If you have more than one version of the mysqld on your system it might
be better for now not to use a global my.cnf file, /etc/my.cnf.
Following the instructions in the tutorial on the MySQL WEBsite, I
copied the contents of my-medium.cnf into /usr/local/mysql/var/my.cnf,
changed the port number [just in case...], and fired up
/usr/local/mysql/bin/mysqld_safe. Seems to work.
-Joseph
===========================================================
On Mon, 2004-07-26 at 10:02, Crouch, Luke H. wrote:
> I do have an Ndb.cfg file in the /usr/local/mysql/var directory. I've been using a
> little different syntax in the Ndb.cfg files, and it's been working fine so far...
>
> OwnProcessId 9
> 10.0.199.150 10000
>
> would it make a difference if I'm running mysqld_safe or mysqld? what about the
> ndbcluster option? can it just stay in the my.cnf, or does it have to be on the mysqld
> command?
>
> thanks for response...I'll try out that different syntax for Ndb.cfg...
>
> -L
>
> > -----Original Message-----
> > From: Joseph E. Sacco, Ph.D. [mailto:jsacco@stripped]
> > Sent: Monday, July 26, 2004 9:00 AM
> > To: Crouch, Luke H.
> > Cc: cluster@stripped
> > Subject: Re: mysqld API overview?
> >
> >
> > Luke,
> >
> > (0) The mysqld needs it's own Ndb.cfg.
> >
> > For example:
> >
> > % cat /usr/local/mysql/var/Ndb.cfg
> > nodeid=9
> > host=10.0.199.150:10000
> >
> > The Ndb.cfg file contains information needed by the mysqld to join the
> > cluster managed by the manager node on 10.0.199.150
> >
> > (1) The mysqld must be running in order for the management node to see
> > it.
> >
> > -Joseph
> >
> > =============================================================
> > On Mon, 2004-07-26 at 09:37, Crouch, Luke H. wrote:
> > > is there a guide explaining how to set up mysqld processes
> > as API to the cluster? I'm running a 4-machine cluster
> > (1,2,3, and 4 are the creative names, eh?). in config.ini,
> > I've specified:
> > >
> > > machine 1: 1 MGM, 1 DB, 1 API
> > > machine 2: 1 DB, 1 API
> > > machine 3: 1 DB, 1 API
> > > machine 4: 1 DB, 1 API
> > >
> > > on machine 1....
> > > /NDB/mgm/config.ini contents:
> > > <snip>
> > > [API DEFAULT]
> > > ArbitratorRank: 1
> > >
> > > [API]
> > > Id: 6
> > > ExecuteOnComputer: 1
> > >
> > > [API]
> > > Id: 7
> > > ExecuteOnComputer: 2
> > >
> > > [API]
> > > Id: 8
> > > ExecuteOnComputer: 3
> > >
> > > [API]
> > > Id:9
> > > ExecuteOnComputer: 4
> > >
> > > /etc/my.cnf contents:
> > > [mysqld]
> > > datadir=/usr/local/mysql/var
> > > ndbcluster
> > > socket=/tmp/mysql.sock
> > >
> > > [client]
> > > socket=/tmp/mysql.sock
> > >
> > > /usr/local/mysql/var/Ndb.cfg contents:
> > > OwnProcessId 6
> > > 10.0.199.150 10000 #mgmt server IP
> > >
> > > but when I run an 'ALL STATUS' in the management, I don't
> > see any API nodes listed at all. is it like the DB nodes,
> > where you have to start them all up to get one of them to
> > start? or have I put my Ndb.cfg in the wrong place? do I have
> > to explicitly use the --ndbcluster option on mysqld, or can I
> > use it in the my.cnf like I am? the API setup part of the
> > cluster admin guide is very shaky...a couple sentences long...
> > >
> > > can anyone help me out?
> > > thanks,
> > > -L
> > --
> > Joseph E. Sacco, Ph.D. <jsacco[at]earthlink[dot]net>
> >
> >
--
Joseph E. Sacco, Ph.D. <jsacco[at]earthlink[dot]net>
