problema mysql-cluster-gpl 6-2-15-0

[lan...@gmail.com]

Salve a tutti,
scrivo per presentare il mio problema. Ho installato un mysql cluster
formato da due nodi più un nodo di management. Faccio partire il
ndb_mgmd sul node = 1 (managament) e l'ndbd sugli altri due. Entrano
in comunicazione, va tutto bene, i mysql si vedono tra di loro, credo
il database su uno che viene replicato sull'altro nodo, ecc. ecc.

Il problema è che dopo un po' (a volte una ventina di minuti, a volte
quaranta) mi si presenta il seguente problema:

Questo sul nodo di management:

2008-08-20 08:40:54 [MgmSrvr] ALERT -- Node 1: Node 2 Disconnected
2008-08-20 08:40:54 [MgmSrvr] ALERT -- Node 3: Node 2 Disconnected
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: Communication to
Node 2 closed
2008-08-20 08:40:54 [MgmSrvr] ALERT -- Node 3: Network partitioning
- arbitration required
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: President restarts
arbitration thread [state=7]
2008-08-20 08:40:54 [MgmSrvr] ALERT -- Node 3: Arbitration won -
positive reply from node 1
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: GCP Take over
started
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: DICT: lock bs: 0
ops: 0 poll: 0 cnt: 0 queue:
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: LCP Take over
started
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: ParticipatingDIH =
0000000000000000
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: ParticipatingLQH =
0000000000000000
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_Counter_DIH = [SignalCounter: m_count=0
0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_Counter_LQH = [SignalCounter: m_count=0
0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: m_LAST_LCP_FRAG_ORD
= [SignalCounter: m_count=0 0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_From_Master_Received = 1
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: GCP Take over
completed
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: kk: 18013/19 0 0
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: LCP Take over
completed (state = 4)
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: ParticipatingDIH =
0000000000000000
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: ParticipatingLQH =
0000000000000000
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_Counter_DIH = [SignalCounter: m_count=0
0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_Counter_LQH = [SignalCounter: m_count=0
0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3: m_LAST_LCP_FRAG_ORD
= [SignalCounter: m_count=0 0000000000000000]
2008-08-20 08:40:54 [MgmSrvr] INFO -- Node 3:
m_LCP_COMPLETE_REP_From_Master_Received = 1
2008-08-20 08:40:54 [MgmSrvr] ALERT -- Node 2: Forced node shutdown
completed. Caused by error 2303: 'System error, node killed during
node restart by other node(Internal error, programming error or
missing error message, please report a bug). Temporary error, restart
node'.
2008-08-20 08:40:55 [MgmSrvr] INFO -- Node 3: Started arbitrator
node 1 [ticket=4a6a0002ded83085]
2008-08-20 08:40:58 [MgmSrvr] INFO -- Node 3: Communication to
Node 2 opened

Questo sugli altri due nodi:

Message: System error, node killed during node restart by other node
(Internal error, programming error or missing error message, please
report a bug)
Error: 2303
Error data: Node 2 killed this node because GCP stop was detected
Error object: NDBCNTR (Line: 235) 0x0000000a
Program: ndbd
Pid: 5112
Trace: /var/lib/mysql-cluster/ndb_2_trace.log.8
Version: mysql-5.1.23 ndb-6.2.15
***EOM***

Message: System error, node killed during node restart by other node
(Internal error, programming error or missing error message, please
report a bug)
Error: 2303
Error data: Node 3 killed this node because GCP stop was detected
Error object: NDBCNTR (Line: 235) 0x0000000a
Program: ndbd
Pid: 28628
Trace: /var/lib/mysql-cluster/ndb_3_trace.log.14
Version: mysql-5.1.23 ndb-6.2.15
***EOM***

Qualche idea?

Ho inoltre un problema nell'installare su un redhat 5 il pacchetto php-
mysql. Avendo installato il pacchetto MySQL-Cluster-gpl-server, allo
yum install php-mysql mi dà l'errore:

Error: MySQL-Cluster-gpl-server conflicts with mysql

qualche consiglio?

Grazie