MySQL Forums
Forum List  »  NDB clusters

ndnd no wait option implications
Posted by: Richard Cruise
Date: November 13, 2020 10:07AM

Hi all,

I'm currently working on an unattended system running 3 servers which each host a NDB management, data & MySQL nodes. Cluster config file is below

[ndbd default]
# Options affecting ndbd processes on all data nodes:
NoOfReplicas=3 # Number of replicas
DataMemory=8192M

[ndb_mgmd]
# Management process options:
HostName=pcamn01 # Hostname or IP address of MGM node
DataDir=/var/lib/mysql-cluster # Directory for MGM node log files
NodeId=1
[ndb_mgmd]
# Management process options:
HostName=pcamn02 # Hostname or IP address of MGM node
DataDir=/var/lib/mysql-cluster # Directory for MGM node log files
NodeId=2
[ndb_mgmd]
# Management process options:
HostName=pcamn03 # Hostname or IP address of MGM node
DataDir=/var/lib/mysql-cluster # Directory for MGM node log files
NodeId=3


[ndbd]
# Options for data node "A":
# # (one [ndbd] section per data node)
HostName=pcamn01 # Hostname or IP address
NodeId=17 # Node ID for this data node
DataDir=/usr/local/mysql/data # Directory for this data node's data files
ServerPort=11001

[ndbd]
# Options for data node "B":
HostName=pcamn02 # Hostname or IP address
NodeId=18 # Node ID for this data node
DataDir=/usr/local/mysql/data # Directory for this data node's data files
ServerPort=11002

[ndbd]
# Options for data node "C":
HostName=pcamn03 # Hostname or IP address
NodeId=19 # Node ID for this data node
DataDir=/usr/local/mysql/data # Directory for this data node's data files
ServerPort=11003

[mysqld]
# SQL node options:
HostName=pcamn01 # Hostname or IP address
NodeId=33 # Node ID for this api node

[mysqld]
# SQL node options:
HostName=pcamn02 # Hostname or IP address
NodeId=34 # Node ID for this api node

[mysqld]
# SQL node options:
HostName=pcamn03 # Hostname or IP address
NodeId=35 # Node ID for this api node


We had an issue recently where one of the clusters didn't start because a data node (Node ID 18) was partitioned during startup.

The other 2 data nodes were then stuck in the starting phase

Since the system is unattended, it would be better if the cluster would start without waiting for other nodes.

My understanding is that the --nowait-nodes parameter can be used for this, in the event a node is unavailable then the other nodes can be told to ignore it and start anyway.

My question is, what are the implications if I set all 3 data nodes to not wait for the others on startup using this option? Is there a risk to data integrity?

Options: ReplyQuote


Subject
Views
Written By
Posted
ndnd no wait option implications
564
November 13, 2020 10:07AM
389
November 16, 2020 03:27AM


Sorry, you can't reply to this topic. It has been closed.

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.