The Advanced Settings section in mPanel allows modification of some of the internal settings of your cloud server. Any modifications to advanced settings requires a server reboot to apply, so be sure to select an appropriate time for a short outage.
Normally there is no requirement to touch any of these settings: if you do not have any issues with your server we recommend that you do not alter any advanced settings.
This document describes what settings are available and what they do.
The "Processor Model" determines the level of host CPU functionality exposed to your cloud server's virtual processor. In other words, it determines the CPU flags - and thus advanced processor features - available for use. It does *not* change the physical processor that your server is running on. There are currently two options available:
- Intel KVM Processor v2
This model is selected for most VPS provisioned since December 2016. It provides the following CPU flags:
fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx rdtscp lm constant_tsc arch_perfmon rep_good nopl eagerfpu pni pclmulqdq ssse3 cx16 sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm xsaveopt fsgsbase smep erms
At a high level compared to the more basic "Common KVM Processor", it adds support for SSE3/SSE4.1/SSE4.2 and AES encryption . The exact set of additions is:
vme rdtscp arch_perfmon rep_good eagerfpu pclmulqdq ssse3 sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand xsaveopt fsgsbase smep erms
This wikipedia page provides a short summary of what each flag means.
Of particular note for many customers the Intel KVM Processor v2 model also enables the "kvmclock" feature, which allows linux VPS to directly use the host node's realtime clock instead of having to keep track of the current time itself. Customers who have experienced issues with timekeeping may wish to change to the newer processor model for this feature alone.
- Common KVM Processor
This is the base model provided by the KVM hypervisor, and was selected for all VPS provisioned before 2017. It provides the following CPU flags:
fpu de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx lm constant_tsc nopl pni cx16 hypervisor lahf_lm
The "kvmclock" feature is disabled on this model, so all VPS must run an NTP service to keep accurate time.
The video card option determines what type of graphics card is emulated. Note that whichever option is selected, neither is a "real" graphics card and the performance is comparatively awful. The video card is used purely for rendering the rescue console from within mPanel. There are two options:
- Standard VGA with VESA extensions. This is the more capable of the two options in terms of resolution and bit-depth, so it is preferable if your operating system has good support for it.
- Cirrus Logic GD5446: This is an emulatation of a very old video card. It is less capable than the first option, but due to its age is well supported by older or more esoteric operating systems.
There is no difference in terms of performance; the only reason to change this option is if your operating system is having issues with the current selection.
HyperV is a hypervisor produced by Microsoft; the KVM hypervisor that BinaryLane uses implements some of its functionality to improve performance for Windows servers. We pre-enable this for all our Windows servers that support it; for other operating systems it defaults to off and there is likely no reason to enable it.
When emulated devices is enabled, the KVM specific "VirtIO" disk drive and network devices are removed, and replaced with emulated versions of physical hardware: an old IDE HDD and an Intel E1000 network card. Emulated devices are much slower than the VirtIO devices, and so this option should not be enabled unless absolutely necessary.
For all our operating system images except the "BYO Operating System (with virtio disabled)", this option is disabled and should remain left off.
When this option is enabled, the functionality necessary to run your own KVM guests within your BinaryLane VPS is enabled. This would allow you to install and operate ProxMox for example. Nested Virtualization is disabled by default.
Note that all the BinaryLane networking limits - one MAC address per VPS, restricted to specific IPs - still apply to public cloud so this is feature is generally only useful in combination with Virtual Private Cloud.