Gateway_less_Fwd: ECMP is not working when IP Fabric is configured as provider network
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R4.1 |
New
|
High
|
Ashok Singh | |||
Trunk |
Fix Committed
|
High
|
Ashok Singh |
Bug Description
ECMP is not working when IP Fabric is configured as provider network
R4.1.0.0 Build 23 Ubuntu 14.04 Mitaka
Steps
———
Create a VN (say left_vn: 10.10.10.0/24)and configure provider network as IP Fabric network
Launch 3 VMs across compute nodes in left_vn
Now, configure AAP (say 10.10.10.100/32) on 2 VMIs with active-active mode
Now, AAP route shows as ECMP properly on left_vn.
But, whenthe AAP route leaked to default routing instance due to IP-Fabric is configured as provider network, it is not ECMP.
Please see the log below:
root@nodek11:~# vif --list
Vrouter Interface Table
Flags: P=Policy, X=Cross Connect, S=Service Chain, Mr=Receive Mirror
Mt=Transmit Mirror, Tc=Transmit Checksum Offload, L3=Layer 3, L2=Layer 2
D=DHCP, Vp=Vhost Physical, Pr=Promiscuous, Vnt=Native Vlan Tagged
Mnp=No MAC Proxy, Dpdk=DPDK PMD Interface, Rfl=Receive Filtering Offload, Mon=Interface is Monitored
Uuf=Unknown Unicast Flood, Vof=VLAN insert/strip offload, Df=Drop New Flows, L=MAC Learning Enabled
Proxy=MAC Requests Proxied Always, Er=Etree Root
vif0/0 OS: em1 (Speed 1000, Duplex 1)
Vrf:0 Flags:L3L2VpEr QOS:-1 Ref:6
RX packets:27157 bytes:2857641 errors:2
TX packets:6929 bytes:4887312 errors:0
Drops:2
vif0/1 OS: vhost0
Vrf:0 Flags:PL3DEr QOS:-1 Ref:7
RX packets:6489 bytes:4744582 errors:0
TX packets:27325 bytes:2906745 errors:0
Drops:1
vif0/2 OS: pkt0
RX packets:1975 bytes:241450 errors:0
TX packets:12668 bytes:1364863 errors:0
Drops:0
vif0/3 OS: tap7fe679ef-96
Vrf:3 Flags:PL3L2DEr QOS:-1 Ref:5
RX packets:1163 bytes:106778 errors:0
TX packets:697 bytes:65390 errors:0
vif0/4 OS: tap0346accc-11
Vrf:0 Flags:PL3L2DProxyEr QOS:-1 Ref:5
RX packets:1161 bytes:105964 errors:0
TX packets:1191 bytes:94846 errors:0
vif0/5 OS: tapb0cdc481-1a
Vrf:0 Flags:PL3L2DProxyEr QOS:-1 Ref:5
RX packets:2901 bytes:355584 errors:0
TX packets:2465 bytes:207514 errors:0
vif0/4350 OS: pkt3
RX packets:0 bytes:0 errors:0
TX packets:0 bytes:0 errors:0
Drops:0
vif0/4351 OS: pkt1
RX packets:981 bytes:82404 errors:0
TX packets:981 bytes:82404 errors:0
Drops:0
root@nodek11:~# rt --dump 2 | grep 10.10.10.100 <=== ECMP on AAP on left_vn VRF
10.10.10.100/32 32 P - 33 -
root@nodek11:~# nh --get 33
Id:33 Type:Composite Fmly: AF_INET Rid:0 Ref_cnt:2 Vrf:2
Valid Hash Key Parameters: Proto,SrcIP,
Sub NH(label): 31(21) 22(26)
Id:31 Type:Tunnel Fmly: AF_INET Rid:0 Ref_cnt:11 Vrf:0
Oif:0 Len:14 Data:00 25 90 93 d2 44 0c c4 7a 32 0a 88 08 00
Id:22 Type:Encap Fmly: AF_INET Rid:0 Ref_cnt:4 Vrf:0
Encap Data: 02 03 46 ac cc 11 00 00 5e 00 01 00 08 00
root@nodek11:~# rt --dump 0 | grep 10.10.10.100 <=== IP-Fabric VN is configured as provider network, no ECMP here
10.10.10.100/32 32 PT - 21 -
root@nodek11:~# nh --get 21
Id:21 Type:Encap Fmly: AF_INET Rid:0 Ref_cnt:6 Vrf:0
Encap Data: 02 03 46 ac cc 11 00 00 5e 00 01 00 08 00
root@nodek11:~#
Changed in juniperopenstack: | |
assignee: | Hari Prasad Killi (haripk) → Ashok Singh (ashoksr) |
information type: | Proprietary → Public |
@haripk @ashoksr
Does this sound fixable or should we disallow active-active AAP
on VMIs in VNs that use ip-fabric as provider?