Is it normal to have near 10x difference in TTFB performance between production and vagrant?
For a site running Twenty Seventeen + WPML + WooCommerce ? I know these plugins are heavy but near 3s for unmodified homepage load on my i7 4710HQ cpu + samsung ssd? Can it be due to running in VMWare instead of VirtualBox?
I have setup NFS and Bind vagrant plugins and they seem to run:
The production server is a dual core VM with ssd backed hdd.
My vagrant machine is using 2 out of 16GB RAM and two out of four cores.
Some of these suggestions are already implemented in the Trellis Vagratnfile and you may find that increasing the vagrant_memory: in vagrant.default.yml is useful.
Yes, I tried most approaches found on the web. It all boils down to NFS server performance on Windows I guess.
Also using WPML + WooCommerce doesn’t seem to help (behemoths).
Lately I’ve been trying CIFS/smb, but it seems even slower that way, albeit a bit more stable.
The thing is I know how much nicer the whole Vagrant/NFS deal is when ran on top of Linux, but I have Visual Studio needs and Windows is still my first choice.
I have one last (crazy) idea to install Hyper-V or ESX as the root hypervisor on my boot drive, and hopefully switch between machines in a less disruptful way than rebooting. Nested virtualization seems possible but could potentially result in an even slower solution.
