Re: 5.1.6-alpha NDB: Could not get apply status share

[jim shnider]

I am having a very similar problem with 5.1.7-beta.

Has any progress been made on this bug?

Allow me describe my situation. Maybe someone can tell me what I am doing wrong:

<description>

192.168.0.212 - CentOS4 x86_64 (Xeon) running ndbd-5.0.19 and mysql-5.1.7 (both using RPM built for the mysql download mirrors - version discrepency resulted from a partial upgrade trying to make Cluster work, but I don't believe this is the problem because the ndbd connects to ndb_mgmd)

192.168.0.20 - FC4 x86_64 (Athlon64) running ndb_mgmd-5.1.7 (RPM from mysql mirrors)

after starting ndb_mgmd and ndbd, the ndb_mgm client shows everything as expected

when I start mysqld, it takes a while (~15 sec) to start, eventually coming up with an error reported in its log file:

<quote>
060328 00:31:01 mysqld started
060328 0:31:11 [Note] Starting MySQL Cluster Binlog Thread
060328 0:31:11 InnoDB: Started; log sequence number 0 43655
060328 0:31:11 [Note] /usr/sbin/mysqld-max: ready for connections.
Version: '5.1.7-beta' socket: '/var/lib/mysql/mysql.sock' port: 3306 MySQL Co
mmunity Edition - Max (GPL)
060328 0:31:11 [ERROR] NDB: Could not get apply status share
</quote>

and the mysqld instance does not show up in the ndb_mgm administration panel

at this point, I can connect with mysql-client and browse existing data; but if I try to create tables, mysql reports an error:

<quote>
mysql> use blah
Database changed
mysql> create table testing ( first int );
ERROR 1050 (42S01): Table 'testing' already exists
</quote>

it might be important to note that mysqld and ndb_mgmd are interacting at least a little; I know this because ndb_mgmd says so in its log after starting mysqld:

<quote>
2006-03-28 00:06:49 [MgmSrvr] INFO -- Mgmt server state: nodeid 5 reserved for ip 192.168.0.212, m_reserved_nodes 0000000000000022.
2006-03-28 00:06:49 [MgmSrvr] INFO -- Node 5: mysqld --server-id=0 2006-03-28 00:09:27 [MgmSrvr] INFO -- Node 1: Node 2 Connected
2006-03-28 00:09:27 [MgmSrvr] INFO -- Node 2: Node shutdown completed. Initiated by signal 15.
2006-03-28 00:10:18 [MgmSrvr] INFO -- Mgmt server state: nodeid 5 freed, m_reserved_nodes 0000000000000002.
2006-03-28 00:13:15 [MgmSrvr] INFO -- Mgmt server state: nodeid 5 reserved for ip 192.168.0.212, m_reserved_nodes 0000000000000022.
2006-03-28 00:13:15 [MgmSrvr] INFO -- Node 5: mysqld --server-id=0
</quote>

and if I terminate ndb_mgmd, mysqld writes to its log:

<quote>
Management server closed connection early. It is probably being shut down (or ha
s crashed). We will retry the connection.
</quote>

which occurs as the next line after the '060328 0:31:11 [ERROR] NDB: Could not get apply status share' line quoted above - with no date or timestamp (seems odd)

it might be relevant to include the config files for ndb_mgmd, and ndbd together with mysqld:

data/api node (/etc/my.cnf):
<quote>
[mysqld]
default-storage-engine=ndbcluster
ndbcluster
ndb-connectstring=192.168.0.20

[mysql_cluster]
ndb-connectstring=192.168.0.20
</quote>

management node (/var/lib/mysql-cluster/config.ini):
<quote>
[NDBD DEFAULT]
NoOfReplicas=3
[NDB_MGMD]
HostName=192.168.0.20

[NDBD]
HostName=192.168.0.212
DataDir=/var/lib/mysql-cluster
[NDBD]
HostName=192.168.0.214
DataDir=/var/lib/mysql-cluster
[NDBD]
HostName=192.168.0.216
DataDir=/var/lib/mysql-cluster

[MYSQLD]
[MYSQLD]
[MYSQLD]
</quote>

</description>

any thoughts or suggestions?