Two processes are writing to fifo and messages are lost.

Hello,

Sometimes when heartbeat starts I get an error: NV failure 
(msgfromsteam). It looks like mach_down script and heartbeat are writing 
to the fifo at the same time and messages are lost. Because of this 
failover doesn't happen.
Is there any solution to this?
Below I attached log from ha-debug.

Thank you in advance.

Best Regards,
Arek

heartbeat[31298]: 2012/06/19_11:12:18 WARN: Logging daemon is disabled 
--enabling logging daemon is recommended
heartbeat[31298]: 2012/06/19_11:12:18 info: **************************
heartbeat[31298]: 2012/06/19_11:12:18 info: Configuration validated. 
Starting heartbeat 2.1.4
heartbeat[31299]: 2012/06/19_11:12:18 info: heartbeat: version 2.1.4
heartbeat[31299]: 2012/06/19_11:12:18 info: Heartbeat generation: 1340024896
heartbeat[31299]: 2012/06/19_11:12:18 info: seed is 1421633575
heartbeat[31299]: 2012/06/19_11:12:18 info: Creating FIFO 
/var/lib/heartbeat/fifo.
heartbeat[31299]: 2012/06/19_11:12:18 info: glib: ping group heartbeat 
started.
heartbeat[31299]: 2012/06/19_11:12:18 info: glib: ucast: write socket 
priority set to IPTOS_LOWDELAY on eth0
heartbeat[31299]: 2012/06/19_11:12:18 info: glib: ucast: bound send 
socket to device: eth0