At 01.30 we rebooted server c2.
The whole process of unmounting the virtual servers (VMs) and virtual functions, powering down the hosting server, powering up the hosting server and remounting all of the virtual functions took 118 seconds.
The downtime to our users, during these 118 seconds was nil.
But one of our customers has asked the question ‘Why do you switch it off and switch it back on again?’
This is quite a clever question.
In the hosting world, where we deal with physical servers and VMs, which are supercomputers when compared to the average high-spec desktop or laptop, we reboot our machines for a few reasons:
- to clear temporary functions out of the processing memory (out of the thinking space)
- to clear temporary files out of the hard disks (out of the storage space), and
- to ‘bed in’ software upgrades and patches to the physical servers and VMs
None of these reasons are unusual; they are ‘business as usual’ tasks that every organisation routinely carries out.
So that’s why.
We will reboot at least one of our servers at least once a week.
And we will continue to offer our users an uninterrupted service (even at 01.30) when we do it.