[ovs-dev] [PATCH ovn v3] tests: Fix occasional failures for test 78

[Dumitru Ceara]

On 4/17/20 11:25 PM, Flavio Fernandes wrote:
> The test case "78: ovn -- 1 LR with distributed router gateway port"
> fails sometimes. This patch fixes this by leveraging the command
> ovn-nbctl sync

Hi Flavio,

Maybe we should change the commit title to something more specific
because test numbers change when tests are added/deleted?

A few minor comments below.

Regards,
Dumitru

> 
> The failed test would look like this:
>    <snip>
>    grep "Port patch-br-int-to-ln-alice" | wc -l`...
>    ovn.at:9356: wait succeeded immediately
> 
>   checking packets in hv3/vif1-tx.pcap against hv3-vif1.expected:
>   ovn.at:12: waiting until $PYTHON "$ovs_srcdir/utilities/ovs-pcap.in" $rcv_pcap > $rcv_text
>           rcv_n=`wc -l < "$rcv_text"`
>           echo "rcv_n=$rcv_n exp_n=$exp_n"
>           test $rcv_n -ge $exp_n...
>   rcv_n=0 exp_n=2
>   <snip>
>   rcv_n=2 exp_n=2
>   ovn.at:12: wait succeeded after 5 seconds
>   ./ovn.at:9367: sort $rcv_text
>       expout      2020-04-17 20:17:08.255531551 +0000
>       /home/vagrant/ovn/tests/testsuite.dir/at-groups/78/stdout   2020-04-17 20:17:08.258531517 +0000
>      -1,2 +1,2
>   -f0000001020400000201020308060001080006040002000002010203ac100101f00000010204ac100103
>   +ffffffffffff00000201020308060001080006040001000002010203ac100101000000000000ac100101
>    ffffffffffff00000201020308060001080006040001000002010203ac100101000000000000ac100101
>   78. ovn.at:9139: 78. ovn -- 1 LR with distributed router gateway port (ovn.at:9139): FAILED (ovn.at:9367)
> 
> Signed-off-by: Flavio Fernandes <flavio at flaviof.com>
> ---
>  tests/ovn.at | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/ovn.at b/tests/ovn.at
> index f83d3f536..e73772300 100644
> --- a/tests/ovn.at
> +++ b/tests/ovn.at
> @@ -9239,9 +9239,9 @@ as hv1 ovs-vsctl set open . external-ids:ovn-bridge-mappings=phys:br-phys
>  as hv3 ovs-vsctl set open . external-ids:ovn-bridge-mappings=phys:br-phys
>  
>  
> -# Allow some time for ovn-northd and ovn-controller to catch up.
> -# XXX This should be more systematic.
> -sleep 2
> +dnl Allow some time for ovn-northd and ovn-controller to catch up.
> +ovn-nbctl --wait=sb --timeout=3 sync
> +ovn-nbctl --wait=hv --timeout=3 sync

"--wait=hv" should be enough because it implies "--wait=sb".

Also, we can skip the "--timeout=3" because OVS_CTL_TIMEOUT is set when
running tests and its value will be used as default.

>  
>  echo "---------NB dump-----"
>  ovn-nbctl show
> @@ -9356,6 +9356,10 @@ as hv2 ovs-vsctl set open . external-ids:ovn-bridge-mappings=phys:br-phys
>  OVS_WAIT_UNTIL([test 1 = `as hv2 ovs-vsctl show | \
>  grep "Port patch-br-int-to-ln-alice" | wc -l`])
>  
> +dnl Allow some time for ovn-controller to catch up.
> +ovn-nbctl --wait=hv --timeout=3 sync
> +ovn-nbctl --wait=sb --timeout=3 sync
> +

Same here.

>  # ARP for router IP address from outside1
>  test_arp 3 1 f00000010204 $outside_ip $rtr_ip 000002010203
>  
>