On Tue, 2017-05-30 at 09:47 -0700, Ben Pfaff wrote: > On Thu, May 25, 2017 at 01:47:32PM -0700, Joe Stringer wrote: > > On 21 May 2017 at 21:55, Yunjian Wang <wangyunj...@huawei.com> wrote: > > > The memory leak was triggered each time on calling netflow_unref() with > > > containing netflow_flows. And flows need to be removed and destroyed. > > > > > > Signed-off-by: Yunjian Wang <wangyunj...@huawei.com> > > > --- > > > > Does this also need to netflow_expire__() the flows? Would it make > > sense for this flow clear path to share code with > > netflow_flow_clear()? > > That would certainly make sense. However, it adds a lot of extra code > to run at the time we're freeing something, which is always a little > risky. Since Greg is having trouble testing this at all, I lean toward > just freeing without trying to send out NetFlow expirations.
I've still had no luck on testing. Over the weekend I ran a continuous loop as suggested by Yunjian that constantly enables and disables netflow and didn't get a break on that line of code. However, I'm looking at adjusting the background network traffic to something that should create more flows and will try again today. - Greg > _______________________________________________ > dev mailing list > d...@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev _______________________________________________ dev mailing list d...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-dev