Oracle revs VirtualBox, mushrooms memory The open source VirtualBox hypervisor for PCs and servers got a major release on Tuesday when Oracle – which controls the VirtualBox project – kicked out version 4.1. The big change with VirtualBox 4.1 is that both the hypervisor and the guest virtual machine partitions that ride atop of the hypervisor can address more main …
I don't care how many widgets, features and fancy frills, if it isn't *rock-sold reliable* then it's an immediate faceplant. Businesses at almost any level need *reliability*, why do you never discuss this?
These articles keep banging on about big iron and fancy this and supercomputer that but never seem to ask what matters.
(which is why I use VMware for personal use, from experience of too much VBox freakery, however, important note to VMware users, found that the graphics driver in the vmware tools when running windows 2008 will randomly hard-hang the VM so don't install it. Just a hint).
I am involved in running a top 500 supercomputer site, and it is reliable. So reliable, in fact, that the customer is saying that they want to manufacturer outages on a certain service so that their users don't get to automatically expect 100% availability.
The main secret as far as I am concerned is the old adage 'if it ain't broke, don't fix it'. Really annoys me when IBM say we *have* to upgrade the software stack to remain in a supported state!
So in answer to the comment, don't tar all services with the same brush.
"It d[oe]s not talk about reliability or advocate use of Virtual Box in the enterprise."
From the article: "It is not clear that Oracle, [...] wants companies to use VirtualBox on those big jobs, but now it is possible – in theory, at least."
"The article merely informs us of a new release"
And the point of informing us of this new release with so many extra features, if not to suggest we use it, is... what? Reliability is a feature. The biggie.
@Peter Gathercole: I don't know what you're doing with it but I managed to break it n times over. I can only speak from my experience; push it too far and it fails. If you install it with a single virtual disk and don't do anythign else then yes, it is very reliable (IME).
Also can't quite get why you want to manufacture outages to be more common than failures on the real machines they run on (can't be more reliable, can they? and they used to use real machines, no?)
@author. On reflection I have to retract at least some of what I said. You've been discussing HPC more often now with fair bit of time spent on energy consumption per flop and theoretical vs peak performance, which is very relevant, so I was somewhat unfair.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
