The so called "meltdown" bug recently discovered in Intel CPU's [[read more]] has raised some questions about how it will affect cloud services powered by Intel CPU's. What we know so far is that it's up to operating system manufacturers to come up with a patch. Windows servers will receive an update next Tuesday while Linux servers will require a kernel update.
Current status is that we've already patched our shared environments such as the servers hosting our websites, API, databases etc.. Operators in charge of the servers acts a bit different tho, some has acknowledged the issue and promised a solution within short while others hasn't said anything. Although we do know that work is being made in the back end and images with updated secure kernels are being rolled out at this moment.
How will this affect deployment of new servers
You may continue to deploy new servers as usual. Insecure images has been removed from stock but could remain in our cache for a while longer. Although even if we have them in cache they will not deploy, instead you'll see an error message like "Invalid image"
How can I upgrade existing servers
Windows servers will receive an update next Tuesday
Linux servers needs a kernel update, this may be solved differently among different distributions. Time before a patch is available may differ as well.
Anything else I should do to prepare
As the physical nodes requires an upgrade as well a planned host node reboot will occur sometime during this weekend or next week. A host node reboot has the same impact as rebooting your VPS but slower which means a few minutes of downtime during this maintenance frame. Prepare your running services and make sure they handle a reboot properly. If you cannot connect to your server or a running service on it make sure the service is running.
Will I get a discount
Even tho it's not our responsibility we've decided to compensate you all for this mess a little, by continue running the 25% cryptocurrency discount until February instead of changing it back to 10% which we initially intended to do at 2018-01-01.