So on the console server, it supports the concepts of groups. Currently
I've done it similar to xCAT where if you add a group that does not exist
to a node through something akin to 'nodech', the group is created
implicitly. At the same time, changing a group to add a node that doesn't
exist will create the node (not possible in xCAT today).
Some recent events have made me consider requiring a node exist before it
can be added to a group and a group to exist before it can be added to a
node. I've seen typos bite people in commands enough that I thought
warning them that 'ipmp' is not a group would be better than just making a
new group 'ipmp' and completely missing the inheritence they intended..
The good news is that I don't think any additional tedium needs to happen
to make a new group compared today. Today I'd do 'nodech <noderange>
groups,=newgroup' to make a new group and in this model, a command like
'nodegroupadd newgroup nodes=<noderange>' would be similarly terse and to
the point.
Thoughts?
------------------------------------------------------------------------------
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user
