Just to check, is it the connection between the mysql console and the mysqld that's being
lost or between the mysqld and the Data Nodes? It sounds as though it's the first case as
you say you can still connect when using SSH. Do you have a firewall on these machines and
if so, have you opened up port 3306 (or if you've changed the default, whatever port
you've selected). What mysql command are you using to connect (e.g. mysql -h 192.168.0.15
-P 3306 -u root).
Have you looked at the Cluster log file for the period when the connection is lost?
> -----Original Message-----
> From: artfire-kyle [mailto:kyleerror@stripped]
> Sent: 18 November 2010 20:43
> To: cluster@stripped
> Subject: MySQL CLuster Data Nodes Lose Connection
> I have a cluster up and running, but I am experiencing some strange
> Cluster is fully up and running. Show in management mode confirms all
> management, data and API nodes are connected.
> When i first bring the cluster up I can properly access the API nodes via
> mysql console remotely. After a what seems random amount of time, the data
> nodes are not accessable from remote servers.
> Restarting mysqld seems to temp. fix it, but then the same thing. The same
> API node is available via SSH without losing connection.
> When the remote connection is lost, the management console still shows them
> all connected, and I can access API node "A" via the mysql console remotely
> VIA API node b and vise versa, but cannot connect from any other server not
> tied to the cluster.
> Two things that seem to fix this is, 1) restart the mysqldaemon, but then it
> repeats after n time, 2) tcpdump (possibly any net traffic) seems to "wake"
> it up again but not always.
> I am really stumped here and I appreciate any help you may offer.
> View this message in context: http://mysql-
> Sent from the MySQL - Cluster mailing list archive at Nabble.com.
> MySQL Cluster Mailing List
> For list archives: http://lists.mysql.com/cluster
> To unsubscribe: