<SOLVED> Config not syncing

[Michael_Kremer]

Hi,

we have a pair of fortiadc's running in two datacenters. The hbdev and datadev are NOT directly connected, but go over a series of switches. Ports 1 and 2 are for the regular traffic, port 3 is the datadev and port 4 is the hbdev. Failover works, but the conf is never synced. When we debug the transfer of the confsync from master to slave with diag debug application hasyncd 31 we see 'no free node'

We are unable to ping the peer with his APIPA Address. Other ADC's on other sites, which are directly interconnected can ping each other over the APIPA-Adress. The interconnecting switches are C3850. Tried changing hbtype to bcast which did not help.

The C3850's are connected to Nexus 5k. We fiddled a bit with the Ethertypes of the three Heartbeat-Params and believe the are the Key-Problem. We changed the 'well-known', with Nexus-Switches colliding, Ethertype but this alone does not help.

What do we have to configure on the cisco side to make the magic happen. Also, can we change the APIPA adresses of the interfaces? Or what Ethertypes should we use on the ADC's

Regards

Michael

[Michael_Kremer]

Hi,

solved, no ip igmp snooping on the Catalyst switches. On the N5k's this had to be done on the vlans that carry hb and data as well. Setting the ha-eth-type to 8895 was important too, because it clashes with some n5k comm.

APIPA-Adresses are not being changed, no idea on how to do that. Have to live with some IP's we don't control in our DC's :(

Regards

Michael