Hi
We have Juniper QFX5100s running randomly dropping MPLS ccc L2 circuits. They are running
17.3R3-S1.5.
Configuration is basically:
set protocols l2circuit neighbor 1.2.3.4 interface xe-0/0/8.0 virtual-circuit-id 418
set protocols l2circuit neighbor 1.2.3.4 interface xe-0/0/8.0 no-control-word
set protocols l2circuit neighbor 1.2.3.4 interface xe-0/0/8.0 mtu 1522
set protocols mpls traffic-engineering mpls-forwarding
set protocols mpls explicit-null
set protocols mpls no-decrement-ttl
set protocols mpls no-cspf
set protocols mpls label-switched-path cisco to 1.2.3.4
set protocols mpls interface all
set protocols mpls interface lo0.0
We see this in the logs at the drop:
Feb 28 14:02:02 qfx fpc0 BRCM_COS_HALP(brcm_cos_classifier_exp_set_switch:1591): brcm_cos_classifier_exp_set_switch: ifl index is zero
...
Feb 28 14:02:02 qfx fpc0 LBCM-L2,pfe_brcm_add_sd_tag_mpls_tpids(),2133:Added TPID 0x8100 on mpls_port 0x18003ff1
All the l2circuits are dropped at this point and then reform about 30 seconds later. From the MPLS neighbour, which is a Cisco, we see this intermittently:
%LDP-5-NBRCHG: LDP Neighbor 1.2.3.3:0 (1) is DOWN (Session KeepAlive Timer expired)There are no layer-2 issues. We see connectivity never drop and no drop packets, interface resets, etc.
Anyone else seen this? This one of numerous MPLS issues we have seen with the QFX platform.
Thanks