[Qemu-discuss] What does "COW" mean?
Hi, I'm new to qemu community, and I'm trying the COW image format (old but simple:-). I have read through its source code, and didn't find anything about 'copy on write'. I wonder wthat "COW" stands for? -- Regards, shhuiw
Re: [Qemu-discuss] network latency for virtio-net-pci
On 09/02/2014 06:05 AM, Zhang Haoyu wrote: We running qemu 1.5, and one of the users complains on high latency on the overlay network of openstack. I done some research and comparation and found those numbers: V - virtual machine with qemu/virtio-net-pci tap device. H - hardware server S - hardware switch O - openvswitch bridge (OVS 2.0) V-O-V - 300 µs V-H - 180 µs H-S-H - 140 µs V-O-S-O-V - 600 µs After doing some math with linear equations and few more tests I found following latencies: internal linux latency - 50 µs hardware switch latency - 40 µs openvswitch (gre-mode) latency 40 µs and, most important: QEMU device - 130 µs Did you use vhost-net? Yep. Here command line for VM (network part) qemu-system-x86_64 -machine accel=kvm:tcg ... -netdev tap,fd=25,id=hostnet0,vhost=on,vhostfd=26 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=fa:16:3e:0e:4c:5c,bus=pci.0,addr=0x3 ... 2. Is any way to reduce latency for qemu network devices? 1) try to patch vhost: add polling mode N.B., you can also introduce similar implementation to guest virtio driver. 2) try to patch Workqueue based vhost work scheduling 3) tune CFS paras for lower latency: sched_min_granularity_ns, sched_latency_ns, sched_wakeup_granularity_ns 4) renice vhost thread to higher priority(e.g., -10) 5) bind vcpu to pcpu Thanks, I'll try.
Re: [Qemu-discuss] What does "COW" mean?
On 02/09/2014 10:33, shhuiw wrote: Hi, I'm new to qemu community, and I'm trying the COW image format (old but simple:-). I have read through its source code, and didn't find anything about 'copy on write'. I wonder wthat "COW" stands for? COW = Copy On Write is a general computing technique where something (in this case a disk image file) doesn't really copy data until someone actual tries to write to (change) their copy, then at that very moment, it makes a real copy for that someone to write to. In qemu context this is used when you make a "snapshot" copy of an entire virtual disk and tells qemu that the running virtual machine should only write to one of the copies while the other copies are kept unchanged for any reason (such as backup). When you tell qemu to do this with a COW disk image format, it just makes copies some data about where different parts of the disk image are stored in the file, and sets some flags such that when the virtual machine actually writes to the disk image, it is given its own copy (in the COW file) of the part of the disk it has written to. For the (usually at least 90%) of the disk that the virtual machine doesn't write to, it just reads back from the original copy without wasting time and disk space on an actual copy. As a bonus feature, some recent versions of the qcow2 file format also uses this technique for any part of the disk that contains all zeroes, which saves space as long as the virtual machine doesn't completely fill its virtual disk. Reading back zeroes from nowhere is faster and cheaper than reading zeroes from a real disk, making this a win-win in most situations. Enjoy Jakob
Re: [Qemu-discuss] What does "COW" mean?
At 2014-09-02 04:33:50, "shhuiw" wrote: > > >Hi, > >I'm new to qemu community, and I'm trying the COW image format (old but >simple:-). >I have read through its source code, and didn't find anything about 'copy on >write'. >I wonder wthat "COW" stands for? Sorry for my unclear expression. I mean when copy-on-write happens if COW image format is used, and how the COW driver code handles cop-on-write? > > >-- > >Regards, >shhuiw -- Regards, shhuiw
