On 24 June 2014 12:24, Rich <rerc...@acm.jhu.edu> wrote: > Not normal. > > What platform and version are we on? OmniOS? SmartOS? Nexenta? Other?
Based on OpenIndiana 151a9 but locally compiled with a couple of upstream bugfix patches applied related to mpt_sas and zfs. None of the patches touching usr/src/uts/common/io are network related. > There were some fixes for how the aggr distribution worked at some point IIRC. > > - Rich > > On Mon, Jun 23, 2014 at 8:51 PM, Kris Shannon via illumos-discuss > <discuss@lists.illumos.org> wrote: >> We have some Intel 82599ES cards (ixgbe driver 10Gbit cards) >> >> When we assign an IP directly to one of these nics we can consistently >> achieve near-linerate speeds with iperf (e.g. 9814Mbit) >> >> If we create an aggregate (even a one port aggregate) and assign the >> IP to it, our iperf throughput drops to 480Mbit. >> >> Is this normal? All the old blog posts about Nemo/GLDv3 seemed to >> imply that the aggegate should have minimal overhead. >> >> >> ------------------------------------------- >> illumos-discuss >> Archives: https://www.listbox.com/member/archive/182180/=now >> RSS Feed: https://www.listbox.com/member/archive/rss/182180/21253829-71dc39ef >> Modify Your Subscription: https://www.listbox.com/member/?&; >> Powered by Listbox: http://www.listbox.com ------------------------------------------- illumos-discuss Archives: https://www.listbox.com/member/archive/182180/=now RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be Modify Your Subscription: https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4 Powered by Listbox: http://www.listbox.com
