Description

+++++++++++++++++++++++++++++++++++++
Problem Description: 
+++++++++++++++++++++++++++++++++++++
TOPOLOGY : 
3 QFX devices connected to each other in a linear topology 

DUT: 
Model: qfx5100-48t-6q
Junos: 21.4R3-S5.17

ISSUE description:

QFX (VLAN-121; IP:10.254.248.2)<--->(ae0) DUT (xe-0/0/1)<---> QFX2( VLAN-121 ; IP: 10.254.248.1)


DUT switch is configured as L2 switch . 

Configuration on ae0 : 

set interfaces xe-0/0/0 ether-options 802.3ad ae0
set interfaces ae0 apply-groups GK-CUSTOMER-CPE-UPLINKTRUNK-L2-INTERFACE-TRUNK
set interfaces ae0 flexible-vlan-tagging
set interfaces ae0 encapsulation flexible-ethernet-services
set interfaces ae0 aggregated-ether-options link-speed 10g
set interfaces ae0 aggregated-ether-options local-minimum-links-threshold 1
set interfaces ae0 aggregated-ether-options lacp active
set interfaces ae0 aggregated-ether-options lacp periodic fast
set interfaces ae0 unit 0 description T=CPE:C=AE-IP:NN=mr01.lab1:NI=ae0:GI=GK-60041:PT=TRUNK:NNI-SERVICE
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN2_CPEMGMT_VLAN
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN99_CLOUDPING_CTR
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN111_CGEI-LAXLAB-CUST1_CTR
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN112_CGEI-LAXLAB-CUST1_CTR
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN113_CGEI-LAXLAB-CUST1_CTR
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN121_SPEgn_SPEpL3_1234
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN122_SPEgn_CTR
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN323_k8s_TEST_BGP
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN324_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN327_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN328_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN329_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN330_IPv6_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN331_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN332_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching vlan members VLAN333_k8s_TEST
set interfaces ae0 unit 0 family ethernet-switching filter input ICMP-ECHO-REQ
set interfaces ae0 unit 1000 vlan-id 1000
set interfaces ae0 unit 1000 family inet address 169.254.1.2/24 arp 169.254.1.3 mac 00:00:00:12:34:56
set interfaces irb unit 2 description T=CPE:NN=mr01.lab1:NI=ae0.2:GI=GK-60051:VLAN=2:CPE-MGMT-VLAN
set protocols uplink-failure-detection group GK-CUSTOMER-CPE-UFD-GROUP-NNI-SERVICE link-to-monitor ae0


Configuration on xe-0/0/1 : 


set interfaces xe-0/0/1 description _csw01.av1.lab1/xe-0/0/1-cf04.laba/et-30/0/2
set interfaces xe-0/0/1 unit 0 family ethernet-switching interface-mode trunk
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members all
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN111_CGEI-LAXLAB-CUST1_CTR
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN121_SPEgn_SPEpL3_1234
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN122_SPEgn_CTR
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN323_k8s_TEST_BGP
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN325_k8s_TEST
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN327_k8s_TEST
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN328_k8s_TEST
set interfaces xe-0/0/1 unit 0 family ethernet-switching vlan members VLAN329_k8s_TEST
set interfaces xe-0/0/1 unit 0 family ethernet-switching filter input ICMP-ECHO-REPLY
set protocols sflow interfaces xe-0/0/1.0



ISSUE: 

- Initiated a rapid ping from QFX1 to QFX2

-Perform commit (even a blank commit) on DUT. DUT is a L2 transit device. 

-We see packet drops in transit traffic. 

{master:0}
labroot@jtac-qfx5100-48t-6q-r2419> ping 10.254.248.1 rapid count 10000    
PING 10.254.248.1 (10.254.248.1): 56 data bytes
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!............!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!^C
--- 10.254.248.1 ping statistics ---
670 packets transmitted, 657 packets received, 1% packet loss
round-trip min/avg/max/stddev = 20.225/39.021/172.010/19.591 ms


Symptoms

OBSERVATIONS: 
Wherenever there is a commit, on DUT, MAC address flushes, this is causing the ping drops: 
root@switch# run show ethernet-switching mac-learning-log 
Jun 07 06:34:38
Tue May 21 03:20:33 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was learned on xe-0/0/0.0 with flags: 0x2001f
Tue May 21 03:20:37 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Mon May 27 07:21:00 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x1000
Mon May 27 07:21:00 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was deleted from xe-0/0/0.0 with flags: 0x1000
Mon May 27 07:24:31 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was learned on xe-0/0/0.0 with flags: 0x2001f
Mon May 27 07:24:31 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Wed Jun  5 07:16:01 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Wed Jun  5 07:16:01 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was deleted with flags: 0x181080
Fri Jun  7 06:00:06 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:00:59 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was deleted from ae0.0 with flags: 0x181080
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:02:41 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was deleted from ae0.0 with flags: 0x181080
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:07:24 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f

{master:0}[edit]
root@switch# commit                                          
configuration check succeeds
commit complete

{master:0}[edit]
root@switch# run show ethernet-switching mac-learning-log    
Jun 07 06:35:03
Tue May 21 03:20:33 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was learned on xe-0/0/0.0 with flags: 0x2001f
Tue May 21 03:20:37 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Mon May 27 07:21:00 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x1000
Mon May 27 07:21:00 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was deleted from xe-0/0/0.0 with flags: 0x1000
Mon May 27 07:24:31 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was learned on xe-0/0/0.0 with flags: 0x2001f
Mon May 27 07:24:31 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Wed Jun  5 07:16:01 2024 vlan_name test+10 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Wed Jun  5 07:16:01 2024 vlan_name test+10 mac 68:22:8e:64:f5:40 was deleted with flags: 0x181080
Fri Jun  7 06:00:06 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:00:59 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was deleted from ae0.0 with flags: 0x181080
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Fri Jun  7 06:02:36 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:02:41 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was deleted from ae0.0 with flags: 0x181080
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080
Fri Jun  7 06:07:19 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was learned on xe-0/0/1.0 with flags: 0x2001f
Fri Jun  7 06:07:24 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was learned on ae0.0 with flags: 0x2001f
Fri Jun  7 06:35:01 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:64:f5:40 was deleted from ae0.0 with flags: 0x181080
Fri Jun  7 06:35:01 2024 vlan_name VLAN121_SPEgn_SPEpL3_1234+121 mac 68:22:8e:65:21:00 was deleted from xe-0/0/1.0 with flags: 0x181080

 

Solution

  • The issue is seen when interface-mac-limit and mac-table-size are enabled.
  • During the commit,from the kernel outputs,we see some mac learn gencfg messages
  • When sw-learning is enabled/disabled and after that when vlan description knob is changed
  • Generic configuration messages (gencfg msgs) for enabling or disabling sw-learning are retransmitted to the kernel, which triggers a MAC flush.
  • Sending the gencfg again when sw-learning enabled/disabled is not correct,
  • To address this issue, a patch file is provided to prevent the retransmission of gencfg messages when enabling or disabling sw-learning
  • Internal PR raised: https://gnats.juniper.net/web/default/1816049#description_tab

Resolved-In

  • evo:21.4R3-S9-EVO evo:22.2R3-S5-EVO evo:22.4R3-S4-EVO evo:23.2R2-S2-EVO evo:23.4R2-S2-EVO evo:24.2R1-S1-EVO evo:24.2R2-EVO evo:24.3R1-EVO evo:24.4R1-EVO junos:23.2R2-S2 junos:24.2R1-S1 junos:24.2R2 junos:24.3R1



Modification History

2024-08-13 : Article Created

Related Information

Internal PR https://gnats.juniper.net/web/default/1816049#summary_tab