MySQL Forums
Forum List  »  NDB clusters

Mysql NDB cluster - Big import fail
Posted by: Cesar Hernandez
Date: May 31, 2022 05:40AM

Hi

I'm having some issues trying to import a big (12 GB) mysql dump. The dump is from an internal mysql database (no ndb) and I'm trying to insert it to a new mysql ndb cluster (changed all sentences ENGINE=INNODB to ENGINE=NDBCLUSTER).

I've had different issues, some of the errors trying to import were fixed changing configuration parameters, but some others don't. In fact, almost the configuration parameters have been changed as a consequence of an error fired.
For example, the last one that (randomly) appears is:

ERROR 1297 (HY000) at line 9831: Got temporary error 4010 'Node failure caused abort of transaction' from NDBCLUSTER
Warning (Code 1297): Got temporary error 4010 'Node failure caused abort of transaction' from NDB
Error (Code 1297): Got temporary error 4010 'Node failure caused abort of transaction' from NDBCLUSTER
Error (Code 1180): Got error 4010 - 'Unknown error 4010' during COMMIT

Every time an error is fired seems is not related to out of ram, out of disk, or other operating system related issues.

It seems to my that this ndb cluster is too hard to configure, compared to a simple mysql database (no configuration at all and it simply works).

Is there any standard configuration that may fit for this "big" environments? (and I said "big" because my database dump is 12 GB which is not too big).
Or is there any configuration parameters related to these mysql import problems?

This dump can be inserted without issues on a mysql database (no cluster)


My environment is:

- one management node (16 GB RAM)

- two ndb nodes (32 GB RAM each)

- one mysql server node (16 GB RAM)

All mysql cluster versions software are 8.0.29.

config.ini:


[ndbd default]
# Options affecting ndbd processes on all data nodes:
NoOfReplicas=2

#MaxNoOfConcurrentOperations must be greater than MaxNoOfConcurrentTransactions
MaxNoOfConcurrentOperations = 1629496
MaxNoOfConcurrentTransactions = 1629496
MaxDMLOperationsPerTransaction = 1629400
#MaxNoOfConcurrentOperations = 4294967039
#MaxNoOfConcurrentTransactions = 4294900000
#MaxDMLOperationsPerTransaction = 4294967295

DataMemory = 16000M # Memory allocate for data storage
IndexMemory = 2048M # Memory allocate for index storage

#temporal para import inicial
NoOfFragmentLogFiles = 1000

SharedGlobalMemory = 10G

MaxNoOfOrderedIndexes = 1000
MaxNoOfAttributes = 100000

TransactionDeadlockDetectionTimeout = 10000

#para evitar REDO log files overloaded error
TimeBetweenLocalCheckpoints = 2

FragmentLogFileSize = 32M

[tcp default]
#hay un error de Out of SendBufferMemory in sendSignal
SendBufferMemory=16M
ReceiveBufferMemory=16M



[ndb_mgmd]
# Management process options:
hostname=wocu-ha-mysqlcluster-dev-1 # Hostname of the manager
datadir=/var/lib/mysql-cluster # Directory for the log files


[ndbd]
hostname=wocu-ha-mysqlcluster-dev-2 # Hostname/IP of the first data node
NodeId=2 # Node ID for this data node
datadir=/var/lib/mysql-cluster/data # Remote directory for the data files


[ndbd]
hostname=wocu-ha-mysqlcluster-dev-3 # Hostname/IP of the second data node
NodeId=3 # Node ID for this data node
datadir=/var/lib/mysql-cluster/data # Remote directory for the data files


[mysqld]
# SQL node options:
hostname=wocu-ha-mysqlcluster-dev-4



Thanks

Options: ReplyQuote




Sorry, only registered users may post in this forum.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.