On Wed, 2021-06-30 at 10:11 +0500, Ammad Syed wrote:
- Is there any isolation done at the Kernel level of the compute OS?
No, There is no changes made in kernel.
- What does your flavor look like right now? Is the behavior different if you remove the numa constraint?
My flavor has currently below specs set.
dec2e31d-e1e8-4ff2-90d5-955be7e9efd3 | RC-16G | 16384 | 5 | 0 | 8 | True | 0 | 1.0 | hw:cpu_cores='2', hw:cpu_sockets='4' |
if these are the only extra you have added you have not enabled cpu pinning in the falvor
But I have set cpu pinning in compute node. [compute] cpu_shared_set = 2-7,10-15,18-23,26-31
you do not enable cpu pinning in the compute node. you can configure cores to be used for cpu pinning using cpu_dedicated_set but vms will not use those cores unless you request them the cpu_shared_set is used for unpinned vms. it defines the range of cpus over which the vms can float more coments below
If I remove hw:cpu_* from flavor and remove above config from nova.conf of compute. Instance deployment works fine.
You seem to have 4 NUMA as well, but only two physical sockets (8CPUx2threads - 16 vCPUs per socket x 2 = 32)
This is my test compute node have 4 cpu sockets and 4 numa nodes.
root@kvm10-a1-khi01:~# numactl -H available: 4 nodes (0-3) node 0 cpus: 0 1 2 3 4 5 6 7 node 0 size: 31675 MB node 0 free: 30135 MB node 1 cpus: 8 9 10 11 12 13 14 15 node 1 size: 64510 MB node 1 free: 63412 MB node 2 cpus: 16 17 18 19 20 21 22 23 node 2 size: 64510 MB node 2 free: 63255 MB node 3 cpus: 24 25 26 27 28 29 30 31 node 3 size: 64485 MB node 3 free: 63117 MB node distances: node 0 1 2 3 0: 10 16 16 16 1: 16 10 16 16 2: 16 16 10 16 3: 16 16 16 10
The generated XML seems to be set for a 4 socket topology?
Yes I am testing this on my test compute node first.
On Wed, Jun 30, 2021 at 7:15 AM Laurent Dumont <laurentfdumont@gmail.com> wrote:
- Is there any isolation done at the Kernel level of the compute OS? - What does your flavor look like right now? Is the behavior different if you remove the numa constraint?
You seem to have 4 NUMA as well, but only two physical sockets (8CPUx2threads - 16 vCPUs per socket x 2 = 32)
The generated XML seems to be set for a 4 socket topology?
<cpu mode='custom' match='exact' check='partial'> <model fallback='allow'>Opteron_G5</model> <topology sockets='4' cores='2' threads='1'/> </cpu>
On Tue, Jun 29, 2021 at 12:41 PM Ammad Syed <syedammad83@gmail.com> wrote:
Hi Stephen,
I have checked all cpus are online.
root@kvm10-a1-khi01:/etc/nova# lscpu Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian Address sizes: 48 bits physical, 48 bits virtual CPU(s): 32 On-line CPU(s) list: 0-31 Thread(s) per core: 2 Core(s) per socket: 8 Socket(s): 2 NUMA node(s): 4 Vendor ID: AuthenticAMD CPU family: 21
I have made below configuration in nova.conf.
[compute]
cpu_shared_set = 2-7,10-15,18-23,26-31
Below is the xml in nova logs that nova is trying to create domain.
2021-06-29 16:30:56.576 2819 ERROR nova.virt.libvirt.guest [req-c76c6809-1775-43a8-bfb1-70f6726cad9d 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] Error launching a defined domain with XML: <domain type='kvm'> <name>instance-0000026d</name> <uuid>06ff4fd5-b21f-4f64-9dde-55e86dd15da6</uuid> <metadata> <nova:instance xmlns:nova=" http://openstack.org/xmlns/libvirt/nova/1.1"> <nova:package version="23.0.0"/> <nova:name>cpu</nova:name> <nova:creationTime>2021-06-29 16:30:50</nova:creationTime> <nova:flavor name="RC-16G"> <nova:memory>16384</nova:memory> <nova:disk>5</nova:disk> <nova:swap>0</nova:swap> <nova:ephemeral>0</nova:ephemeral> <nova:vcpus>8</nova:vcpus> </nova:flavor> <nova:owner> <nova:user uuid="2af528fdf3244e15b4f3f8fcfc0889c5">admin</nova:user> <nova:project uuid="890eb2b7d1b8488aa88de7c34d08817a">admin</nova:project> </nova:owner> <nova:root type="image" uuid="1024317b-0db6-418c-bc39-de9b61d8ce59"/> <nova:ports> <nova:port uuid="dccf68a2-ec48-4985-94ce-b3487cfc99f3"> <nova:ip type="fixed" address="192.168.100.106" ipVersion="4"/> </nova:port> </nova:ports> </nova:instance> </metadata> <memory unit='KiB'>16777216</memory> <currentMemory unit='KiB'>16777216</currentMemory> <vcpu placement='static' cpuset='2-7,10-15,18-23,26-31'>8</vcpu>
here you can se we have copyied the cpu_share_set value into the set of cpus which this vm is allowed to float over. this will do soft pinning using cgroups but the guest cpu will still float over that range
<cputune> <shares>8192</shares> </cputune> <sysinfo type='smbios'> <system> <entry name='manufacturer'>OpenStack Foundation</entry> <entry name='product'>OpenStack Nova</entry> <entry name='version'>23.0.0</entry> <entry name='serial'>06ff4fd5-b21f-4f64-9dde-55e86dd15da6</entry> <entry name='uuid'>06ff4fd5-b21f-4f64-9dde-55e86dd15da6</entry> <entry name='family'>Virtual Machine</entry> </system> </sysinfo> <os> <type arch='x86_64' machine='pc-i440fx-4.2'>hvm</type> <boot dev='hd'/> <smbios mode='sysinfo'/> </os> <features> <acpi/> <apic/> </features> <cpu mode='custom' match='exact' check='partial'> <model fallback='allow'>Opteron_G5</model> <topology sockets='4' cores='2' threads='1'/> </cpu> <clock offset='utc'> <timer name='pit' tickpolicy='delay'/> <timer name='rtc' tickpolicy='catchup'/> <timer name='hpet' present='no'/> </clock> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/bin/qemu-system-x86_64</emulator> <disk type='file' device='disk'> <driver name='qemu' type='qcow2' cache='none'/> <source file='/var/lib/nova/instances/06ff4fd5-b21f-4f64-9dde-55e86dd15da6/disk'/> <target dev='sda' bus='scsi'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> <controller type='scsi' index='0' model='virtio-scsi'> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </controller> <controller type='usb' index='0' model='piix3-uhci'> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x2'/> </controller> <controller type='pci' index='0' model='pci-root'/> <interface type='bridge'> <mac address='fa:16:3e:f5:46:7d'/> <source bridge='br-int'/> <virtualport type='openvswitch'> <parameters interfaceid='dccf68a2-ec48-4985-94ce-b3487cfc99f3'/> </virtualport> <target dev='tapdccf68a2-ec'/> <model type='virtio'/> <driver name='vhost' queues='8'/> <mtu size='1442'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> <serial type='pty'> <log file='/var/lib/nova/instances/06ff4fd5-b21f-4f64-9dde-55e86dd15da6/console.log' append='off'/> <target type='isa-serial' port='0'> <model name='isa-serial'/> </target> </serial> <console type='pty'> <log file='/var/lib/nova/instances/06ff4fd5-b21f-4f64-9dde-55e86dd15da6/console.log' append='off'/> <target type='serial' port='0'/> </console> <input type='tablet' bus='usb'> <address type='usb' bus='0' port='1'/> </input> <input type='mouse' bus='ps2'/> <input type='keyboard' bus='ps2'/> <graphics type='spice' autoport='yes' listen='0.0.0.0'> <listen type='address' address='0.0.0.0'/> </graphics> <video> <model type='qxl' ram='65536' vram='65536' vgamem='16384' heads='1' primary='yes'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <memballoon model='virtio'> <stats period='10'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </memballoon> <rng model='virtio'> <backend model='random'>/dev/urandom</backend> <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/> </rng> </devices> </domain> : libvirt.libvirtError: Unable to write to '/sys/fs/cgroup/cpuset/machine.slice/machine-qemu\x2d1\x2dinstance\x2d0000026d.scope/emulator/cpuset.cpus': Permission denied
the xml below is for an unpinend guest and has no numa toplogy defiend. this error indicates a issue wiht the set of cgroups that libvirt can use. its either a limisation of your kernel or you have alstered the default set of cgroups in your libvirt config. this is contoled by the following in /etc/libvirt/qemu.conf # What cgroup controllers to make use of with QEMU guests # # - 'cpu' - use for scheduler tunables # - 'devices' - use for device whitelisting # - 'memory' - use for memory tunables # - 'blkio' - use for block devices I/O tunables # - 'cpuset' - use for CPUs and memory nodes # - 'cpuacct' - use for CPUs statistics. # # NB, even if configured here, they won't be used unless # the administrator has mounted cgroups, e.g.: # # mkdir /dev/cgroup # mount -t cgroup -o devices,cpu,memory,blkio,cpuset none /dev/cgroup # # They can be mounted anywhere, and different controllers # can be mounted in different locations. libvirt will detect # where they are located. # #cgroup_controllers = [ "cpu", "devices", "memory", "blkio", "cpuset", "cpuacct" ] the libvirt default which are commented out should be correct. with out cpuset you cant use any form of cpu pinning. so you should confirm that cpuset is included in the cgroup mount point i.e. mount -t cgroup -o devices,cpu,memory,blkio,cpuset none /dev/cgroup
Ammad
On Tue, Jun 29, 2021 at 8:29 PM Stephen Finucane <stephenfin@redhat.com> wrote:
Thanks,, the information is really helpful. I am have set below
On Tue, 2021-06-29 at 17:44 +0500, Ammad Syed wrote: properties to
flavor according to my numa policies.
--property hw:numa_nodes=FLAVOR-NODES \ --property hw:numa_cpus.N=FLAVOR-CORES \ --property hw:numa_mem.N=FLAVOR-MEMORY
you should only ever use hw:numa_cpus.N and hw:numa_mem.N iif you have an unblanced numa toplogy in the guest and 2+ numa nodes. where in is the numa node index starting form 0. if you are using them and you do not have a deep understaing of how that alters the vm xml generation and asignment to the host and you have not messured the performace the its almost alwasy a mistake to specify hw:numa_cpus or hw:numa_mem.
I am having below error in compute logs. Any advise.
libvirt.libvirtError: Unable to write to '/sys/fs/cgroup/cpuset/machine.slice/machine- qemu\x2d48\x2dinstance\x2d0000026b.scope/emulator/cpuset.cpus':
denied 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest Traceback (most recent call last): 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/nova/virt/libvirt/guest.py", line 155, in launch 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest return self._domain.createWithFlags(flags) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/eventlet/tpool.py", line 193, in doit 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest result = proxy_call(self._autowrap, f, *args, **kwargs) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/eventlet/tpool.py", line 151, in
2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest rv = execute(f, *args, **kwargs) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/eventlet/tpool.py", line 132, in execute 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest six.reraise(c, e, tb) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/six.py", line 703, in reraise 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest raise value 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/eventlet/tpool.py", line 86, in tworker 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest rv = meth(*args, **kwargs) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest File "/usr/lib/python3/dist-packages/libvirt.py", line 1265, in createWithFlags 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest if ret == -1: raise libvirtError ('virDomainCreateWithFlags() failed', dom=self) 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest libvirt.libvirtError: Unable to write to '/sys/fs/cgroup/cpuset/machine.slice/machine- qemu\x2d48\x2dinstance\x2d0000026b.scope/emulator/cpuset.cpus': Permission denied 2021-06-29 12:33:10.144 1310945 ERROR nova.virt.libvirt.guest 2021-06-29 12:33:10.146 1310945 ERROR nova.virt.libvirt.driver [req-4f6fc6aa- 04d6-4dc0-921f-2913b40a76a9 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] [instance: ed87bf68-b631- 4a00-9eb5-22d32ec37402] Failed to start libvirt guest:
Permission proxy_call libvirt.libvirtError:
Unable to write to '/sys/fs/cgroup/cpuset/machine.slice/machine- qemu\x2d48\x2dinstance\x2d0000026b.scope/emulator/cpuset.cpus': Permission denied 2021-06-29 12:33:10.150 1310945 INFO os_vif [req-4f6fc6aa-04d6-4dc0-921f- 2913b40a76a9 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] Successfully unplugged vif VIFOpenVSwitch(active=False,address=fa:16:3e:ba:3d:c8,bridge_name='br- int',has_traffic_filtering=True,id=a991cd33-2610-4823-a471- 62171037e1b5,network=Network(a0d85af2-a991-4102-8453-
lete=False,vif_name='tapa991cd33-26') 2021-06-29 12:33:10.151 1310945 INFO nova.virt.libvirt.driver [req-4f6fc6aa- 04d6-4dc0-921f-2913b40a76a9 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] [instance: ed87bf68-b631- 4a00-9eb5-22d32ec37402] Deleting instance files /var/lib/nova/instances/ed87bf68-b631-4a00-9eb5-22d32ec37402_del 2021-06-29 12:33:10.152 1310945 INFO nova.virt.libvirt.driver [req-4f6fc6aa- 04d6-4dc0-921f-2913b40a76a9 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] [instance: ed87bf68-b631- 4a00-9eb5-22d32ec37402] Deletion of /var/lib/nova/instances/ed87bf68-b631-4a00- 9eb5-22d32ec37402_del complete 2021-06-29 12:33:10.258 1310945 ERROR nova.compute.manager [req-4f6fc6aa-04d6- 4dc0-921f-2913b40a76a9 2af528fdf3244e15b4f3f8fcfc0889c5 890eb2b7d1b8488aa88de7c34d08817a - default default] [instance: ed87bf68-b631- 4a00-9eb5-22d32ec37402] Instance failed to spawn:
ba68c5e10b65),plugin='ovs',port_profile=VIFPortProfileOpenVSwitch,preserve_on_de libvirt.libvirtError: Unable
to write to '/sys/fs/cgroup/cpuset/machine.slice/machine- qemu\x2d48\x2dinstance\x2d0000026b.scope/emulator/cpuset.cpus': Permission denied
Any advise how to fix this permission issue ?
I have manually created the directory machine-qemu in /sys/fs/cgroup/cpuset/machine.slice/ but still having the same error.
I have also tried to set [compute] cpu_shared_set AND [compute] cpu_dedicated_set they are also giving the same error.
There are quite a few bugs about this [1][2]. It seems most of them are caused by CPUs being offlined. Have you offline CPUs? Are the CPUs listed in the mask all available?
Stephen
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1609785 [2] https://bugzilla.redhat.com/show_bug.cgi?id=1842716
Using ubuntu20.04 and qemu-kvm 4.2.
Ammad
On Fri, 2021-06-25 at 10:02 +0500, Ammad Syed wrote: > Hi, > > I am using openstack wallaby on ubuntu 20.04 and kvm. I am working to make > optimized flavor properties that should provide optimal
> reviewing the document below. > > https://docs.openstack.org/nova/wallaby/admin/cpu-topologies.html > > I have two socket AMD compute node. The workload running on nodes are mixed > workload. > > My question is should I use default nova CPU topology and NUMA node that > nova deploys instance by default OR should I use hw:cpu_sockets='2' > and hw:numa_nodes='2'. the latter hw:cpu_sockets='2' and hw:numa_nodes='2' should give you better performce however you should also set hw:mem_page_size=small or hw:mem_page_size=any when you enable virtual numa policies we afinities the guest memory to host numa nodes. This can lead to Out of memory evnet on the the host numa nodes which can result in vms being killed by the host kernel memeory reaper if you do not enable numa aware memeory trackign iin nova which is done by setting hw:mem_page_size. setting hw:mem_page_size has the side effect of of disabling memory over commit so you have to bare that in mind. if you are using numa toplogy you should almost always also use hugepages which are enabled using hw:mem_page_size=large this however requires you to configure hupgepages in the host at boot. > > Which one from above provide best instance performance ? or any other > tuning should I do ?
in the libvirt driver the default cpu toplogy we will genergated is 1 thread per core, 1 core per socket and 1 socket per flavor.vcpu. (technially this is an undocumeted implemation detail that you should not rely on, we have the hw:cpu_* element if you care about the toplogy)
this was more effincet in the early days of qemu/openstack but has may issue when software is chagne per sokcet or oepreating systems have a limit on socket supported such as windows.
generally i advies that you set hw:cpu_sockets to the typical number of sockets on the underlying host. simialrly if the flavor will only be run on host with SMT/hypertreading enabled on you shoudl set hw:cpu_threads=2
the flavor.vcpus must be devisable by the product of hw:cpu_sockets, hw:cpu_cores and hw:cpu_threads if they are set.
so if you have hw:cpu_threads=2 it must be devisable by 2 if you have hw:cpu_threads=2 and hw:cpu_sockets=2 flavor.vcpus must be a multiple of 4 > > The note in the URL (CPU topology sesion) suggests that I should stay with > default options that nova provides. in generaly no you should aling it to the host toplogy if you have similar toplogy across your data center. the default should always just work but its not nessisarly optimal and window sguest might not boot if you have too many sockets. windows 10 for exmple only supprot 2 socket so you could only have 2 flavor.vcpus if you used the default toplogy.
> > Currently it also works with libvirt/QEMU driver but we don’t recommend it > in production use cases. This is because vCPUs are actually running in one > thread on host in qemu TCG (Tiny Code Generator), which is the backend for > libvirt/QEMU driver. Work to enable full multi-threading support for TCG > (a.k.a. MTTCG) is on going in QEMU community. Please see this MTTCG project > <http://wiki.qemu.org/Features/tcg-multithread> page for detail. we do not gnerally recommende using qemu without kvm in produciton. the mttcg backend is useful in cases where you want to emulate other
On Fri, Jun 25, 2021 at 10:54 AM Sean Mooney <smooney@redhat.com> wrote: performance. I was plathform
but that usecsae is not currently supported in nova. for your deployment you should use libvirt with kvm and you should also consider if you want to support nested virtualisation or not. > > > Ammad
-- Regards,
Syed Ammad Ali