2

In the early days of Microsoft Windows, after an application crashed it was often a good idea to reboot the system. The reason why is that Windows would often fail to release resources that were used by the application that crashed. If you didn't reboot, Windows would eventually run out of resources, causing other applications to fail or for the OS itself to experience problems.

With Windows 7, 8.x, and 10, is there any reason to reboot Windows after an application crashes or is force-terminated by the user due to a hang?

2 Answers 2

2

No, systems based on the NT kernel (including 10, 7, Vista, XP, 2000) are structured completely differently from Win95/98 and generally avoid such problems. Kernel objects are reference-counted by the Object Manager, and all GDI resources are associated with the process that holds them.

3
  • Thank you for your answer. I hope your answer is correct, but John has written a starkly different answer. The reasons I hope your answer is correct is because it is easier, less time consuming, and doesn't involve any interruptions in workflow. I've been on SuperUser for a number of years, and I'm surprised I don't remember reading your answers before (and I don't recall you ever answering one of my questions before). I think your staggering number of SE rep points and badges largely speak for themselves, and so I'm leaning towards your answer over John's, but John knows his stuff too. Commented Nov 25, 2020 at 14:26
  • Your answer and John's are almost polar opposites, and I hope we can work together to figure out the truth. Commented Nov 25, 2020 at 14:26
  • I was quite tired and busy with several tasks when I wrote the words "the truth". A much better choice of words would have been "best practices". Commented Nov 25, 2020 at 23:54
0

Usually no, with rare exceptions.

@user1686 is right that modern Windows manages resources way more reliably than pre-2000 DOS-based versions. But in some cases there's more to a crash than just resource management:

  • Complex software with bugs. Sometimes software consists of multiple components that have to work together and one of them crashing leaves others in inconsistent state from which they can't recover. Manually hunting down and restarting all components can be hard, but a restart should just work.

  • Bad virtualization crashes. It happened to me once that a virtual machine crashed so bad that all VMs refused to start and the host failed to reboot properly (OS shut down, but the PC didn't reboot). I had to hold the power button unless the system powered off, then cold boot it.

  • Hardware-related failures. Hardware left in inconsistent state either due to bad driver or firmware could malfunction and not work properly until a reboot.

  • Probably more...

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .