crash node

Hi,
I have installed percona xtradb cluster on CentOS 6.

Percona-XtraDB-Cluster-server-5.5.28-23.7.369.rhel6.x86_64
Percona-XtraDB-Cluster-galera-2.0-1.117.rhel6.x86_64

There are still problems, whith which I need your help:

During the restore DB with LOAD DATA INFILE

nod options:
wsrep_provider=/usr/lib64/libgalera_smm.so
wsrep_sst_receive_address=172.31.6.x:3320
wsrep_node_incoming_address=172.31.6.x
wsrep_slave_threads=2
wsrep_cluster_name=trimethylxanthine
wsrep_provider_options = “gmcast.listen_addr=tcp://172.31.6.x:3330;”
wsrep_sst_method=rsync
wsrep_node_name=node3330
wsrep_max_ws_size=4G

if transaction is bigger than 4G, it will end and these two messages
will be displayed and node always crashed:
ERROR 1213 (40001) at line 1: Deadlock found when trying to get lock;
try restarting transaction
ERROR 2013 (HY000) at line 1: Lost connection to MySQL server during query

Sometimes it happens during the smaller transaction, too:
ERROR 1047 (08S01) at line 1: Unknown command
When I run the same command later again, it will be processed.

Could you give me an advise, what can be wrong, please?

Can you provide your MySQL error log and/or innodb/global status prior to the crash if you can repeat this?