Re: [Qemu-devel] [PATCH v8 00/54] Postcopy implementation

["Dr. David Alan Gilbert" <dgilbert@redhat.com>]

* Bharata B Rao (bharata@linux.vnet.ibm.com) wrote:
> On Mon, Sep 28, 2015 at 05:51:39PM +0100, Dr. David Alan Gilbert (git) wrote:
> > From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
> > 
> >   This is the 8th cut of my version of postcopy.
> > 
> > The userfaultfd linux kernel code is now in the upstream kernel
> > tree, and so 4.3-rc3 can be used without modification.
> > 
> > This qemu series can be found at:
> > https://github.com/orbitfp7/qemu.git
> > on the wp3-postcopy-v8 tag
> > 
> > 
> > Testing status:
> >   * Tested heavily on x86
> >   * Smoke tested on aarch64 (so it does work on different page sizes)
> >   * Power is unhappy for me (but gets further than the htab problem
> >     v7 used to have) (I get a kvm run failed)
> 
> Seems to be completing successfully on Power. But it takes 2min for the
> migration status to transition from setup to active.
> 
> Host: 4.3.0-rc3+
> Guest: 4.3.0-rc3+
> QEMU: wp3-postcopy-v8 of your tree.
> 
> # ./ppc64-softmmu/qemu-system-ppc64 --enable-kvm --nographic -machine pseries -m 8G,slots=32,maxmem=32G -device virtio-blk-pci,drive=rootdisk -drive file=/home/bharata/F20-snap1,if=none,cache=none,id=rootdisk,format=qcow2 -vga none -net nic,model=virtio -net user -redir tcp:2000::22 -smp 16,maxcpus=32 -serial pty
> 
> (qemu) migrate_set_capability x-postcopy-ram on
> (qemu) migrate -d tcp:localhost:4444
> (qemu) info migrate
> capabilities: xbzrle: off rdma-pin-all: off auto-converge: off zero-blocks: off compress: off events: off x-postcopy-ram: on 
> Migration status: setup
> total time: 0 milliseconds
> 
> same status for around 2min...

That's interesting; I saw that behaviour on my aarch64 box, but not on
my power box or on x86.  Can you try using tcp:127.0.0.1:4444 to force
ipv4 (that fixed it for me on aarch64).  On the aarch box I found that
it still happened with head of tree qemu and so decided it wasn't my
postcopy world; I'm assuming what's happening is that it's trying
to connect to the IPv6 address, timing out and then trying IPv4.


> (qemu) info migrate
> capabilities: xbzrle: off rdma-pin-all: off auto-converge: off zero-blocks: off compress: off events: off x-postcopy-ram: on 
> Migration status: active
> total time: 130089 milliseconds
> expected downtime: 300 milliseconds
> setup: 24 milliseconds
> transferred ram: 79454 kbytes
> throughput: 50.09 mbps
> remaining ram: 7670688 kbytes
> total ram: 8388864 kbytes
> duplicate: 160684 pages
> skipped: 0 pages
> normal: 18860 pages
> normal bytes: 75440 kbytes
> dirty sync count: 1
> 
> (qemu) migrate_start_postcopy
> 
> (qemu) info migrate
> capabilities: xbzrle: off rdma-pin-all: off auto-converge: off zero-blocks: off compress: off events: off x-postcopy-ram: on 
> Migration status: postcopy-active
> total time: 135627 milliseconds
> expected downtime: 43 milliseconds
> setup: 24 milliseconds
> transferred ram: 338598 kbytes
> throughput: 74.02 mbps
> remaining ram: 1688384 kbytes
> total ram: 8388864 kbytes
> duplicate: 1600406 pages
> skipped: 0 pages
> normal: 75258 pages
> normal bytes: 301032 kbytes
> dirty sync count: 0
> dirty pages rate: 98 pages
> 
> (qemu) info migrate
> capabilities: xbzrle: off rdma-pin-all: off auto-converge: off zero-blocks: off compress: off events: off x-postcopy-ram: on 
> Migration status: completed
> total time: 136898 milliseconds
> downtime: 685 milliseconds
> setup: 24 milliseconds
> transferred ram: 1194196 kbytes
> throughput: 72.00 mbps
> remaining ram: 0 kbytes
> total ram: 8388864 kbytes
> duplicate: 1810921 pages
> skipped: 0 pages
> normal: 286839 pages
> normal bytes: 1147356 kbytes
> dirty sync count: 2

Great; is the guest happy?

Dave

> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK