Re: Disk Data files and restarting data nodes using --initial

On Thu, 2007-08-16 at 13:12 -0500, Al wrote:
> Hi,
>
> I'd like to know what the recommended practice is for adding new nodes
> to the cluster on a disk-based system.
>
> From how I've understood things so far, any scale out we do on the
> data node side will require that we back up our data as we would
> expect to lose all our data when starting up ndb on each node using
> --initial. And as from what I gather in a separate,
> currently-ongoing thread, --initial is required whenever a hardware
> change occurs.

no. initial only *required* on some specific configuration changes.

> A few moments ago, we did just that. We ran ndbd with the --initial
> flag on all data nodes and saw that disk data files are not removed
> when restarting a node using --initial. Fine, but then what is our
> recourse now since the two older nodes have these disk data files and
> the two new ones don't? Does the cluster recognize these data files,
> and if so, can it handle the fact that these files only exist in two
> of the storage nodes and not the other nodes?

you should manually remove these files. This is an irritation/bug at the moment.

-- 
Stewart Smith, Senior Software Engineer
MySQL AB, www.mysql.com
Office: +14082136540 Ext: 6616
VoIP: 6616@sip.us.mysql.com
Mobile: +61 4 3 8844 332

Jumpstart your cluster:
http://www.mysql.com/consulting/packaged/cluster.html


application/pgp-signature attachment: This is a digitally signed message part