Re: MySQL client not connecting to management server?

Matthew,

You hit the nail right on the head! The problem was that I had defined 2 ndbd nodes and I was trying to initially start my cluster up with 1 node. I imagined that I could start my cluster up with 1 node and when the future nodes are available then the data will automatically be copied over to them when they come up. I guess this is not true.

I learned that when bringing new nodes into the cluster I would have to backup all my data and re-perform an ndbd initial startup which means starting from scratch.

Anyways I wish the logs would have printed why my api clients weren't able to connect but all is well since its figured out now.

Thanks,
- Jake

On 9/28/07, Matthew Montgomery <mmontgomery@mysql.com> wrote:
> On Thu, 2007-09-27 at 07:08 -0700, Jake Conk wrote:
> > I only have 1 data node at the moment and I'm working on building
> > another one but for the one that I do have ndbd was fully connected
> > before I tried to connect the NDB APE client. All I need is ndbcluster
> > and ndb-connectstring=XX in my my.cnf right?
>
> Based on the attached output of 'SHOW', your config.ini has two [NDBD]
> blocks defined which dictates that two datanodes constitutes a full
> cluster. For a "initial system start" all defined NDBD nodes must be
> present. For a "system start", NDBD requires that at least one member
> from each node group is started before the cluster as a whole may
> start.
>
> You haven't stated what your NoOfReplicas value is. This is fairly
> important for understanding how start up is handled. If NoOfReplicas =
> 1 you have one node per node group. This setting makes it such that all
> nodes must be present for a "system start". If NoOfReplicas = 2 then
> the cluster will wait for at least one node from each node group and
> StartPartialTimeout and StartPartitionedTimeout to expire before
> attempting to do a "system start".
>
> Defaults for these are:
> StartPartialTimeout = 30000 milliseconds
> StartPartitionedTimeout = 60000 milliseconds
>
> Regarding NDB API connections, the ndb-connectstring directs the
> application to where the management node is located. The management
> node then notifies the API node when the cluster is started and which
> data nodes are members of the cluster. The API node then connects to
> the data nodes using various randomly assigned ports.
>
>
> > Are there any other methods of debugging since this isn't giving me
> > any errors or any clues to help troubleshoot my problem of none of my
> > ndb clients are able to connect?
> >
>
> Within the Datadir for ndb_mgmd on 192.168.20.1 will be a
> ndb_1_cluster.log file this should show you messages from the starting
> node "id=2 @192.168.20.30" as to why it is delaying start up. Likely it
> is because this node is waiting for its peer.
>
> > Thanks,
> > > >>> [ndbd(NDB)] 2 node(s)
> > > >>> id=2 @192.168.20.30 (Version: 5.0.45, starting, Nodegroup:0)
> > > >>> id=3 (not connected, accepting connect from 192.168.20.40)
>
> --
> Matthew Montgomery, Support Engineer
> San Antonio, TX
> MySQL Inc, www.mysql.com
>
>

-- 
MySQL Cluster Mailing List
For list archives: 
http://lists.mysql.com/cluster
To unsubscribe:    
http://lists.mysql.com/cluster?unsub=lists@pantek.com