I'm sorry, but how is it okay to bring a production server down completely
and spend the 30 minutes it's gonna take to get a flash archive on a
server, and then the other hour or more it'll take you to regression test
EVERYTHING, since who knows what really happened ... to update a single
package? If you've got thousands of servers to manage and upgrade, you'll
have to simply hire update minions, I know that I sure don't have time for
that.
Of course it's not okay to just go and bring a production server down.
And you're not supposed to be doing regression tests in production. That's
what the integration & testing, and then product testing and acceptance
phases are for.
BTW, that's the "three track" testing model. It's extremely effective and
delivers very high stability, but the downside is that you need at least
twice as many systems. With Solaris zones that number can now be
significantly reduced though.
Finally, think about what happens when your production system goes down.
You've got a problem, possibly a very big problem. But do you really know
what the problem is?
It's not that the system went down.
It's not that it was a production system.
The problem is of architectural nature, or lack thereof. There should have
been no production systems deployed without some means of either HA (High
Availability) or failover, for example, a cluster of two or more nodes. And
the architecture, or lack thereof, did not account nor provide for that.
That's what IT and CS is supposed to be all about: not ad-hoc,
hack-it-together-just-so-it-works-we-don't-have-enough-time, but planning,
architecting, designing, engineering, testing. Making IT an asset not a
liability to the company.
_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
