Hi folks.
I'm speccing a system at my NT-zealot work. I have no chance of getting
this system built in the deadline we have on anything but NT--we just
don't have the coding expertise in-house. However, I wish to use it as
additional ammunition in my war against braindead technology decisions.
So what I'd like to do is specify a reasonable level of availability for
what the system is to be: a reasonably heavily utilised, database-driven
web server.
Now how does one arrive at figures like what IBM like to quote like 99.9%?
Does that exclude planned maintenance? If so, does planned maintenance
include software upgrades such kernel upgrades? How about application
upgrades (reboot necessary for NT, not generally for Unix)? Or is it just
harware upgrades?
So what is a reasonable level of availability, without going to Local
Director-based load balancing or something overboard like that? This
isn't for a bank or the military--so I need to be reasonable about it.
My reason for this is because I just get dumbfounded when these souped up,
enormously fault-tolerant, super-fast, super overpowered web servers fall
over with fscking crap software failures. We won't talk about the file,
print, proxy or directory servers... I'll start foaming at the mouth :)
---
Rev Simon Rumble Politics -- the gentle art of getting votes from
[EMAIL PROTECTED] the poor and campaign funds from the rich by
http://www.rumble.net by promising to protect each from the other.
-- Oscar Ameringer
--
SLUG - Sydney Linux Users Group Mailing List - http://www.slug.org.au
To unsubscribe send email to [EMAIL PROTECTED] with
unsubscribe in the text
