Because my organization has yet to implement Linux virtualization on the
mainframe, I tend to use this list as a way to stay abreast of
developments with little to contribute. However, as a networking
professional designing global networks I must state that your advise,
Mr. Boyes, to the list is absolutely BEST PRACTICE. While Mr. Cox's
comments are techincally accurate, for a list composed of primarily
operating system professionals whose focus is never exclusively
networking his suggestions are close to being malicious. Such
techniques, while possible, are application availability denial of
service attacks waiting to happen, even for networking professionals
such as Mr. Cox who might know how or when to use them. Obviously, a
lesson you and I have both learned.
Harold Grovesteen
David Boyes wrote:
O> I'd argue that you will have more problems trying to make this work
reliably than just doing it the way I described. If you've got time
to
debug this and all the paths have equivalent permissions and
usability
characteristics, then yes, it's technically possible. You just have
to
have a lot of free time to figure out what happened when it doesn't
work, and personally, I've got better things to do.
The paths can be completely different, IP does not care.
Certainly. But you're going to spend a lot more time figuring out why it
doesn't work when it doesn't.
I can make it work -- been there, done that. I just have learned from
bitter experience with similar setups that it's a rotten idea if you can
avoid it, and it's usually not a good idea to design based on it. It's
more trouble than it's worth for 99% of cases.
If what you're after is that it can be made to work, then you are
correct, it can work. Is it a good idea? I'd say no, and it definitely
isn't a best practice. Too many opportunities for things and people to
be confused, especially when there are ways easier to implement and
maintain.
Sounds like a new added bug as you describe it.
Not from a diagnostic standpoint.
Well if a stack prohibits valid configurations I'd call it words like
buggy or flawed, or if I'm trying to set up such a configuration
"unfit
for purpose".
We'll have to agree to disagree, then. While there are reasons for
wanting to do this, there are alternative methods to accomplish the same
effect that don't require this "feature" and that (at least IMHO) are
simpler to understand and maintain. If I need to do this with one of the
IBM stacks, I can always front-end the IBM stack with a Linux guest and
not have to have further discussion about it.
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
