Hi, I got VPP crashed when trying to add more nsh map. My case is this:
client -> router1 (vpp, 192.168.1.34) -> node1 (vpp, 192.168.1.37) -> ... -> node2 (vpp, 192.168.1.40) -> ... node1/2 work as nodes of a nfv cluster, ,and router1 has two vxlan-gpe tunnels with node1/2: vpp# show vxlan-gpe [0] lcl 192.168.1.34 rmt 192.168.1.37 vni 100 fib-idx 0 sw-if-idx 4 decap-next-protocol nsh [1] lcl 192.168.1.34 rmt 192.168.1.40 vni 101 fib-idx 0 sw-if-idx 11 decap-next-protocol nsh vpp# And router1 works as SFF to redirect client's icmp request to node2, trace like: Packet 6 00:04:13:626134: af-packet-input af_packet: hw_if_index 1 next-index 4 tpacket2_hdr: status 0x81 len 172 snaplen 172 mac 66 net 80 sec 0x5da8216b nsec 0x2708ea01 vlan 0 vlan_tpid 0 00:04:13:626138: ethernet-input IP4: 02:fe:c0:a8:01:28 -> 02:fe:c0:a8:01:22 00:04:13:626141: ip4-input UDP: 192.168.1.40 -> 192.168.1.34 tos 0x00, ttl 253, length 158, checksum 0x39b4 fragment id 0x0000 UDP: 4790 -> 4790 length 138, checksum 0x0000 00:04:13:626144: ip4-lookup fib 0 dpo-idx 5 flow hash: 0x00000000 UDP: 192.168.1.40 -> 192.168.1.34 tos 0x00, ttl 253, length 158, checksum 0x39b4 fragment id 0x0000 UDP: 4790 -> 4790 length 138, checksum 0x0000 00:04:13:626146: ip4-local UDP: 192.168.1.40 -> 192.168.1.34 tos 0x00, ttl 253, length 158, checksum 0x39b4 fragment id 0x0000 UDP: 4790 -> 4790 length 138, checksum 0x0000 00:04:13:626148: ip4-udp-lookup UDP: src-port 4790 dst-port 4790 00:04:13:626149: vxlan4-gpe-input VXLAN-GPE: tunnel 1 next 4 error 0 00:04:13:626151: nsh-input nsh ver 0 ttl 3 len 6 (24 bytes) md_type 1 next_protocol 3 service path 14 service index 254 c1 1 c2 2 c3 3 c4 4 00:04:13:626156: error-drop rx:vxlan_gpe_tunnel1 00:04:13:626157: drop nsh-input: no mapping for nsh key vpp# Existing nsh rules are: vpp# show nsh map nsh entry nsp: 16 nsi: 255 maps to nsp: 16 nsi: 255 nsh_action push encapped by VXLAN GPE intf: 4 nsh entry nsp: 16 nsi: 254 maps to nsp: 16 nsi: 254 nsh_action pop encap-none nsh entry nsp: 15 nsi: 255 maps to nsp: 15 nsi: 255 nsh_action push encapped by VXLAN GPE intf: 4 nsh entry nsp: 15 nsi: 254 maps to nsp: 15 nsi: 254 nsh_action pop encap-none nsh entry nsp: 14 nsi: 255 maps to nsp: 14 nsi: 255 nsh_action push encapped by VXLAN GPE intf: 11 vpp# Since the packet trace complain that no expected nsp/nsi found, I add the missing nsp/nsi, using following cmd: vpp# create nsh map nsp 14 nsi 254 mapped-nsp 14 mapped-nsi 254 nsh_action pop encap-none 0 0 And once the client sent icmp packet again, the VPP would crashed immediately, and restarted, with nothing left. No other clues have been found currently, I'm guessing according to the packet trace, #1 this packet is expected to be sent to `vxlan_gpe_tunnel1`, but `vxlan4-gpe-input` shows that next idx is 4, which refer to `vxlan_gpe_tunnel0`, do I misuse vxlan-gpe tunnel here or miss some other rules? #2 if this CRASH is a known issue, what can I do to workaround this? If anybody needs any more information, I'd be glad to provide.
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#14192): https://lists.fd.io/g/vpp-dev/message/14192 Mute This Topic: https://lists.fd.io/mt/34696238/21656 Mute #vpp: https://lists.fd.io/mk?hashtag=vpp&subid=1480452 Mute #nsh: https://lists.fd.io/mk?hashtag=nsh&subid=1480452 Mute #vxlan-gpe: https://lists.fd.io/mk?hashtag=vxlan-gpe&subid=1480452 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
