Friday, February 18, 2005
« IE news | Main | Fire bad, tree pretty »
Why we REALLY need stateless server clusters
I was just IMing with a friend. He's working with a client that has an interesting IT staff. The person he's working with for instance, recently shut down a main server for maintenance - in the middle of the day, without warning the active users.
 
This is why we need fault-tolerant, stateless server clusters. Not to stop downtime from accidents or hardware failure, but rather to overcome the limitations of IT staffing.
Friday, February 18, 2005 9:11:51 AM (Central Standard Time, UTC-06:00)  #    Disclaimer  |  Comments [4]  | 

Friday, February 18, 2005 9:57:16 AM (Central Standard Time, UTC-06:00)
Good to know I'm not alone. Our IT guy turned on automatic updates for all of our servers a while back and had it automatically install them as well. It took us a bit to figure out why the ERP server would suddenly reboot itself in the middle of the day once in a while.

Oy.
David Totzke
Monday, February 21, 2005 10:15:34 AM (Central Standard Time, UTC-06:00)
Before we go concluding what an idiot the guy was at the client, we need to also consider the possibility that management declined his request for overtime or comp time when he suggested he needed to come in on Sunday to do this off-hours, or something like that.

An example of how the "limitations of IT staffing" are often caused by the limitations of IT management as much as by the competency of the staff pool.
Daniel Billingsley (DansDreams)
Monday, February 21, 2005 2:03:23 PM (Central Standard Time, UTC-06:00)
Or they could have been very junior (though in this case neither is true - the person is just bascially without clue...).

But when I was junior I did some seriously bad stuff. Due to incompetence certainly, but potentially excused due to lack of experience? :)

Perhaps the worst was when I changed the password policy on our VAX to require password changes. This had the side-effect (unknown to me until then) of invalidating all sorts of existing passwords. It took us two days to get all the user and batch processing accounts back to working status. Very ugly, and very preventable - if only I had then the experience I have now...
Rockford Lhotka
Tuesday, February 22, 2005 2:31:32 PM (Central Standard Time, UTC-06:00)
Overtime?!? In IT? It's been my experience that we're on salary for the most part and OT and weekends is just part of the job. Not that I agree with that, mind you, but it does seem to be the norm.

You're not with the "Union" are you brother Daniel? < g >

Dave
David Totzke
Comments are closed.