After assigning IP address to router port via ifconfig ping to host from the same subnet is not working

[oleksandrivantsiv]

Platform: Mellanox SN2700
Description:
After assigning IP address to router port via ifconfig ping to host from the same subnet is not working. ARP packets are handled, neighbor is added on both sides. ICMP packets on device are sent but not trapped. Possible root cause - trap route for router port is not created (sx_api_router_uc_routes_dump_all.py doesn't show this route). When IP address is set in config and deployed by Ansible ping works as expected.

[oleksandrivantsiv]

Steps to reproduce are the following:

• Configure interfaces IP address in /etc/network/interfaces file:

vim /etc/network/interfaces
...
#
auto Ethernet96
iface Ethernet96 inet static
    address 20.20.20.25
    netmask 255.255.255.0
...

• Restart device to apply configuration.
• Configure IP address from the same subnet on host connected to switch.
• Ping host from device and vice-versa:

root@arc-switch1025:/home/admin# ifconfig Ethernet96
Ethernet96 Link encap:Ethernet  HWaddr 00:02:03:04:05:00
          inet addr:20.20.20.25  Bcast:20.20.20.255  Mask:255.255.255.0
          inet6 addr: fe80::202:3ff:fe04:500/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:9216  Metric:1
          RX packets:2 errors:0 dropped:0 overruns:0 frame:0
          TX packets:9 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:136 (136.0 B)  TX bytes:742 (742.0 B)

root@arc-switch1025:/home/admin# ping 20.20.20.100
PING 20.20.20.100 (20.20.20.100) 56(84) bytes of data.
64 bytes from 20.20.20.100: icmp_seq=1 ttl=64 time=0.394 ms
64 bytes from 20.20.20.100: icmp_seq=2 ttl=64 time=0.359 ms
64 bytes from 20.20.20.100: icmp_seq=3 ttl=64 time=0.332 ms
64 bytes from 20.20.20.100: icmp_seq=4 ttl=64 time=0.278 ms
64 bytes from 20.20.20.100: icmp_seq=5 ttl=64 time=0.280 ms
^C
--- 20.20.20.100 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4000ms
rtt min/avg/max/mdev = 0.278/0.328/0.394/0.049 ms
root@arc-switch1025:/home/admin#

1. Change IP address if interface via ifconfig command:
   root@arc-switch1025:/home/admin# ifconfig Ethernet96 21.21.21.25/24

The following error is observed in logs

000001 1030000000001 1050000000001 1070000000001 1090000000001 10b0000000001 10d00000
Oct 05 07:33:52 arc-switch1025 docker[937]: Oct 05 07:33:52 NOTICE  SAI_RIF: mlnx_sai_rif.c[398]- mlnx_remove_router_interface: Remove rif rif 25
Oct 05 07:33:52 arc-switch1025 docker[937]: Oct 05 07:33:52 ERROR   ROUTER: hwi/rif/rif_be.c[499]- sdk_router_interface_set: Failed to delete rif, this rif(25) is being used, err (Resource is in use)
Oct 05 07:33:52 arc-switch1025 docker[937]: Oct 05 07:33:52 ERROR   SAI_RIF: mlnx_sai_rif.c[415]- mlnx_remove_router_interface: Failed to delete router interface - Resource is in use.
Oct 05 07:33:52 arc-switch1025 docker[729]: terminate called without an active exception
Oct 05 07:33:52 arc-switch1025 docker[937]: terminate called without an active exception
Oct 05 07:33:52 arc-switch1025 kernel: Core dump to |/usr/bin/coredump-compress orchagent 107 pipe failed
Oct 05 07:33:52 arc-switch1025 kernel: Core dump to |/usr/bin/coredump-compress syncd 40 pipe failed
Oct 05 07:33:52 arc-switch1025 docker[937]: /etc/init.d/syncd: line 39:    40 Aborted                 start-stop-daemon --start --quiet --exec $DAEMON -- $DAEMON_ARGS
▒Oct 05 07:33:53 arc-switch1025 ntpd[663]: Listen normally on 43 Ethernet96 21.21.21.25 UDP 123
Oct 05 07:33:53 arc-switch1025 ntpd[663]: Deleting interface #33 Ethernet96, 20.20.20.25#123, interface stats: received=0, sent=0, dropped=0, active_time=259 secs
Oct 05 07:33:53 arc-switch1025 ntpd[663]: peers refreshed

[oleksandrivantsiv]

Issue was closed by mistake

[lguohan]

Thanks a lot for reporing this issue. It looks like the orchagent is trying to delete a rif while it is been referenced. I think current orchagent does not have the logic to protect a rif been removed when it is still in use. We probably need to add the logic there.

[oleksandrivantsiv]

Issue fixed.