Open Stack

2013/5/22 Gary Kotton <gkotton at redhat.com>:
> I have not looked at the specific case. If the backport is backward
> compatible then why should it make a difference. That is, the default value
> should reflect the current state.
> In some cases a configuration variable is required for certain bug fixes.

Agreed, adding config parameter on a stable branch would be allowed if
default stays the same and required for the bugfix, but in this case
it is not required.

> What am I missing?

This:
>> The flag is totally unreleated and actually adds a new "feature",
>> which is the possibility for admins to forbid multi-attach. However,
>> the flag has a sane default and doesn't change the default behavior
>> but it isn't required to fix the bug.

Cheers,
Alan