HA Cluster Not synchronised with master

[one_co_il]

Hi

I'm using 2 FG 100D with OS 5.4.4 as a Cluster, Slave Unit not sync with master

I try to force them sync via CLI using command mention here: https://kb.fortinet.com/kb/documentLink.do?externalID=FD36494

but still system not sync (event after recalculate checksum)

download and compare the Units conf file and something isn't clear.

the same setting are exssist but not in the same order:

for example:

Primary unit:

config system DHCP server

edit 3

XXX

XXX

XXX

next

edit 2

XXX

XXX

XXX

next

--------------------------

Slave unit:

config system DHCP server

edit 2

XXX

XXX

XXX

next

edit 3

XXX

XXX

XXX

next

also running command "get sys ha status" show errors on HA interfaces (I'm not sure if it is OK or NOT).

HBDEV stats:

    FG100DXXX(updated 3 seconds ago):

        ha1: physical/1000full, up, rx-bytes/packets/dropped/errors=273267071/50                                                              3402/0/0, tx=362985753/591158/0/0

        ha2: physical/1000full, up, rx-bytes/packets/dropped/errors=253297452/34                                                              6508/0/0, tx=261769107/347635/0/0

[Toshi_Esumi]

Not ok to see any error at the top of "get sys ha status". Looks like heartbeat interface is having some kind of issues. If you have multiple links, just start with one link.

But order of DHCP servers (2 and 3) could be reversed depending on how they're created, or removed&recreated, while they're connected (they might be copied even when the units are not completely in-sync). That wouldn't cause checksum mismatch between both units.

[one_co_il]

Hi

thank you for reply, the errors i saw on the HA1 and HA2 interface was rx-bytes. I didn't read it properly.

I manage to make the slave unit sync with master. its a workaround but worked for me.

I Break the Cluster.

[ol]

Download config from master edit the Config:

rename Unit name

rename unit SN

change HA section to Stand alone

upload config to salve unit

reconnect to cluster[/ol]

It seem that units are working fine and synced