I welcome any fix this improves the intermittent testsuite failures :D On 23 December 2014 at 09:03, Alex Wang <al...@nicira.com> wrote: > This commit fixes the intermittent failure of test (ofproto-dpif, > balance-tcp bonding, different recirc flow), by waiting for all > bond interfaces to be enabled, and by waiting for datapath flow > installation finish before purging it. > > Signed-off-by: Alex Wang <al...@nicira.com> > --- > tests/ofproto-dpif.at | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at > index 55c3e90..1dc971d 100644 > --- a/tests/ofproto-dpif.at > +++ b/tests/ofproto-dpif.at > @@ -174,6 +174,9 @@ OVS_VSWITCHD_START( > AT_CHECK([ovs-appctl netdev-dummy/set-admin-state up], 0, [OK > ]) > > +# Waits for all ifaces enabled. > +OVS_WAIT_UNTIL([test `ovs-appctl bond/show | grep -- "may_enable: true" | > wc -l` -ge 4]) > + > # The dl_vlan flow should not be ever matched, > # since recirculation should not change the flow handling. > AT_DATA([flows.txt], [dnl > @@ -186,6 +189,8 @@ AT_CHECK([ovs-ofctl add-flows br-int flows.txt]) > # Should generate recirc_id(0x12d),dp_hash(0xc1261ba2/0xff). > AT_CHECK([ovs-appctl netdev-dummy/receive p5 > "in_port(5),eth(src=50:54:00:00:00:05,dst=50:54:00:00:01:00),eth_type(0x0800),ipv4(src=10.0.0.2,dst=10.0.0.1)"]) > > +ovs-appctl time/warp 5000 100 > + > # Collects flow stats. > AT_CHECK([ovs-appctl revalidator/purge], [0])
If we're waiting on flow installation to finish, do you think that it might be a little more robust to use something like the below? ovs-appctl vlog/set dpif:dbg (send traffic...) OVS_WAIT_UNTIL(test `cat ovs-vswitchd.log | FILTER_FLOW_DUMP | wc-l` -ge 4]) Otherwise, LGTM Acked-by: Joe Stringer <joestrin...@nicira.com> _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
