When will Windows become sophisticated enough to...

[Rubycon]

Not require a reboot after a program/install/change/update? It's annoying that you have to reboot after all these years - and we're still doing it! :|

 

[us3rnotfound]

I am also curious.

 

[Spineshank]

Most of the time i dont reboot and the program works fine.

 

[Markbnj]

I install a fair bit of software, and see this very rarely anymore. And it could be that some of those cases are just poorly written installers, as Gnrslash4life alludes. A reboot might be required is if an existing component needs to be updated, and was already running or had its disk resources locked when the installer ran. The installer is able to create a record telling Windows to perform the file replacement on the next restart. Another case can be an update to registry values that are known to be reloaded only on startup. But again, these are all rare cases now.

 

[nerp]

I see reboots infrequently. Call them rareboots.

I even didn't have to reboot after installing the latest ATI Catalyst drivers for my video card and IGP. How about that?

 

[BlueAcolyte]

I guess its because the changes need to be made to files that are open?

 

[George P Burdell]

Originally posted by: Gnrslash4life
Most of the time i dont reboot and the program works fine.

this

 

[corkyg]

Reboot is required when changes are made in the registry in order to execute them. If no registry changes are required, no reboot is usually the path.

The focus is on the registry. When booted, the existing regitry is loaded into memory. Changes cannot be made to the registry and have them be active unless there is a reboot.

 

[Markbnj]

Originally posted by: corkyg
Reboot is required when changes are made in the registry in order to execute them. If no registry changes are required, no reboot is usually the path.

The focus is on the registry. When booted, the existing regitry is loaded into memory. Changes cannot be made to the registry and have them be active unless there is a reboot.

This doesn't make sense, corkyg, since I can write to the registry and read the changes back at any time. The registry hives are stored in files on disk, but the registry API manages write-through to those files, and they should be consistent at any given time. What you say would be true if program A is making a change to a part of the registry used by program B, and program B only reads the registry at startup. When program B is an OS component then a restart might well be required. Otherwise I think reboots are almost always caused by an inability to replace a file on disk due to is being locked by a running process.

 

[Nothinman]

Reboot is required when changes are made in the registry in order to execute them. If no registry changes are required, no reboot is usually the path.

The focus is on the registry. When booted, the existing regitry is loaded into memory. Changes cannot be made to the registry and have them be active unless there is a reboot.

That's so wrong it's not funny. I don't know how often the registry is sync'd to disk but any changes made to the registry are available immediately. Whether the service using those settings needs restarted or not to notice them is a completely separate issue.

The focus is on replacing in use files. Windows doesn't let you rename or delete in use files so to update those files a temporary copy of the new one is stored and a request to replace the file is made in the registry (ironically, considering your post) so that when the system boots up next time the old file is replaced with the new one.

 

[VinDSL]

Originally posted by: Rubycon
Not require a reboot after a program/install/change/update?

Heh! That's a loaded question!

And, it begs the question...

Why would you NOT want to "reboot after a program/install/change/update"?

I run a Slackware web server (personal) here at the house, and a CentOS server (production) in Atlanta.

Depending on what I'm 'installing/changing/updating', these NuX boxes require a reboot from time-to-time too!

I would judge that the same degree of exigency exists in Linux boxes and Windows machines - correct me if I'm wrong...

If you keep diddling around with your operating system - installing/uninstalling/changing/updating core components and applications on a near daily basis, what do you expect?!?!?!?

Bottom line: If you leave your computer (any computer) alone, it will probably run for years without a reboot. If you 'rearrange the furniture' every day, you should expect to see a few 'dust bunnies'.

 

[Nothinman]

I would judge that the same degree of exigency exists in Linux boxes and Windows machines - correct me if I'm wrong...

You are wrong, the only thing in Linux that requires a reboot is a kernel change. Some things like libc suggest a reboot so that you're 100% sure everything is using the new version of the library but if you just restart all of the affected services you'll be fine.

 

[VinDSL]

Originally posted by: nerp
I see reboots infrequently. Call them rareboots.

I even didn't have to reboot after installing the latest ATI Catalyst drivers for my video card and IGP. How about that?

Or... rarebits.

OMG!!! That makes me hungry!

I digress...

Ever have Welsh Rarebit???

For Americans, that would be - a grilled cheese sandwich with pork n' beans and Tabasco poured on top... :thumbsup:

Speaking of sloppy things - I removed ATI Catalyst Control Center from ALL my machines, a while back, and you wouldn't believe how much better and faster they run. It's like a virus! And, once you install it, it's practically impossible to get rid of. It requires the use of a special utility that's well hidden on the ATI web site.

What's the point?!?!?

Even when you run something as seemingly innocuous as ATI CC - which in reality manipulates many hardware functions on your machine - you should expect to reboot after you "install/change/update" it.

 

[VinDSL]

Originally posted by: Nothinman
You are wrong, the only thing in Linux that requires a reboot is a kernel change.

I see...

Ever heard of kexec?

 

[corkyg]

Originally posted by: Nothinman
The focus is on replacing in use files. Windows doesn't let you rename or delete in use files so to update those files a temporary copy of the new one is stored and a request to replace the file is made in the registry (ironically, considering your post) so that when the system boots up next time the old file is replaced with the new one.

We are saying the same thing in different words. If my synrax is wrong - my bad - but what you said is what I tried to say. Sorry about that.

 

[Maximilian]

Originally posted by: Rubycon
Not require a reboot after a program/install/change/update? It's annoying that you have to reboot after all these years - and we're still doing it! :|

You are the highly technical mod, you tell us I was hoping vista would never require a reboot, but it does.

 

[Cogman]

Driver updates are usually the only thing the I religiously reboot on. Other then that, not much else (ok, security updates as well). The only reason to reboot is so that an program using the outdated version of a driver/dll will be restarted using the new version of it, rather then continuing to rely on whats in memory.

 

[Nothinman]

Ever heard of kexec?

Yea and AFAIK it's not reliable enough to be used in production. Some of the hibernation people have been working on using it for that too but it's still got some pretty ugly warts to be worked out.

We are saying the same thing in different words. If my synrax is wrong - my bad - but what you said is what I tried to say. Sorry about that.

We didn't say the same thing at all, you were talking about the registry which is most certainly updateable at runtime. Some services may not be smart enough to notice the changes without a restart but that's a completely separate issue.

 

[corkyg]

All I tried to say was that while registry changes are entered immediately, they are not acted upon or take effect until until reboot.

"Making Registry Changes Effective
When you use an administrative tool to change the configuration of a system feature or service, the change usually takes effect either immediately or soon.
However, if you make the same change by editing the registry, for the change to take effect you might need to log off and log on again, restart the service, or restart Windows.
In general, if you change the value of any entry in the CurrentControlSet subkey, you must restart the computer for the changes to take effect. Also, if you use the registry editor Regedit.exe to change values for most entries in HKEY_CURRENT_USER, you must log off and log on again for the changes to take effect.
When a change to the value of an entry is not immediate, the description of the entry contains an Activation Method, which describes what you need to do to make the change effective."

From: Microsoft

 

[Nothinman]

All I tried to say was that while registry changes are entered immediately, they are not acted upon or take effect until until reboot.

Which isn't what you said and in a lot of cases isn't true. Most of the things that installers do like register filetypes, COM crap, etc do take affect at runtime.

 

[corkyg]

Are you disagreeing with the MS Tech Bulletin in principal or citing exceptions to the rule?

I accept that file types, and other data stuff do take effect at runtime. But those are not programmtic changes.

 

[Jeff7181]

When developers start developing applications that don't hook themselves into various parts of Windows.

 

[pallejr]

Originally posted by: corkyg
All I tried to say was that while registry changes are entered immediately, they are not acted upon or take effect until until reboot.

"Making Registry Changes Effective
When you use an administrative tool to change the configuration of a system feature or service, the change usually takes effect either immediately or soon...

That is something completely else. That is because the service/application that uses those rules only reads them when it starts.

 

[pallejr]

In vista they have also developed a system, where you can update critial system components runtime...

 

[zerogear]

To be fair, since XP and evenmoreso with Vista, a lot of drivers can be handled user-based, and doesn't need reboot after driver install, if you have anyone to blame, you really should blame the driver makers.