No access to management IP after add/del mgmt VRF

[wangxin]

Description

Simply run below commands on SONiC switch, then access to management IP failed, including ping and SSH.

sudo config vrf add mgmt
sudo config vrf del mgmt

This issue is not observed on all platforms. So far, I have tested 3 platforms, two of them have this issue:

• Platform: x86_64-arista_7260cx3_64, HwSKU: Arista-7260CX3-D108C8, ASIC: broadcom
• Platform: x86_64-cel_seastone-r0, HwSKU: Celestica-DX010-C32, ASIC: broadcom

Below platform is OK:

• Platform: x86_64-mlnx_msn2700-r0, HwSKU: Mellanox-SN2700, ASIC: mellanox

Steps to reproduce the issue:

1. Run command sudo config vrf add mgmt on switch.
2. Run command sudo config vrf del mgmt on switch.
3. Ping management IP from another host. Ping failed.
4. Try to establish SSH connection to management IP of the switch. SSH connection failed.
5. Reboot the switch from console.
6. After reboot, the switch is back to normal.

Describe the results you received:

• After command sudo config vrf del mgmt is executed, the current SSH connection is stuck.
• Try to ping management IP of the switch from anther host, ping failed.
• Try to establish a new SSH connection to management IP of the switch, SSH connection failed.

Describe the results you expected: After the mgmt VRF add/del operation, we should be able to access the management IP of the switch as usual.

Additional information you deem important (e.g. issue happens only occasionally):

**Output of `show version`:**

$ show version

SONiC Software Version: SONiC.20191130.47
Distribution: Debian 9.13
Kernel: 4.9.0-11-2-amd64
Build commit: 272885b17
Build date: Thu Aug 27 22:29:12 UTC 2020
Built by: sonicbld@jenkins-slave-phx-2

Platform: x86_64-arista_7260cx3_64
HwSKU: Arista-7260CX3-D108C8
ASIC: broadcom
Serial Number: SSJ17444852
Uptime: 08:02:40 up 0 min,  1 user,  load average: 2.49, 0.69, 0.24

Docker images:
REPOSITORY                 TAG                 IMAGE ID            SIZE
docker-snmp-sv2            20191130.47         ac73d27334db        348MB
docker-snmp-sv2            latest              ac73d27334db        348MB
docker-fpm-frr             20191130.47         a30e044767ea        335MB
docker-fpm-frr             latest              a30e044767ea        335MB
docker-acms                20191130.47         18a3e7d99d25        182MB
docker-acms                latest              18a3e7d99d25        182MB
docker-lldp-sv2            20191130.47         e2b73929aae0        312MB
docker-lldp-sv2            latest              e2b73929aae0        312MB
docker-orchagent           20191130.47         923ba3c4daab        333MB
docker-orchagent           latest              923ba3c4daab        333MB
docker-teamd               20191130.47         6537f4cb9bf5        315MB
docker-teamd               latest              6537f4cb9bf5        315MB
docker-syncd-brcm          20191130.47         58907962b1e6        437MB
docker-syncd-brcm          latest              58907962b1e6        437MB
docker-platform-monitor    20191130.47         8bc71c7776a3        357MB
docker-platform-monitor    latest              8bc71c7776a3        357MB
docker-sonic-telemetry     20191130.47         660da51474e9        353MB
docker-sonic-telemetry     latest              660da51474e9        353MB
docker-database            20191130.47         a6998344b386        289MB
docker-database            latest              a6998344b386        289MB
docker-dhcp-relay          20191130.47         74e646e284e6        299MB
docker-dhcp-relay          latest              74e646e284e6        299MB
docker-router-advertiser   20191130.47         0f3dac3a2c1e        289MB
docker-router-advertiser   latest              0f3dac3a2c1e        289MB
k8s.gcr.io/pause           3.2                 80d28bedfe5d        683kB

**Attach debug file `sudo generate_dump`:**

sonic_dump_str-7260cx3-acs-1_20200828_075247.tar.gz

[lguohan]

@Anandaraj-Maharajan , can dell look at this issue? it is related to mgmt vrf.