When there is a connection flap between the Junos space and the device, the Device down alarm will appear. However, it will not disappear automatically, even when the device is reachable in Network Director 7.1R3

This is a bug in ND 7.1R3.

The issue will not be reproduced when the interface is deactivated in the Device CLI. Because the connection still shows as established in Space even after interface deactivation. Hence, we needed to kill the sshd process in the device CLI to reproduce the issue in your lab.

Steps:

show system processes extensive | match sshd
21225 labroot      20    0    18M    14M select   1   0:12   0.00% sshd
20589 labroot      20    0    17M    13M select   1   0:12   0.00% sshd
20587 root         52    0    16M    12M select   0   0:01   0.00% sshd
21223 root         52    0    16M    12M select   1   0:01   0.00% sshd
 
start shell
kill -9 21225
kill -9 20589

When we kill the sshd deamon, it will flap the connection.

The fix will be part of ND 7.1R3HPv1, and the ETA is tentatively in November.

As a workaround, we can acknowledge the alarm manually to clear it