Currently the "testing drop-stats - sampling action" test terminates
vswitchd before error messages can be logged. If the testsuite exit
hangs for a second then vswitchd will log some more errors and the
test will fail at check_logs.
./drop-stats.at:324: check_logs "/sending to collector failed/d"
dpif(revalidator18)|WARN|dummy@ovs-dummy: failed to flow_del (No
such file or directory) ufid:<UUID>
dpif(revalidator15)|WARN|dummy@ovs-dummy: failed to flow_del (No
such file or directory) ufid:<UUID>
This can be simulated by inserting a sleep in check_logs. These errors
are caused by the revalidator sweeping flows that were already deleted
through ovs-appctl dpctl/del-flows. This patch adds adds a revalidator
purge before flow deletion to suppress these log messages.
Fixes: 516569d31fbf ("ofproto: xlate: Make sampled drops explicit.")
Reported-at: https://redhat.atlassian.net/browse/FDP-1812
Signed-off-by: Mike Pattrick <[email protected]>
---
tests/drop-stats.at | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tests/drop-stats.at b/tests/drop-stats.at
index 748eaed35..a28a88056 100644
--- a/tests/drop-stats.at
+++ b/tests/drop-stats.at
@@ -287,6 +287,7 @@ AT_CHECK([ovs-appctl time/warp 5000], [0], [ignore])
CHECK_COVERAGE([drop_action_of_pipeline], [3])
+AT_CHECK([ovs-appctl revalidator/purge])
AT_CHECK([ovs-appctl dpctl/del-flows])
for i in $(seq 1 3); do
@@ -304,6 +305,7 @@ AT_CHECK([ovs-appctl time/warp 5000], [0], [ignore])
CHECK_COVERAGE([drop_action_of_pipeline], [6])
+AT_CHECK([ovs-appctl revalidator/purge])
AT_CHECK([ovs-appctl dpctl/del-flows])
for i in $(seq 1 3); do
--
2.54.0
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
