Share
Embed
Transcript
71 slides
-
ԾԽڥͷެฏੑʁ • KVMͷήετΛͨ͘͞Μ্ཱͪ͛ͨ࣌ɺϊʔυؒ ͷ௨৴ੑೳCPUར༻࣌ؒͷެฏੑͲͷఔอ ͨΕ͍ͯΔͷ͔ • CPUར༻࣌ؒ • ௨৴ྔ •
ϨΠςϯγ • VMลΓͷෛՙͷΒ͖ͭ 1211݄20Ր༵ -
ੑೳଌఆڥ • ̎ͷLinuxػɾ10G NIC • 1ʙ64ͷKVMήετɾ128ϓϩηεͷ netperf 1VM→VM͋ͨΓ128ϑϩʔ 2VM→VM͋ͨΓ64ϑϩʔ 4VM→VM͋ͨΓ32ϑϩʔ…
64VM→VM͋ͨΓ2ϑϩʔ • TCP Request/ResponseϞʔυ 1byteͷύέοτΛϐϯϙϯ VMϗετ ήετ ςετػ 10G NIC 10G NIC ήετ netperf netserver netperf 1211݄20Ր༵ -
ϋʔυʗιϑτͷεϖοΫ Distribution Ubuntu Server 12.10 Linux Kernel 3.5.0-18-generic QEMU-KVM 1.2.0
Netperf 2.5.0 CPU(VMϗετ) Intel Core i7 980 (3.33GHz) Memory(VMϗετ) 24GB CPU(ςετػ) Intel Core i7 860 (2.8GHz) Memory(ςετػ) 8GB NIC Intel 82599(ixgbe) ཧ6ίΞɺཧ12ίΞ 1211݄20Ր༵ -
ήετϚγϯͷઃఆ • ԾCPUɿ1ͭ • ԾNICͷઃఆ • macvtapͰbridge / vhost-net •
SR-IOV • 512MBͷϝϞϦׂΓͯ 1211݄20Ր༵ -
͓͞Β͍ɿSR-IOV • ཧNIC͕VMʹରͯ͠ԾNICΛఏڙɹIOϋΠύʔόΠβʔΛհࡏͤͣʹߦΘΕΔ • ׂࠐΈ͚ͩԾԽग़དྷ͍ͯͳ͍ͷͰKVMΛ௨ͯ͡సૹ͞ΕΔ • ࠷ੑೳ͕ߴ͍͕ϋʔυରԠ͕ඞཁʗήετͷύέοτΛϑΟϧλʔɾվม͢Δࣄ͍͠ NIC Χʔωϧ Χʔωϧ
ήετ Ϣʔβ PF PFυϥΠό qemu kvm VF υϥΠό TCP/IP ελοΫ NIC2 ϑΥϫʔυ ׂΓࠐΈ ཧׂΓࠐΈ ׂΓࠐΈ ׂࠐΈ ϋϯυϥ VF %."ύεεϧʔ ϋΠύʔόΠβ 1211݄20Ր༵ -
macvtap <interface type='direct'> <mac address='52:54:00:6b:28:01'/> <source dev='eth1' mode='vepa'/> <model type='virtio'/>
<address type='pci' domain='0x0000' bus='0x00' slot='0x07' function='0x0'/> </interface> 1211݄20Ր༵ -
SR-IOV <hostdev mode='subsystem' type='pci' managed='yes'> <source> <address domain='0x0000' bus='0x04' slot='0x10'
function='0x2'/> </source> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </hostdev> 1211݄20Ր༵ -
ൺֱରͱͯ͠ͷ ࣮ػੑೳଌఆ • ̎ͷLinuxػɾ10G NIC • KVMήετͰnetserver࣮ߦ • 128ϓϩηεͷnetperf •
TCP Request/ResponseϞʔυ 1byteͷύέοτΛϐϯϙϯ VMϗετ ςετػ 10G NIC 10G NIC netperf netserver netperf netserver 1211݄20Ր༵ -
ඵؒτϥϯβΫγϣϯ 0 1375.00 2750.00 4125.00 5500.00 1 2 4 8
16 32 64 vhost-net SR-IOV baremetal VM transaction/sec ཧίΞʹ͍ۙลΓͰ ੑೳ࠷ߴ 1211݄20Ր༵ -
ϨΠςϯγ 0 750.00 1500.00 2250.00 3000.00 1 2 4 8
16 32 64 vhost-net SR-IOV baremetal VM ϨΠςϯγ ཧίΞʹ͍ۙลΓͰ ϨΠςϯγ࠷ద 1211݄20Ր༵ -
CPUෛՙ 0 25.00 50.00 75.00 100.00 1 2 4 8
16 32 64 vhost-net SR-IOV baremetal VM ύʔηϯςʔδ ෛՙΛVMͷίΞʹ ͔͠ࢄͰ͖ͯͳ͍ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ඵؒτϥϯβΫγϣϯʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 vhost-net SR-IOV VM ύʔηϯςʔδ vhost-net < 10% SR-IOV < 30% 1211݄20Ր༵ -
VMؒͷภΓ ʢ ϨΠςϯγʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 vhost-net SR-IOV VM ύʔηϯςʔδ vhost-net < 10% SR-IOV < 30% 1211݄20Ր༵ -
VMؒͷภΓ ʢ CPUෛՙʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 vhost-net SR-IOV VM ύʔηϯςʔδ vhost-net < 10% SR-IOV < 20% 1211݄20Ր༵ -
ൺֱ݁Ռ • ཧίΞʹ͍ۙVMͩͱੑೳ͕Ұ൪ྑ͍Α͏ʹ ݟ͑Δ • ཧίΞͷ10ഒͷVMΛ࣮ߦͯ͠ɺ͜ͷෛՙύλ ʔϯͰۃͳϨΠςϯγͷѱԽෛՙͷภΓͳͲ ݟΒΕͳ͔ͬͨ • ཧڥͱൺ͔ͯͳΓੑೳ͕ѱ͍
→ෛՙΛෳίΞࢄग़དྷ͍ͯͳ͍ࣄ͕ݪҼͷީ ิͱͯ͠ߟ͑ΒΕΔ 1211݄20Ր༵ -
Ծڥ্ͰίΞࢄʁ • ԾCPUΛ1-16·Ͱ૿ͯ͠ΈΔ • ίΞΛ૿͚ͨͩ͠ͰωοτϫʔΫε λοΫ͕ฒྻͰΒͳ͍ • ✕ RSSɿԾNICɾVF͕ະରԠ •
◦ RPSɿήετଆLinuxͷઃఆͰ༗ޮԽ͠ ͯΈΔ 1211݄20Ր༵ -
RPS $ echo "f" > /sys/class/net/eth1/queues/rx-0/ rps_cpus $ echo 4096
> /sys/class/net/eth1/queues/ rx-0/rps_flow_cnt $ echo 32768 > /proc/sys/net/core/ rps_sock_flow_entries 1211݄20Ր༵ -
ඵؒτϥϯβΫγϣϯ ʢvhost-netʣ 0 1375.00 2750.00 4125.00 5500.00 1 2 4
8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM transaction/sec VM =< 4ͳΒվળ 1211݄20Ր༵ -
ඵؒτϥϯβΫγϣϯ ʢvhost-net, RPSʣ 0 1375.00 2750.00 4125.00 5500.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM transaction/sec VM =< 4ͳΒRPSͰ ੑೳվળ 1211݄20Ր༵ -
ඵؒτϥϯβΫγϣϯ ʢSR-IOVʣ 0 1375.00 2750.00 4125.00 5500.00 1 2 4
8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM transaction/sec VM =< 4ͳΒվળ 1211݄20Ր༵ -
ඵؒτϥϯβΫγϣϯ ʢSR-IOV, RPSʣ 0 1375.00 2750.00 4125.00 5500.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM transaction/sec VM =< 4ͳΒRPSͰ ੑೳվળ 1211݄20Ր༵ -
ϨΠςϯγ ʢvhost-netʣ 0 750.00 1500.00 2250.00 3000.00 1 2 4
8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM ϨΠςϯγ tpsͱ͍͍ͩͨಉ͡ 1211݄20Ր༵ -
ϨΠςϯγ ʢvhost-net, RPSʣ 0 750.00 1500.00 2250.00 3000.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM ϨΠςϯγ tpsͱ͍͍ͩͨಉ͡ 1211݄20Ր༵ -
ϨΠςϯγ ʢSR-IOVʣ 0 750.00 1500.00 2250.00 3000.00 1 2 4
8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM ϨΠςϯγ tpsͱ͍͍ͩͨಉ͡ 1211݄20Ր༵ -
ϨΠςϯγ ʢSR-IOV, RPSʣ 0 750.00 1500.00 2250.00 3000.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 baremetal VM ϨΠςϯγ tpsͱ͍͍ͩͨಉ͡ 1211݄20Ր༵ -
CPUෛՙ ʢvhost-netʣ 0 25.00 50.00 75.00 100.00 1 2 4
8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ VM =< 4ͳΒɺcpu = 1 ͷ࣌ΑΓෛՙΛࢄͰ ͖͍ͯΔʁ 1211݄20Ր༵ -
CPUෛՙ ʢvhost-net, RPSʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ RPSͰΑΓෛՙΛࢄ 1211݄20Ր༵ -
CPUෛՙ ʢSR-IOVʣ 0 25.00 50.00 75.00 100.00 1 2 4
8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ 1211݄20Ր༵ -
CPUෛՙ ʢSR-IOV, RPSʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ඵؒτϥϯβΫγϣϯʗvhost-netʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ cpu >= 8, VM >=16ͷ࣌ʹ 15%Λ͑ΔภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ඵؒτϥϯβΫγϣϯ/vhost-net, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ cpu >= 8, VM >=16ͷ࣌ʹ 15%Λ͑ΔภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ඵؒτϥϯβΫγϣϯʗSR-IOVʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ૯εϨου͕ཧCPUΛ ͑ͨลΓͰେ͖ͳภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ඵؒτϥϯβΫγϣϯʗSR-IOV, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ૯εϨου͕ཧCPUΛ ͑ͨลΓͰେ͖ͳภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ϨΠςϯγʗvhost-netʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ tpsͱ͍͍ͩͨಉ͡ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ϨΠςϯγʗvhost-net, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ϨΠςϯγʗSR-IOVʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ૯εϨου͕ཧCPUΛ ͑ͨลΓͰେ͖ͳภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ ϨΠςϯγʗSR-IOV, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ૯εϨου͕ཧCPUΛ ͑ͨลΓͰେ͖ͳภΓ 1211݄20Ր༵ -
VMؒͷภΓ ʢ CPUෛՙʗvhost-netʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ͋·Γେ͖ͳภ ΓݟΒΕͳ͍ 1211݄20Ր༵ -
VMؒͷภΓ ʢ CPUෛՙʗvhost-net, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 64 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ ͋·Γେ͖ͳภ ΓݟΒΕͳ͍ 1211݄20Ր༵ -
VMؒͷภΓ ʢ CPUෛՙʗSR-IOVʣ 0 25.00 50.00 75.00 100.00 1 2
4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ VMͷ্ঢͰ एׯ૿Ճ 1211݄20Ր༵ -
VMؒͷภΓ ʢ CPUෛՙʗSR-IOV, RPSʣ 0 25.00 50.00 75.00 100.00 1
2 4 8 16 32 cpu1 cpu2 cpu4 cpu8 cpu16 VM ύʔηϯςʔδ VMͷ্ঢͰ एׯ૿Ճ 1211݄20Ր༵ -
ൺֱ݁Ռ • ԾϚγϯͰ࣮ػੑೳͷҎ্Λग़ͤͨͷVM͕͘͝গͳ͘ vCPU͕͔ͳΓଟ͍࣌ͷΈ → ͜ͷ݅Ͱ͔͠RPSͰੑೳ͕ग़ͳ͍ • ૯vCPUΛଟͯ͘͠ҙ֎ͱੑೳԼ͕Βͳ͍ • SR-IOVͷ߹ʹVF͕ཧίΞΛ͑ͨลΓͰ͔ͳΓੑೳʹภ
Γ͕ͰΔ • NIC্ͷΩϡʔʹରͯ͠ड৴ॲཧʹׂΓͯΔίΞ͕ෆͯ͠ ͜ͷΑ͏ͳঢ়گʹͳΔͱਪଌ • vhost-netͰͦͷΑ͏ͳݱ͕ݟΒΕͣɺ૯vCPUΛ૿ͯ͠ ൺֱతެฏʹॲཧ͞Ε͍ͯΔΑ͏ʹݟ͑Δ 1211݄20Ր༵ -
$ ls /sys/fs/cgroup blkioɹϒϩοΫI/Oͷ੍ݶ cpuɹCPUϦιʔεͷ੍ݶ cpuacctɹCPUͷߦ࣌ؒͳͲͷใऩू cpusetɹCPUͷׂΓ͚ʹؔ͢Δ੍ݶ devicesɹσόΠεͷΞΫηεڐՄɾڋ൱ freezerɹάϧʔϓʹଐ͢ΔϓϩηεͷҰ࣌ఀࢭɾ࠶։ memoryɹϝϞϦϦιʔεͷ੍ݶ
net_clsɹૹग़ύέοτʹtc༻ͷλάΛઃఆ nsɹωʔϜεϖʔεʢίϯςφ༻ʣ 1211݄20Ր༵ -
cd /sys/fs/cgroup/cpu $ mkdir grp_a $ echo ‘12254’ > grp_a/tasks
$ echo ‘2184’ > grp_a/tasks $ echo ‘512’ > grp_a/cpu.shares $ mkdir grp_b $ echo ‘9012’ > grp_b/tasks $ echo ‘1024’ > grp_b/cpu.shares άϧʔϓ࡞ άϧʔϓॴଐϓϩη εΛՃʢPIDʣ άϧʔϓͷλεΫͰ༻ग़དྷ ΔCPU࣌ؒͷ૬ରత 1211݄20Ր༵ -
virshͱcgroup $ virsh schedinfo vm0 Scheduler : posix cpu_shares :
1024 vcpu_period : 100000 vcpu_quota : -1 $ virsh schedinfo --set cpu_shares=512 vm0 1211݄20Ր༵ -
̎ͭͷάϧʔϓʹެฏʹϦ ιʔεΛׂΓͯͯΈΔ • /sys/fs/cgroup/grp_a • VMΠϯελϯεɿ̒ݸ • cpu.sharesɿσϑΥϧτʢ̍̌̎̐ʣ • /sys/fs/cgroup/grp_b
• VMΠϯελϯεɿ̍̌ݸ • cpu.sharesɿσϑΥϧτʢ̍̌̎̐ʣ 1211݄20Ր༵ -
άϧʔϓͷ࡞Γํ $ virsh start <VM໊> $ mkdir /sys/fs/cgroup/cpu/grp_a $ cat
/sys/fs/cgroup/cpu/libvirt/qemu/<VM໊>/tasks > tmp $ cat /sys/fs/cgroup/cpu/libvirt/qemu/<VM໊>/ vcpu*/tasks >> tmp $ cat tmp > /sys/fs/cgroup/cpu/grp_a/tasks 1211݄20Ր༵ -
άϧʔϓ͝ͱͷ߹ܭൺֱ ʢcgroup༻࣌ʣ 0 0.50 1.00 1.50 2.00 CPU Net grp_a
grp_b 1.10 1.14 1211݄20Ր༵ -
ʦࢀߟʧάϧʔϓ͝ͱͷ߹ܭൺֱ ʢcgroupະ༻࣌ʣ 0 0.50 1.00 1.50 2.00 CPU Net grp_a
grp_b 1.64 1.65 1211݄20Ր༵ -
άϧʔϓ͝ͱͷ߹ܭൺֱ ʢquota=200000,period=50000ʣ 0 0.50 1.00 1.50 2.00 CPU Net grp_a
grp_b 1.01 0.63 1211݄20Ր༵ -
ݕূ݁Ռ • ͔֬ʹάϥϑ্៉ྷʹͳΔͷ͕ͩɺCPU͕ͲΜͳʹ ۭ͍͍ͯͯquota·ͰߦͬͨΒઈର࣮ߦ͞Εͳ͍ • ͔ͱ͍࣮ͬͯߦ͖͠Εͳ͍ఔͷquotaΛ༩͑Δͱઃఆ͠ ͳ͍߹ͱมΘΒͳ͘ͳΔ • CPUωοτϫʔΫIOͪΐ͏ͲΑ͘ͽͬͨΓ1.0ͱ͍ ͏༁ʹ͍͔ͳ͔ͬͨ
• period͕͗͢ΔͱVM͕͍࣮ؒߦ͞Εͳͯ͘ྑ͘ͳ ͍͜ͱ͕ى͖ͦ͏ 1211݄20Ր༵ -
net_cls $ tc class add dev virbr0 parent 10: classid
10:1 htb rate 24Mbit $ echo 0x100001 > /sys/fs/cgroup/net_cls/ grp_a/net_cls.classid ϓϩηε͔Βૹ৴͞ΕΔύέοτʹtcͷ λάΛ͚ΔࣄʹΑΓଳҬ੍ݶ͕ग़དྷΔ 1211݄20Ր༵ -
·ͱΊ • KVMʹΑΔԾԽڥͷϦιʔεׂΓͯͷެฏੑʹ͍ͭͯ ݕূͨ͠ • ࠓճͷ࣮ݧͷ݅ԼͰɺ࣮CPUΛେ͖͑͘ΔVMͰ ۃͳϨΠςϯγͷѱԽෛՙͷภΓͳͲݟΒΕͳ͔ ͬͨ • ୠ͠ɺSR-IOVͷ߹ʹVF͕ཧίΞΛ͑ͨลΓͰ͔
ͳΓੑೳʹภΓ͕ൃੜͨ͠ • cgroupΛ༻͍ͯϚϧνϢʔβ࣌ʹϢʔβؒͷCPUׂΓͯΛ ެฏʹग़དྷΔࣄΛ֬ೝͨ͠ 1211݄20Ր༵