On 3/17/23 16:04, Ilya Maximets wrote: > On 3/17/23 11:48, Ales Musil wrote: >> The CT flush was enabled by default for every LB, add >> config option called "ct_flush" that allows >> users to enable/disable the CT flush. The CT flush option >> is set to "false" by default. >> >> Reported-at: https://bugzilla.redhat.com/2178962 >> Signed-off-by: Ales Musil <[email protected]> >> --- >> v2: Make the feature opt-in. >> Store the flag in 'struct ovn_controller_lb'. >> v3: Update the NEWS. >> Address nits from Dumitru. >> --- >> NEWS | 4 ++++ >> controller/ovn-controller.c | 6 ++++-- >> lib/lb.c | 1 + >> lib/lb.h | 1 + >> ovn-nb.xml | 6 ++++++ >> tests/ovn.at | 28 +++++++++++++++++++++++++--- >> tests/system-ovn.at | 36 ++++++++++++++++++++++++++++++------ >> 7 files changed, 71 insertions(+), 11 deletions(-) >> > > <snip> > >> diff --git a/ovn-nb.xml b/ovn-nb.xml >> index 73f707aa0..f5d7c4e50 100644 >> --- a/ovn-nb.xml >> +++ b/ovn-nb.xml >> @@ -2041,6 +2041,12 @@ or >> the affinity timeslot. Max supported affinity_timeout is 65535 >> seconds. >> </column> >> + >> + <column name="options" key="ct_flush" type='{"type": "boolean"}'> >> + The value indicates whether ovn-controller should flush CT entries >> + that are related to this LB when the backends are removed. > > This part is talking about beckend removal, but it should probably cover > the removal of the whole LB as well. To be consistent and explicitly > define the behavior. It might not be fully clear what will happen otherwise. > Something like 'Same applies to removal of the load balancer as a whole.' > or 'when the backends or the entire LB record is removed.' > > Thoughts?
Good catch! Yes, we should explicitly mention the LB removal case. > > Best regards, Ilya Maximets. > _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
