Comments on: Deploying MySQL Cluster over multiple hosts

By: neel

neel — Wed, 28 Dec 2011 08:24:41 +0000

very very good article…each n everything is mentioned in a clear way…helpful for the beginers….

thanks…
Neel

By: admin

admin — Fri, 17 Jun 2011 15:04:25 +0000

IKetut,

to get more information, if you were using the start /B approach then try without it (you’ll need more windows to cope with the number of processes). After that check the log files that are created – there should be some clues there.

Regards, Andrew.

By: IKetut gunawan

IKetut gunawan — Fri, 17 Jun 2011 08:18:49 +0000

Hi Andrew ,
I have done to try .
Fist time clustering runing well , but after restart my window and runing again the second server mgm error “failed to report event to event log, error: 1502

can you give sugestion .

Thanks you .
IKetutG

By: aktheus

aktheus — Thu, 31 Mar 2011 20:27:20 +0000

I’m a newbie, jeje, but i found your post very clear, thanks a lot. I’m going to use it for my very first distributed architecture proyect (I’m a student). Hope everything goes well.

By: admin

admin — Thu, 17 Feb 2011 15:19:58 +0000

Thomas, the first thing I would do is to replace 127.0.0.1 with the real IP address (192.168.180.??). 127.0.0.1 only makes sense on that machine whereas the configuration data needs to be shared and interpreted on all hosts in the Cluster.

Please let me know how you get on.

Regards, Andrew.

By: Thomas

Thomas — Wed, 16 Feb 2011 12:32:41 +0000

I’ve set up my MySQL-Cluster as you expected (adding some additional variables), it started fine, but:

root@zm2:/etc/mysql# ndb_mgm
– NDB Cluster — Management Client –
ndb_mgm> show
Connected to Management Server at: localhost:1186
Cluster Configuration
———————
[ndbd(NDB)] 2 node(s)
id=1 @192.168.180.44 (mysql-5.1.39 ndb-7.0.9, Nodegroup: 0, Master)
id=2 @127.0.0.1 (mysql-5.1.39 ndb-7.0.9, Nodegroup: 0)

[ndb_mgmd(MGM)] 2 node(s)
id=101 @127.0.0.1 (mysql-5.1.39 ndb-7.0.9)
id=102 @127.0.0.1 (mysql-5.1.39 ndb-7.0.9)

[mysqld(API)] 2 node(s)
id=51 (not connected, accepting connect from 192.168.180.44)
id=52 (not connected, accepting connect from 192.168.180.45)

ndb_mgm>

mysqld is running. And I have configured:
[mysqld]
ndb-nodeid = 51
server-id = 51
ndb-cluster
ndb-connectstring = 192.168.180.44

[mysql_cluster]
ndb-connectstring = 192.168.180.44

Since there is no error message and I couldn’t find anything — why doesn’t connect mysqld to the management daemon?

A little bit lost …

By: admin

admin — Thu, 03 Feb 2011 10:14:55 +0000

somecallmemike,

The reason why the management node(s) shouldn’t be co-located with the data nodes is that by default one of the management nodes acts as an arbitrator in the case that one set of data nodes become isolated from the rest. I work through several use cases of how arbitration works and the impact of where you run the management node in this post.

Regards, Andrew.

By: somecallmemike

somecallmemike — Thu, 03 Feb 2011 05:20:37 +0000

Great post got it working right away. I was hoping you could explain why the cluster warns that running the management server on the same host as the data node is not advised? I did some googling and couldn’t find a definitive answer.

Thanks!

By: admin

admin — Wed, 24 Nov 2010 18:28:41 +0000

Hi Tamer,

Great to hear that you’re up and running.

Did you set the DataMemory parameter in your config.ini? If not then the default is 80Mbytes and you may have hit that: http://dev.mysql.com/doc/refman/5.1/en/mysql-cluster-ndbd-definition.html#ndbparam-ndbd-datamemory

To check, run “ALL REPORT MEMORY” from within the ndb_mgm tool.

You may need to increase IndexMemory too.

Once you’ve edited your config.ini file, shut down the ndb_mgmd process and then start it again with the –initial option. Then restart each of your data nodes in turn (wait for the 1st data node to come back up before restarting the next one to avoid an outage) – do *not* use the –initial option when starting the data nodes unless you want to remove all of your data.

Andrew.

By: Tamer

Tamer — Wed, 24 Nov 2010 17:21:55 +0000

I found this very helpfull. I was able to get the cluster setup and working but at 18,080 rows in one of my tables. I get a table full error.

I am using 3 linux boxes running Ubuntu 9. and each box has 4gb of ram. Any help resolving this is greatly appreciated.

Thanks