With puppet of course.
Or worse yet. Some just write down the password in a place that's easy to find.
Is that so bad a good password that's written down is far better against a network based attack than a poor password that's remembered?
I often tell users to write their password on a postit and put it in their wallet imo that's safer than stored badly encrypted (think password protected excell spreadsheets) on a system thats on and network connected
I don't know's on third, What's the name of the guy on second.
Dr Mario is not a real doctor under no cirucmstances should he be permitted to examine your special places.
It depends what you mean by fixup.
If it used to work.. but was broken when sendmail was re-installed that can be fixed within an hr or so.
but if it's changing the function of a controled production system that's a major change that requires a CIP to be filled out. two appearances in front of CAB. and 30 days lead time
I spend FAR more of my time doing paperwork than I do doing actual techwork.
Right.. but at some point in asking for some non trivial ammount of resourcing to fix something you're going to need to identify and explain what change needs to be done.. what the risk that is being mitigated is. and what the probability of that risk eventuating is.
IT suffers pretty badly from building Taj Mahals and misconstruing technical risk as business risk.
"We need to fix this and it's going to cost us X if we don't" has to be based on established trust and faith that you've evaluated X correctly. you get there by proving it with data and reports (things IT also sucks at supprisingly hard)
The biggest problem that IT is that people think "This is going to cost us X if we don't fix it."where in reality what's needed for decision making is "We should fix Y because there's a Z chance of costing X if we don't"
RAND isn't as reasonable or as non-discriminatory as most people seem to think
http://en.wikipedia.org/wiki/Reasonable_and_Non_Discriminatory_Licensing
In particular RAND allows for per unit licencing costs for implementations which sounds the death knell for any free software implementation of a standard that is patent encumbered even if the patents are availible under a RAND scheme.
The issue is that as free software is by definition redistributable you need to licence the patent for every potential user.
Intel CPUs are not defective, they just act that way. -- Henry Spencer