net_ctl: got bad state MASTER from peer

[askthelist]

anyone know why this situation would come up in sasyncd and/or help me
pinpoint the root cause? seems to correspond with a hiccup in traffic
flowing through the vpn. it corrects itself after a few minutes but it has
occurred on multiple occasions about the same time of day. Non-VPN traffic
still flows normally. My C skills are non-existant, but I've managed to
track it down to this bit of code below inside
src/usr.sbin/sasyncd/net_ctl.c where it seems that I have two masters? I
have four machines, two routers connected to multiple peers, and two
firewalls in active-standby configuration. The two routers run BGP and all
four boxes are running OSPF between one another. This setup has worked out
pretty well but this little situation keeps popping up. Can anyone help me
track down the root cause of this?

net_ctl_check_state(struct syncpeer *p, enum RUNSTATE nstate)
{
    if (nstate < INIT || nstate > FAIL) {
        log_msg(0, "net_ctl: got bad state %d from peer \"%s\"",
            nstate, p->name);
        net_ctl_send_error(p, CTL_STATE);
        return -1;
    }
 *   if (cfgstate.runstate == MASTER && nstate == MASTER) {
        log_msg(0, "net_ctl: got bad state MASTER from peer \"%s\"",
            p->name);
        net_ctl_send_error(p, CTL_STATE);
        return -1;
    }*
    if (p->runstate != nstate) {
        p->runstate = nstate;
        log_msg(1, "net_ctl: peer \"%s\" state change to %s", p->name,
            carp_state_name(nstate));
    }
    return 0;
}