ok, do you REALLY think that if enough people in America got together to form a revolt, and they actually had enough people to win........do you really think that they wouldn't be able to also win an election?
Let's see...
Few thousands of people with a small number of tanks, military airplanes and nukes, can turn Washington, DC and New York, NY into piles of rubble, if that is the course of action they choose over all others. It would turn US into something completely different, as current government and financial companies would no longer exist, so if that would happen to be their goal, they would be successful.
Same few thousands of people could've sell tanks, planes and nuke parts, choose some acceptable-looking figureheads out of themselves, and probably buy some election ads on TV. No matter what their goals are, they would have zero impact on government or its policies, unless they coincide with goals of people already in power, or close to power.
I am not saying that anyone is in such position now, or will be any soon, or it is likely that such group of people would accomplish anything positive, but here is an example where violence works and elections don't.
Look, life is full of people at all skill levels. I deal with C++ programmers who have never cracked a case, and have only a vague familiarity with the relationship between voltage, current, and power and have never touched assembly. Guess what? They can still program in C++ and C# competently, and we all make money.
And I have seen various other people who "made money" without any knowledge of what was supposed to be their job. They are called frauds, and what they did was called fraud.
You would have some resemblance of a point if highly skilled and knowledgeable people were extremely rare. They are not. They just drown in the sea of scammers and idiots such as yourself. Die in a fire and don't forget to kill all your friends.
There is no "radical shift". Multiple services were running on Unix servers since the beginning of Unix development. The idea of "single service per server" is entirely from Windows system administration, and combined with VMs it can be implemented -- but it's an extremely stupid idea.
Unix design is the opposite to this -- clusters of servers may be involved in implementing a single service, and design of such service will take into account a particular model of distributing the load, keeping shared state, locking, etc. However at the same time one server may handle multiple related or unrelated services. In Windows world such a design is unthinkable -- so VMs are used as crutches to allow huge multicore servers run multiple single-threades piece of shit application, each in its special padded room with its own copy of the OS, its own emulated hardware up to the keyboard controller and SB16 sound card to emit beeps that no one is going to hear. The last thing we need is this model spreading to sanely designed systems.
This only happens when server is being administered by a Windows admin. And if it's Linux running on top of VM running on Windows, it is not a Linux server at all, it's an idiotic contraption that has absolutely no excuse to exist.
Windows admins are dumb. Their job is dumb because Windows provides no tools or interfaces to perform its administration in intelligent way. Mark Russinovich is probably the closest thing to a smart Windows admin/developer that exists, but he is dumb, too (and he works for Microsoft).
Problems happen when Windows admins are trying to work on non-Windows systems -- they bring dumb Windows way of administering a server, and they are still nothing but Windows admins. They can be easily recognized by their use of VMware products to run Linux in production environment.
Study how? If the problem did not happen within 10 seconds of boot-up, it had to be somehow triggered, likely by things that only happen in production environment.
It's also very unlikely that VMware jockey keeps an exact copy of production network and a copy of all production data, that are necessary to run a production server's snapshot anywhere outside of the production environment in the first place.
All of these places use a tactic very similar to mine, and any other competent sysadmin. They restore a working copy of it so that users are uninterrupted and then perform forensics on the system that went down so they can solve the problem.
No. They keep THE REST OF SERVERS running, and apply the fix to all of them once the problem is found. One server being offline means nothing for them.
1. Unless the problem is trivial, it only shows itself under certain conditions, that are unlikely to be imitated outside of the production environment. 2. Whatever the problem is, it's usually related to data -- your current production data.
And then your system (maintained by hacks and idiots) corrupts the data on that SAN. And does that every time you restore data from backup (minus all transaction that were completed after last backup) and re-image the system. Now what?
Unix sysadmins are supposed to be smart not because they are maintaining a more complex system. They are supposed to be smart because they are should use all tools available to them, to keep system running reliably and efficiently, and fix the problems that appear or are discovered while running it. "Restoring" a broken system from the state when its brokenness is discovered to the state before the brokenness is discovered, is for VMware jockeys, not admins.
Titles or work and characters are usually trademarked, and adaptations are derived works, so both trademarks and copyright still require movie authors to get permission.
Unix file descriptor exhibits polymorphism. And its implementations are usually based on inheritance. That's 1970's.
At the time when BSD sockets were made (clearly polymorphic and based on the above mentioned file descriptors), they were criticized as not sufficiently generalized interface when it comes to implementation, so "more pure OO" design of SysV was proposed -- and then rejected as pointless windbaggery, and fortunately remains so. In the next decade Microsoft, in its usual infinite wisdom, did not see a value in implementing file, device, network socket and various forms of IPC descriptors as alternative forms of the same base object but re-used BSD sockets for networking, thus giving birth to the fundamental split between Windows' "everything is special" design and "generalize as much as possible" Unix design, that we still can clearly see now.
So the "question" you are talking about, is being "discussed" in this manner for at least 40 years already!
Slashdot Mirror
Translation: It's OK for us to conquest a country if it was ruled by people we don't like.
ok, do you REALLY think that if enough people in America got together to form a revolt, and they actually had enough people to win........do you really think that they wouldn't be able to also win an election?
Let's see...
Few thousands of people with a small number of tanks, military airplanes and nukes, can turn Washington, DC and New York, NY into piles of rubble, if that is the course of action they choose over all others. It would turn US into something completely different, as current government and financial companies would no longer exist, so if that would happen to be their goal, they would be successful.
Same few thousands of people could've sell tanks, planes and nuke parts, choose some acceptable-looking figureheads out of themselves, and probably buy some election ads on TV. No matter what their goals are, they would have zero impact on government or its policies, unless they coincide with goals of people already in power, or close to power.
I am not saying that anyone is in such position now, or will be any soon, or it is likely that such group of people would accomplish anything positive, but here is an example where violence works and elections don't.
Why "getting information out of the country" is necessary? To call the foreign enemies so they can take over and install their puppets?
Look, life is full of people at all skill levels. I deal with C++ programmers who have never cracked a case, and have only a vague familiarity with the relationship between voltage, current, and power and have never touched assembly. Guess what? They can still program in C++ and C# competently, and we all make money.
And I have seen various other people who "made money" without any knowledge of what was supposed to be their job. They are called frauds, and what they did was called fraud.
You would have some resemblance of a point if highly skilled and knowledgeable people were extremely rare. They are not. They just drown in the sea of scammers and idiots such as yourself. Die in a fire and don't forget to kill all your friends.
Stock can be used:
1. To buy other companies.
2. To "reward" executives without paying real money.
Both things are usually detrimental to the overall conditions on the market.
And not a single fuck will be given that day by anyone who is not a disgusting parasite.
I do software development and system administration in a professional capacity for 20 years already.
You, on the other hand, "know computers".
I did, and I do. As opposed to you, I know what I am doing.
By using package manager and rsync, you idiot!
lol
As a former Windows SMS Administrator and Desktop Engineer
...you should better shut up when people are discussing real servers.
There is no "radical shift". Multiple services were running on Unix servers since the beginning of Unix development. The idea of "single service per server" is entirely from Windows system administration, and combined with VMs it can be implemented -- but it's an extremely stupid idea.
Unix design is the opposite to this -- clusters of servers may be involved in implementing a single service, and design of such service will take into account a particular model of distributing the load, keeping shared state, locking, etc. However at the same time one server may handle multiple related or unrelated services. In Windows world such a design is unthinkable -- so VMs are used as crutches to allow huge multicore servers run multiple single-threades piece of shit application, each in its special padded room with its own copy of the OS, its own emulated hardware up to the keyboard controller and SB16 sound card to emit beeps that no one is going to hear. The last thing we need is this model spreading to sanely designed systems.
This only happens when server is being administered by a Windows admin. And if it's Linux running on top of VM running on Windows, it is not a Linux server at all, it's an idiotic contraption that has absolutely no excuse to exist.
Then you read bullshit.
Windows admins are dumb. Their job is dumb because Windows provides no tools or interfaces to perform its administration in intelligent way. Mark Russinovich is probably the closest thing to a smart Windows admin/developer that exists, but he is dumb, too (and he works for Microsoft).
Problems happen when Windows admins are trying to work on non-Windows systems -- they bring dumb Windows way of administering a server, and they are still nothing but Windows admins. They can be easily recognized by their use of VMware products to run Linux in production environment.
Study how? If the problem did not happen within 10 seconds of boot-up, it had to be somehow triggered, likely by things that only happen in production environment.
It's also very unlikely that VMware jockey keeps an exact copy of production network and a copy of all production data, that are necessary to run a production server's snapshot anywhere outside of the production environment in the first place.
Congratulations, you are a incompetent!
Virtualization is good for two things:
1. Running Windows.
2. Software development.
Anyone who is running VMware or its ilk as a replacement for version control system, backups or package manager, is incompetent.
If you have failures of that kind, you are doing something terribly, terribly wrong. Servers do not go down by themselves.
All of these places use a tactic very similar to mine, and any other competent sysadmin. They restore a working copy of it so that users are uninterrupted and then perform forensics on the system that went down so they can solve the problem.
No. They keep THE REST OF SERVERS running, and apply the fix to all of them once the problem is found. One server being offline means nothing for them.
Okay, so let's assume you've got a big cluster of servers for some random task, and one of them breaks.
You find the problem and fix it on all servers -- because if one is broken and all are identical, then all of them are broken.
It NEVER works.
1. Unless the problem is trivial, it only shows itself under certain conditions, that are unlikely to be imitated outside of the production environment.
2. Whatever the problem is, it's usually related to data -- your current production data.
lol wut
And then your system (maintained by hacks and idiots) corrupts the data on that SAN. And does that every time you restore data from backup (minus all transaction that were completed after last backup) and re-image the system. Now what?
Unix sysadmins are supposed to be smart not because they are maintaining a more complex system. They are supposed to be smart because they are should use all tools available to them, to keep system running reliably and efficiently, and fix the problems that appear or are discovered while running it. "Restoring" a broken system from the state when its brokenness is discovered to the state before the brokenness is discovered, is for VMware jockeys, not admins.
Titles or work and characters are usually trademarked, and adaptations are derived works, so both trademarks and copyright still require movie authors to get permission.
Unix file descriptor exhibits polymorphism. And its implementations are usually based on inheritance. That's 1970's.
At the time when BSD sockets were made (clearly polymorphic and based on the above mentioned file descriptors), they were criticized as not sufficiently generalized interface when it comes to implementation, so "more pure OO" design of SysV was proposed -- and then rejected as pointless windbaggery, and fortunately remains so. In the next decade Microsoft, in its usual infinite wisdom, did not see a value in implementing file, device, network socket and various forms of IPC descriptors as alternative forms of the same base object but re-used BSD sockets for networking, thus giving birth to the fundamental split between Windows' "everything is special" design and "generalize as much as possible" Unix design, that we still can clearly see now.
So the "question" you are talking about, is being "discussed" in this manner for at least 40 years already!