Re: Impossible to recover from datanode down

[Jan-willem van Eys]

We're running version MySQL-Max-5.0.24-0.glibc23 (64bit version) installed from rpm.

Our config.ini:

[NDBD DEFAULT]
NoOfReplicas=2
DataMemory=3500M
IndexMemory=1024M
MaxNoOfOrderedIndexes=800
MaxNoOfAttributes=8000
MaxNoOfTables=600

#MaxNoOfUniqueHashIndexes=512
#MaxNoOfTriggers=2000

NoOfFragmentLogFiles=50

[TCP DEFAULT]
portnumber=2202

[NDB_MGMD]
hostname=192.168.100.4
datadir=/var/lib/mysql-cluster

[NDBD]
hostname=192.168.100.2
datadir=/mysql/mysql-cluster1

[NDBD]
hostname=192.168.100.3
datadir=/mysql/mysql-cluster1

[MYSQLD]
hostname=192.168.100.3
[MYSQLD]
hostname=192.168.100.1
[MYSQLD]
[MYSQLD]
[MYSQLD]

We killed -9 the ndbd process, and tried to restart it by entering 'ndbd'

There weren't any messages in the logs AFAIK. (We 'recycled' the test machines, and I didn't save any logs).

It feels as if the reconnecting node can't catch up with the transactions it's getting...