I have OTV running and have implemented EEM largely as per the Cisco design guide:
event manager applet WatchBFDdown
description "Monitors BFD status, if it goes down, bring OVERLAY int down"
event syslog pattern "BFDFSM-6-BFD_SESS_DOWN" period 1
action 1.0 cli command "enable"
action 2.0 cli command "config t"
action 2.1 syslog msg "EEM: WatchBFDdown will shut int Overlay10"
action 3.0 cli command "interface Overlay10"
action 4.0 cli command "shutdown"
action 5.0 syslog msg "EEM WatchBFDdown COMPLETE ..."
event manager applet WatchBFDup
description "Monitors BFD status, if it goes up, bring OVERLAY int up"
event syslog pattern "BFDFSM-6-BFD_SESS_UP" period 1
action 1.0 cli command "enable"
action 2.0 cli command "config t"
action 2.1 syslog msg "EEM: WatchBFDdown will no shut int Overlay10"
action 3.0 cli command "interface Overlay10"
action 4.0 cli command "no shutdown"
action 5.0 syslog msg "EEM WatchBFDup COMPLETE ..."
However the result is this:
Jun 8 00:10:25.379: %BFDFSM-6-BFD_SESS_DOWN: BFD-SYSLOG: BFD session ld:4107 handle:1,is going Down Reason: ECHO FAILURE
Jun 8 00:10:25.407: %BFDFSM-6-BFD_SESS_UP: BFD-SYSLOG: BFD session ld:4107 handle:1 is going UP
Jun 8 00:10:25.428: %HA_EM-6-LOG: WatchBFDup: EEM: WatchBFDdown will no shut int Overlay10
Jun 8 00:10:25.502: %HA_EM-6-LOG: WatchBFDdown: EEM: WatchBFDdown will shut int Overlay10
The result is the Overlay is shut down and stays shut down. It looks like the syslog polling process does not enforce a strict FIFO.
What are my options to fix this? I'm thinking of configuring a trigger delay on the "no shut" applet.
Increasing the BFD interval could improve things but won't necessary completely avoid this scenario.
Thanks.
Peter
