Bug #48669 ndb_mgmd needs --no-wait option

• Submitted: 10 Nov 2009 13:25
• Modified: 5 Dec 2009 18:45
• Reporter: Geert Vanderkelen
• Status: Closed
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S4 (Feature request)
• Version: mysql-5.1-telco-7.0
• OS: Any
• Assigned to: Magnus Blåudd
• CPU Architecture: Any

[10 Nov 2009 13:25] Geert Vanderkelen

Description:
MySQL Cluster 7.0, when you have 2 management nodes configured, starting the first one, it will need the other up and running as well. This is because it wants to be make sure the configuration is identical.
However, there are a few situations where it makes sense only to start 1 management node and let the other ndb_mgmd join later.
So, starting a MySQL Cluster 7.0 (and up) configured with 2 management nodes, you should be able to start it with only 1 ndb_mgmd. (like in earlier versions)

How to repeat:
(feature request)

Suggested fix:
* Add an option (similar to) --no-wait
* Check whether the configuration is in-sync when the other ndb_mgmd joins the cluster.

[13 Nov 2009 11:24] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/90333

[13 Nov 2009 13:00] Magnus Blåudd

Have extended the test to see what happens when two nodes are started with different config using --nowait-nodes and the will bot stay up and  ignore each other.
That's because the config check  protocol only compares generation and state, will add checksum as well so that the config check will shutdown one of the nodes, letting the on with lowest nodeid win.

[17 Nov 2009 18:14] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/90736

[17 Nov 2009 18:23] Bugs System

Pushed into 5.1.39-ndb-7.0.10 (revid:[e-mail]-20091117181657-7bhcfhlx8cvprub0) (version source revid:[e-mail]-20091117181657-7bhcfhlx8cvprub0) (merge vers: 5.1.39-ndb-7.0.10) (pib:13)

[17 Nov 2009 18:23] Bugs System

Pushed into 5.1.39-ndb-7.1.0 (revid:[e-mail]-20091117181724-9ua9h7g0lid2x929) (version source revid:[e-mail]-20091117181724-9ua9h7g0lid2x929) (merge vers: 5.1.39-ndb-7.1.0) (pib:13)

[17 Nov 2009 23:23] Magnus Blåudd

Pushed to 7.0 and 7.1

[5 Dec 2009 18:45] Jon Stephens

Documented in the NDB-7.0.10 changelog as follows:

        Added the ndb_mgmd --nowait-nodes option, which
        allows a cluster that is configured to use multiple management
        servers to be started using fewer than the number configured.
        This is most likely to be useful when a cluster is configured
        with two management servers and you wish to start the cluster
        using only one of them.

        For more information, see "ndb_mgmd — The MySQL Cluster 
        Management Server Daemon".

Also updated ndb_mgmd program description, added new option to ndb_mgmd options table, added reference in Cluster Roadmap section.

Closed.