I think you're right. I'll write up a patch.
Ethan
On Wed, Jan 29, 2014 at 4:39 PM, YAMAMOTO Takashi
wrote:
>> Before this patch, OVS tried to guess an optimal max idle time for
>> datapath flows based on the number of datapath flows relative to the
>> limit. This caused instability because th
> Before this patch, OVS tried to guess an optimal max idle time for
> datapath flows based on the number of datapath flows relative to the
> limit. This caused instability because the limit was based on the
> dump duration which was affected by the max idle time. This patch
> chooses instead to
Looks good to me. This certainly makes it easier to get reproducible
performance results.
Acked-by: Joe Stringer
On 27 January 2014 16:50, Ethan Jackson wrote:
> Before this patch, OVS tried to guess an optimal max idle time for
> datapath flows based on the number of datapath flows relative
Before this patch, OVS tried to guess an optimal max idle time for
datapath flows based on the number of datapath flows relative to the
limit. This caused instability because the limit was based on the
dump duration which was affected by the max idle time. This patch
chooses instead to hardcode t
