scottk/wlan-lanforge-scripts
1
0
Fork 0
mirror of https://github.com/Telecominfraproject/wlan-lanforge-scripts.git synced 2025-11-03 12:18:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

l3_cxprofile.py : added quiesce_cx option to allow the cx to drain then stop

Browse Source 
realm.py : added drain_stop_cx (could not get quiesce_cx name to work)
test_l3.py : added quiesce_cx

Signed-off-by: Chuck SmileyRekiere <chuck.smileyrekiere@candelatech.com>
This commit is contained in:
Chuck SmileyRekiere
2022-03-28 16:33:34 -06:00
committed by shivam
parent 04ff6d7cb4
commit c6a8627240
3 changed files with 68 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
py-scripts/test_l3.py
View File

@@ -47,6 +47,29 @@ Example using wifi_settings
--endp_type lf_udp --rates_are_totals --side_a_min_bps=20000 --side_b_min_bps=300000000
--test_rig CT-US-001 --test_tag 'test_L3'
NOTES:
command composer : <lanforge ip>:8080/help
Drop RX - in port manager is not the same are layer 3 cx Rx Drop %
Notes from Ben Greear
Port Mgr Drop Rx
unlikely you will see rx drop on port mgr
layer-3 rx drop is related to gaps in rx sequence numbers,
which means network under test dropped frames somehow and/or re-ordered them.
Layer 3 cx Rx Drop %
cx mgr drop% is based on number peer inpoint transmitted vs what we received.
It would ignore reodering problem, and also detect complete path loss
(where rx drop based on seq numbers would not since there is no gap).
But, packets in-flight through network under test may be counted as dropped
when really they are probably not dropped.
you can quiesce a test at the end, which means stop transmitter
but leave received going to drain packets from the network and get better drop% calculations.
Stopping a test would record packets in flight at the moment it is stopped as dropped.
COPYRIGHT:
Copyright 2021 Candela Technologies Inc
@@ -376,13 +399,13 @@ class L3VariableTime(Realm):
total_dl_rate += int(endp["rx rate"])
total_dl_rate_ll += int(endp["rx rate ll"])
total_dl_pkts_ll += int(endp["rx pkts ll"])
dl_rx_drop_percent = int(endp["rx drop %"])
dl_rx_drop_percent = round(endp["rx drop %"],2)
# -B upload side
else:
total_ul_rate += int(endp["rx rate"])
total_ul_rate_ll += int(endp["rx rate ll"])
total_ul_pkts_ll += int(endp["rx pkts ll"])
ul_rx_drop_percent = int(endp["rx drop %"])
ul_rx_drop_percent = round(endp["rx drop %"],2)
return lat, jit, total_dl_rate, total_dl_rate_ll, total_dl_pkts_ll, dl_rx_drop_percent, total_ul_rate, total_ul_rate_ll, total_ul_pkts_ll, ul_rx_drop_percent
@@ -616,6 +639,10 @@ class L3VariableTime(Realm):
(self.station_count, self.cx_count))
# Run the main body of the test logic.
# todo there may be a need to start all existing stations on lanforge
# sta_json = super().json_get(
# "port/1/{resource}/list?field=alias".format(resource=self.resource))['interfaces']
def start(self, print_pass=False):
logger.info("Bringing up stations")
@@ -755,7 +782,7 @@ class L3VariableTime(Realm):
self.reset_port_check()
self.epoch_time = int(time.time())
new_rx_values, rx_drop_percent, endps, total_dl_bps, total_ul_bps, total_dl_ll_bps, total_ul_ll_bps = self.__get_rx_values()
endp_rx_map, endp_rx_drop_map, endps, total_dl_bps, total_ul_bps, total_dl_ll_bps, total_ul_ll_bps = self.__get_rx_values()
log_msg = "main loop, total-dl: {total_dl_bps} total-ul: {total_ul_bps} total-dl-ll: {total_dl_ll_bps}".format(
total_dl_bps=total_dl_bps, total_ul_bps=total_ul_bps, total_dl_ll_bps=total_dl_ll_bps)
@@ -832,8 +859,9 @@ class L3VariableTime(Realm):
# Stop connections.
self.cx_profile.stop_cx()
self.multicast_profile.stop_mc()
# TODO the passes and expected_passes are not checking anything
if passes == expected_passes:
# Sets the pass indication
self._pass(
"PASS: Requested-Rate: %s <-> %s PDU: %s <-> %s All tests passed" %
(ul, dl, ul_pdu, dl_pdu), print_pass)
@@ -995,7 +1023,13 @@ class L3VariableTime(Realm):
self.csv_results_writer.writerow(row)
self.csv_results_file.flush()
# quiesce the cx : quiesce a test at the end, which means stop transmitter
# but leave received going to drain packets from the network and get better drop% calculations.
# Stopping a test would record packets in flight at the moment it is stopped as dropped.
def quiesce_cx(self):
self.cx_profile.quiesce_cx()
# Stop traffic and admin down stations.
def stop(self):
self.cx_profile.stop_cx()
@@ -1676,6 +1710,9 @@ Setting wifi_settings per radio
parser.add_argument('--no_stop_traffic', help='leave traffic running',
action='store_true')
parser.add_argument('--quiesce_cx', help='--quiesce store true, allow the cx to drain then stop so as to not have rx drop pkts',
action='store_true')
# logging configuration
parser.add_argument(
"--lf_logger_config_json",
@@ -1894,8 +1931,8 @@ Setting wifi_settings per radio
quit(1)
station_list = LFUtils.portNameSeries(
prefix_="sta",
start_id_=1 + index * 1000 + int(args.sta_start_offset),
end_id_=number_of_stations + index * 1000 + int(args.sta_start_offset),
start_id_=0 + index * 1000 + int(args.sta_start_offset),
end_id_=number_of_stations - 1 + index * 1000 + int(args.sta_start_offset),
padding_number_=10000,
radio=radio_name_)
station_lists.append(station_list)
@@ -1981,7 +2018,12 @@ Setting wifi_settings per radio
if args.no_stop_traffic:
logger.info("--no_stop_traffic set leave traffic running")
else:
ip_var_test.stop()
if args.quiesce_cx:
ip_var_test.quiesce_cx()
time.sleep(3)
# Quisce cx
else:
ip_var_test.stop()
if not ip_var_test.passes():
logger.warning("Test Ended: There were Failures")
logger.warning(ip_var_test.get_fail_message())