Mike Mason wrote:
> I would strongly advise that you do NOT "reinstall the server and re-add
> it to the cluster with the same name as before, it will pick up the
> former ID."
>
> I ran into some very serious problems that occurred (either directly or
> indirectly) when we did this where shared pools got wiped out. Problem
> is that the SBD contains the Node name of the server that died and if
> you try to add it again you will have a dup Node name on the SBD.

[snip]
> As I said before, IMHO there is a problem with NCS 1.7 because it never
> removes servers from the SBD that have been removed from the cluster.
> The only way to remove a formerly active NCS server from the cluster is
> to recreate the SBD.

[snip]
>>>> Andrew C Taubman<ataubmanILoveSpam@novell.AndSpamLovesMe.co m>

> 8/3/2004 8:08:43 PM >>>
> Just delete the node object.
>
> If it's to be permanent then clean up the non-clustering stuff as normal.
>
> Otherwise reinstall the server and re-add it to the cluster with the
> same name as before, it will pick up the former ID.


Can anyone comment further on this issue on NW65SP5? I have just removed
a crashed node from a two-node cluster and reinstalled it successfully,
but have not re-created the SBD, as mentioned above. Before I re-add it
to the cluster again, I want to be sure there won't have any surprises.

Thanks,

Craig

--
Craig Holyoak
IT Support Officer @ Redlands College
cholyoak@redlands.qld.edu.au
http://www.redlands.qld.edu.au/