The Xen Project will bring non-disruptive patching to its hypervisor, version 4.7 of which is set to debut onJune 3rd, 2016.
That date may wobble a little, because the feature freeze scheduled for April 1st was put back a week in order to let contributors sort themselves out over Easter.
But it's known that a sub-project called xSplice will make it in to the release, bringing with it non-disruptive patching so that users can apply repairs to the hypervisor without having to restart virtual machines.
We even know that release candidate one of the new release is being held up while xSplice is merged, thanks to this thread in which Xen project peeps debate what to put in the press release when 4.7 launches.
“xSplice is definitively the first one on my list,” writes project contributor Wei Liu. “I'm actually holding off cutting RC1 waiting for it to be merged.”
Consensus seems to be the project will get most column inches if it makes a big deal about Coarse-grain lock-stepping Virtual Machines for Non-stop Service (COLO), a high-availability effort that sees two VMs run in parallel but on different servers. The idea is to make sure there's always a failover resource ready to run – indeed running the very same workload the primary VM runs. COLO won't be on by default, but looks like a fine idea for users who need high availability.
Adding the Advanced Configuration and Power Interface (ACPI) on ARM processors gets lots of support. As it should because one of ARM-powered devices' advantages is low power consumption so making sure Xen can use ACPI's power management features makes an awful lot of sense.
The Register's virtualisation desk expects xSplice will make the press release, if only because Xen's had bugs a-plenty of late so non-disruptive patching will show the project's knocked a problem on the head. ®