacquiere.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
----- Original Message -----
From: "Robert Blacquiere" <freebsd-net@blacquiere.nl>
> Hi Marek,
>
> On Fri, Apr 25, 2014 at 09:48:36AM +0200, Marek Salwerowicz wrote:
>> Hi list,
>>
> <snip>
>> Both boxes are connected to the same switch (HP 1910-48G)
>>
>> I need to transfer around 10 TB of data from storage1 to storage2
>> I obeserve that during copying, only one NIC (instead of all 4) is used.
>>
>> Boxes are not stressed during copying
>>
>> What's more, apart from having 1 NIC saturated (transfer around 120
>> MB/s), I observe transfer rate on level of 70-80 MB/s
>
> Default lacp 802.3ad works with mac based hashes to loadbalance traffic.
> So single host (mac) will be transfered by one ethernet adaptor as you
> have seen.
In FreeBSD 10 onwards the default is l2,l3 & l4 hash this is not true for
the switch side, so may need to tune the switch.
Regards
Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6312D8095E3B480DAEDDAB9C39A4B581>