I’m spinning up a VM using qemu-system-x86_64 where I’m passing the VM ~500
interfaces.   Watching the boot process I see that the PCI probing/boot
phase itself takes 10s of minutes.    I see multiple qemu processes running
extensively (eg at 200% and 100%).   The VM has 14 vcpu and 24GB of RAM.



I don’t have PCI clue and so I’m wondering what is actually happening
during this phase that would take so long?    Besides reducing the number
of interfaces, is there a way to reduce the time the VM takes to boot
during this PCI phase?



Thanks

Reply via email to