Management VRF not responding

[jlmickens]

I have a Catalyst 9200 with the Mgmt-vrf configured.  Over the weekend, it stopped responding to our monitoring platform.  I can no longer ssh into the switch either.  It is at a remote site so I do not have direct access to it.  Yesterday I managed to get someone on site to connect to the console port with a laptop.  Everything appears normal and seems to be working, however the management vrf is still not responding.  No traffic is coming out of it.  I can do a packet capture on the link connected to the management port, and can see the traffic going in, but nothing coming out.  Has anyone seen this before?  No configuration changes were made.

Is there a way to restart the management vrf without rebooting the switch? 

[MHM Cisco World]

I will check the mgmt VLAN is it UP or not 

try ping to SW 

MHM

[jlmickens]

The port is configured with an IP directly - routed, not access.  I can't ping either.  Responses never come back from the switch.

[MHM Cisco World]

try use traceroute see where the traffic end 

sometime one engineer add ACL or filter prefix in path 

MHM

[Preston Chilcote · Accepted solution]

There's no such thing as "restarting a VRF".  Are there any other IPs on the box that you can try to ping as a test?

What's the IOS-xe version? Whats the uptime?  Any suspicious syslogs?  If you don't want to spend a lot of time troubleshooting, this is probably a good excuse to upgrade to the latest recommended release, as a reload will likely fix the issue. 

[jlmickens]

I'm still waiting on someone on site to get another laptop hooked up to the console again, but the other switch that it's paired with (not a VC) is running IOS XE Lite version 17.06.04, so I suspect that they are the same.  Unfortunately that is the only IP configured on the switch, as its generally only used for L1/L2 connectivity.  Devices in the same subnet are not able to get to it.  Uptime is over a year, close to a year and a half, judging by the other switch that I can get to.

[jlmickens]

For those following along, a reload of the switch recovered the management access.