I'm using two ASA 5540 in active-standby failover configuration. These boxes (primary and secondary) are used to establish some L2L and RA (Remote Access) VPN. The active box run OSPF process.
The problem is when failover occurs (just shuting down the active box, or running 'failover active' in a secondary box) all L2L don't be reestablished in a secondary box. The unique way that I can do this (reestablish the connection) is removing the RRI (Reverse Route Injection) configuration (eg. 'no crypto map rprbbe_map 3 set reverse-route') and putting the rri configuration ( 'crypto map rprbbe_map 3 set reverse-route'). After do this the connection is reestablished.
In RA clients the session persists, on a failover event, but the client lost the access. To solve this, the client need to disconnect and reconnect.
A lot of guess work but there are two potential bugs:
I would check if you see rerr or txerr increasing in "show failover" on any of the firewall.
I'd check "fsck" on both units and try doing "write standby" on active afterwards.
Short of that ... either upgrad to 8.3.2 when it's out (august?) or downgrade to 8.2 and test again. Or open a TAC case.
Behavior looks buggy.
What version are you running?