1 [[chapter_virtual_machines]]
10 qm - Qemu/KVM Virtual Machine Manager
16 include::qm.1-synopsis.adoc[]
22 Qemu/KVM Virtual Machines
23 =========================
28 // http://pve.proxmox.com/wiki/Container_and_Full_Virtualization
29 // http://pve.proxmox.com/wiki/KVM
30 // http://pve.proxmox.com/wiki/Qemu_Server
32 Qemu (short form for Quick Emulator) is an open source hypervisor that emulates a
33 physical computer. From the perspective of the host system where Qemu is
34 running, Qemu is a user program which has access to a number of local resources
35 like partitions, files, network cards which are then passed to an
36 emulated computer which sees them as if they were real devices.
38 A guest operating system running in the emulated computer accesses these
39 devices, and runs as it were running on real hardware. For instance you can pass
40 an iso image as a parameter to Qemu, and the OS running in the emulated computer
41 will see a real CDROM inserted in a CD drive.
43 Qemu can emulate a great variety of hardware from ARM to Sparc, but {pve} is
44 only concerned with 32 and 64 bits PC clone emulation, since it represents the
45 overwhelming majority of server hardware. The emulation of PC clones is also one
46 of the fastest due to the availability of processor extensions which greatly
47 speed up Qemu when the emulated architecture is the same as the host
50 NOTE: You may sometimes encounter the term _KVM_ (Kernel-based Virtual Machine).
51 It means that Qemu is running with the support of the virtualization processor
52 extensions, via the Linux kvm module. In the context of {pve} _Qemu_ and
53 _KVM_ can be used interchangeably as Qemu in {pve} will always try to load the kvm
56 Qemu inside {pve} runs as a root process, since this is required to access block
60 Emulated devices and paravirtualized devices
61 --------------------------------------------
63 The PC hardware emulated by Qemu includes a mainboard, network controllers,
64 scsi, ide and sata controllers, serial ports (the complete list can be seen in
65 the `kvm(1)` man page) all of them emulated in software. All these devices
66 are the exact software equivalent of existing hardware devices, and if the OS
67 running in the guest has the proper drivers it will use the devices as if it
68 were running on real hardware. This allows Qemu to runs _unmodified_ operating
71 This however has a performance cost, as running in software what was meant to
72 run in hardware involves a lot of extra work for the host CPU. To mitigate this,
73 Qemu can present to the guest operating system _paravirtualized devices_, where
74 the guest OS recognizes it is running inside Qemu and cooperates with the
77 Qemu relies on the virtio virtualization standard, and is thus able to present
78 paravirtualized virtio devices, which includes a paravirtualized generic disk
79 controller, a paravirtualized network card, a paravirtualized serial port,
80 a paravirtualized SCSI controller, etc ...
82 It is highly recommended to use the virtio devices whenever you can, as they
83 provide a big performance improvement. Using the virtio generic disk controller
84 versus an emulated IDE controller will double the sequential write throughput,
85 as measured with `bonnie++(8)`. Using the virtio network interface can deliver
86 up to three times the throughput of an emulated Intel E1000 network card, as
87 measured with `iperf(1)`. footnote:[See this benchmark on the KVM wiki
88 http://www.linux-kvm.org/page/Using_VirtIO_NIC]
91 [[qm_virtual_machines_settings]]
92 Virtual Machines Settings
93 -------------------------
95 Generally speaking {pve} tries to choose sane defaults for virtual machines
96 (VM). Make sure you understand the meaning of the settings you change, as it
97 could incur a performance slowdown, or putting your data at risk.
100 [[qm_general_settings]]
104 [thumbnail="gui-create-vm-general.png"]
106 General settings of a VM include
108 * the *Node* : the physical server on which the VM will run
109 * the *VM ID*: a unique number in this {pve} installation used to identify your VM
110 * *Name*: a free form text string you can use to describe the VM
111 * *Resource Pool*: a logical group of VMs
118 [thumbnail="gui-create-vm-os.png"]
120 When creating a VM, setting the proper Operating System(OS) allows {pve} to
121 optimize some low level parameters. For instance Windows OS expect the BIOS
122 clock to use the local time, while Unix based OS expect the BIOS clock to have
130 Qemu can emulate a number of storage controllers:
132 * the *IDE* controller, has a design which goes back to the 1984 PC/AT disk
133 controller. Even if this controller has been superseded by recent designs,
134 each and every OS you can think of has support for it, making it a great choice
135 if you want to run an OS released before 2003. You can connect up to 4 devices
138 * the *SATA* (Serial ATA) controller, dating from 2003, has a more modern
139 design, allowing higher throughput and a greater number of devices to be
140 connected. You can connect up to 6 devices on this controller.
142 * the *SCSI* controller, designed in 1985, is commonly found on server grade
143 hardware, and can connect up to 14 storage devices. {pve} emulates by default a
144 LSI 53C895A controller.
146 A SCSI controller of type _VirtIO SCSI_ is the recommended setting if you aim for
147 performance and is automatically selected for newly created Linux VMs since
148 {pve} 4.3. Linux distributions have support for this controller since 2012, and
149 FreeBSD since 2014. For Windows OSes, you need to provide an extra iso
150 containing the drivers during the installation.
151 // https://pve.proxmox.com/wiki/Paravirtualized_Block_Drivers_for_Windows#During_windows_installation.
152 If you aim at maximum performance, you can select a SCSI controller of type
153 _VirtIO SCSI single_ which will allow you to select the *IO Thread* option.
154 When selecting _VirtIO SCSI single_ Qemu will create a new controller for
155 each disk, instead of adding all disks to the same controller.
157 * The *VirtIO Block* controller, often just called VirtIO or virtio-blk,
158 is an older type of paravirtualized controller. It has been superseded by the
159 VirtIO SCSI Controller, in terms of features.
161 [thumbnail="gui-create-vm-hard-disk.png"]
162 On each controller you attach a number of emulated hard disks, which are backed
163 by a file or a block device residing in the configured storage. The choice of
164 a storage type will determine the format of the hard disk image. Storages which
165 present block devices (LVM, ZFS, Ceph) will require the *raw disk image format*,
166 whereas files based storages (Ext4, NFS, CIFS, GlusterFS) will let you to choose
167 either the *raw disk image format* or the *QEMU image format*.
169 * the *QEMU image format* is a copy on write format which allows snapshots, and
170 thin provisioning of the disk image.
171 * the *raw disk image* is a bit-to-bit image of a hard disk, similar to what
172 you would get when executing the `dd` command on a block device in Linux. This
173 format does not support thin provisioning or snapshots by itself, requiring
174 cooperation from the storage layer for these tasks. It may, however, be up to
175 10% faster than the *QEMU image format*. footnote:[See this benchmark for details
176 http://events.linuxfoundation.org/sites/events/files/slides/CloudOpen2013_Khoa_Huynh_v3.pdf]
177 * the *VMware image format* only makes sense if you intend to import/export the
178 disk image to other hypervisors.
180 Setting the *Cache* mode of the hard drive will impact how the host system will
181 notify the guest systems of block write completions. The *No cache* default
182 means that the guest system will be notified that a write is complete when each
183 block reaches the physical storage write queue, ignoring the host page cache.
184 This provides a good balance between safety and speed.
186 If you want the {pve} backup manager to skip a disk when doing a backup of a VM,
187 you can set the *No backup* option on that disk.
189 If you want the {pve} storage replication mechanism to skip a disk when starting
190 a replication job, you can set the *Skip replication* option on that disk.
191 As of {pve} 5.0, replication requires the disk images to be on a storage of type
192 `zfspool`, so adding a disk image to other storages when the VM has replication
193 configured requires to skip replication for this disk image.
195 If your storage supports _thin provisioning_ (see the storage chapter in the
196 {pve} guide), and your VM has a *SCSI* controller you can activate the *Discard*
197 option on the hard disks connected to that controller. With *Discard* enabled,
198 when the filesystem of a VM marks blocks as unused after removing files, the
199 emulated SCSI controller will relay this information to the storage, which will
200 then shrink the disk image accordingly.
203 The option *IO Thread* can only be used when using a disk with the
204 *VirtIO* controller, or with the *SCSI* controller, when the emulated controller
205 type is *VirtIO SCSI single*.
206 With this enabled, Qemu creates one I/O thread per storage controller,
207 instead of a single thread for all I/O, so it increases performance when
208 multiple disks are used and each disk has its own storage controller.
209 Note that backups do not currently work with *IO Thread* enabled.
216 [thumbnail="gui-create-vm-cpu.png"]
218 A *CPU socket* is a physical slot on a PC motherboard where you can plug a CPU.
219 This CPU can then contain one or many *cores*, which are independent
220 processing units. Whether you have a single CPU socket with 4 cores, or two CPU
221 sockets with two cores is mostly irrelevant from a performance point of view.
222 However some software licenses depend on the number of sockets a machine has,
223 in that case it makes sense to set the number of sockets to what the license
226 Increasing the number of virtual cpus (cores and sockets) will usually provide a
227 performance improvement though that is heavily dependent on the use of the VM.
228 Multithreaded applications will of course benefit from a large number of
229 virtual cpus, as for each virtual cpu you add, Qemu will create a new thread of
230 execution on the host system. If you're not sure about the workload of your VM,
231 it is usually a safe bet to set the number of *Total cores* to 2.
233 NOTE: It is perfectly safe if the _overall_ number of cores of all your VMs
234 is greater than the number of cores on the server (e.g., 4 VMs with each 4
235 cores on a machine with only 8 cores). In that case the host system will
236 balance the Qemu execution threads between your server cores, just like if you
237 were running a standard multithreaded application. However, {pve} will prevent
238 you from assigning more virtual CPU cores than physically available, as this will
239 only bring the performance down due to the cost of context switches.
241 [[qm_cpu_resource_limits]]
245 In addition to the number of virtual cores, you can configure how much resources
246 a VM can get in relation to the host CPU time and also in relation to other
248 With the *cpulimit* (``Host CPU Time'') option you can limit how much CPU time
249 the whole VM can use on the host. It is a floating point value representing CPU
250 time in percent, so `1.0` is equal to `100%`, `2.5` to `250%` and so on. If a
251 single process would fully use one single core it would have `100%` CPU Time
252 usage. If a VM with four cores utilizes all its cores fully it would
253 theoretically use `400%`. In reality the usage may be even a bit higher as Qemu
254 can have additional threads for VM peripherals besides the vCPU core ones.
255 This setting can be useful if a VM should have multiple vCPUs, as it runs a few
256 processes in parallel, but the VM as a whole should not be able to run all
257 vCPUs at 100% at the same time. Using a specific example: lets say we have a VM
258 which would profit from having 8 vCPUs, but at no time all of those 8 cores
259 should run at full load - as this would make the server so overloaded that
260 other VMs and CTs would get to less CPU. So, we set the *cpulimit* limit to
261 `4.0` (=400%). If all cores do the same heavy work they would all get 50% of a
262 real host cores CPU time. But, if only 4 would do work they could still get
263 almost 100% of a real core each.
265 NOTE: VMs can, depending on their configuration, use additional threads e.g.,
266 for networking or IO operations but also live migration. Thus a VM can show up
267 to use more CPU time than just its virtual CPUs could use. To ensure that a VM
268 never uses more CPU time than virtual CPUs assigned set the *cpulimit* setting
269 to the same value as the total core count.
271 The second CPU resource limiting setting, *cpuunits* (nowadays often called CPU
272 shares or CPU weight), controls how much CPU time a VM gets in regards to other
273 VMs running. It is a relative weight which defaults to `1024`, if you increase
274 this for a VM it will be prioritized by the scheduler in comparison to other
275 VMs with lower weight. E.g., if VM 100 has set the default 1024 and VM 200 was
276 changed to `2048`, the latter VM 200 would receive twice the CPU bandwidth than
279 For more information see `man systemd.resource-control`, here `CPUQuota`
280 corresponds to `cpulimit` and `CPUShares` corresponds to our `cpuunits`
281 setting, visit its Notes section for references and implementation details.
286 Qemu can emulate a number different of *CPU types* from 486 to the latest Xeon
287 processors. Each new processor generation adds new features, like hardware
288 assisted 3d rendering, random number generation, memory protection, etc ...
289 Usually you should select for your VM a processor type which closely matches the
290 CPU of the host system, as it means that the host CPU features (also called _CPU
291 flags_ ) will be available in your VMs. If you want an exact match, you can set
292 the CPU type to *host* in which case the VM will have exactly the same CPU flags
295 This has a downside though. If you want to do a live migration of VMs between
296 different hosts, your VM might end up on a new system with a different CPU type.
297 If the CPU flags passed to the guest are missing, the qemu process will stop. To
298 remedy this Qemu has also its own CPU type *kvm64*, that {pve} uses by defaults.
299 kvm64 is a Pentium 4 look a like CPU type, which has a reduced CPU flags set,
300 but is guaranteed to work everywhere.
302 In short, if you care about live migration and moving VMs between nodes, leave
303 the kvm64 default. If you don’t care about live migration or have a homogeneous
304 cluster where all nodes have the same CPU, set the CPU type to host, as in
305 theory this will give your guests maximum performance.
307 Meltdown / Spectre related CPU flags
308 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
310 There are two CPU flags related to the Meltdown and Spectre vulnerabilities
311 footnote:[Meltdown Attack https://meltdownattack.com/] which need to be set
312 manually unless the selected CPU type of your VM already enables them by default.
314 The first, called 'pcid', helps to reduce the performance impact of the Meltdown
315 mitigation called 'Kernel Page-Table Isolation (KPTI)', which effectively hides
316 the Kernel memory from the user space. Without PCID, KPTI is quite an expensive
317 mechanism footnote:[PCID is now a critical performance/security feature on x86
318 https://groups.google.com/forum/m/#!topic/mechanical-sympathy/L9mHTbeQLNU].
320 The second CPU flag is called 'spec-ctrl', which allows an operating system to
321 selectively disable or restrict speculative execution in order to limit the
322 ability of attackers to exploit the Spectre vulnerability.
324 There are two requirements that need to be fulfilled in order to use these two
327 * The host CPU(s) must support the feature and propagate it to the guest's virtual CPU(s)
328 * The guest operating system must be updated to a version which mitigates the
329 attacks and is able to utilize the CPU feature
331 In order to use 'spec-ctrl', your CPU or system vendor also needs to provide a
332 so-called ``microcode update'' footnote:[You can use `intel-microcode' /
333 `amd-microcode' from Debian non-free if your vendor does not provide such an
334 update. Note that not all affected CPUs can be updated to support spec-ctrl.]
337 To check if the {pve} host supports PCID, execute the following command as root:
340 # grep ' pcid ' /proc/cpuinfo
343 If this does not return empty your host's CPU has support for 'pcid'.
345 To check if the {pve} host supports spec-ctrl, execute the following command as root:
348 # grep ' spec_ctrl ' /proc/cpuinfo
351 If this does not return empty your host's CPU has support for 'spec-ctrl'.
353 If you use `host' or another CPU type which enables the desired flags by
354 default, and you updated your guest OS to make use of the associated CPU
355 features, you're already set.
357 Otherwise you need to set the desired CPU flag of the virtual CPU, either by
358 editing the CPU options in the WebUI, or by setting the 'flags' property of the
359 'cpu' option in the VM configuration file.
363 You can also optionally emulate a *NUMA*
364 footnote:[https://en.wikipedia.org/wiki/Non-uniform_memory_access] architecture
365 in your VMs. The basics of the NUMA architecture mean that instead of having a
366 global memory pool available to all your cores, the memory is spread into local
367 banks close to each socket.
368 This can bring speed improvements as the memory bus is not a bottleneck
369 anymore. If your system has a NUMA architecture footnote:[if the command
370 `numactl --hardware | grep available` returns more than one node, then your host
371 system has a NUMA architecture] we recommend to activate the option, as this
372 will allow proper distribution of the VM resources on the host system.
373 This option is also required to hot-plug cores or RAM in a VM.
375 If the NUMA option is used, it is recommended to set the number of sockets to
376 the number of sockets of the host system.
381 Modern operating systems introduced the capability to hot-plug and, to a
382 certain extent, hot-unplug CPUs in a running systems. Virtualisation allows us
383 to avoid a lot of the (physical) problems real hardware can cause in such
385 Still, this is a rather new and complicated feature, so its use should be
386 restricted to cases where its absolutely needed. Most of the functionality can
387 be replicated with other, well tested and less complicated, features, see
388 xref:qm_cpu_resource_limits[Resource Limits].
390 In {pve} the maximal number of plugged CPUs is always `cores * sockets`.
391 To start a VM with less than this total core count of CPUs you may use the
392 *vpus* setting, it denotes how many vCPUs should be plugged in at VM start.
394 Currently only this feature is only supported on Linux, a kernel newer than 3.10
395 is needed, a kernel newer than 4.7 is recommended.
397 You can use a udev rule as follow to automatically set new CPUs as online in
401 SUBSYSTEM=="cpu", ACTION=="add", TEST=="online", ATTR{online}=="0", ATTR{online}="1"
404 Save this under /etc/udev/rules.d/ as a file ending in `.rules`.
406 Note: CPU hot-remove is machine dependent and requires guest cooperation.
407 The deletion command does not guarantee CPU removal to actually happen,
408 typically it's a request forwarded to guest using target dependent mechanism,
409 e.g., ACPI on x86/amd64.
416 For each VM you have the option to set a fixed size memory or asking
417 {pve} to dynamically allocate memory based on the current RAM usage of the
420 .Fixed Memory Allocation
421 [thumbnail="gui-create-vm-memory-fixed.png"]
423 When setting memory and minimum memory to the same amount
424 {pve} will simply allocate what you specify to your VM.
426 Even when using a fixed memory size, the ballooning device gets added to the
427 VM, because it delivers useful information such as how much memory the guest
429 In general, you should leave *ballooning* enabled, but if you want to disable
430 it (e.g. for debugging purposes), simply uncheck
431 *Ballooning Device* or set
435 in the configuration.
437 .Automatic Memory Allocation
438 [thumbnail="gui-create-vm-memory-dynamic.png", float="left"]
440 // see autoballoon() in pvestatd.pm
441 When setting the minimum memory lower that the memory, {pve} will make sure that the
442 minimum amount you specified is always available to the VM, and if RAM usage on
443 the host is below 80%, will dynamically add memory to the guest up to the
444 maximum memory specified.
446 When the host is becoming short on RAM, the VM will then release some memory
447 back to the host, swapping running processes if needed and starting the oom
448 killer in last resort. The passing around of memory between host and guest is
449 done via a special `balloon` kernel driver running inside the guest, which will
450 grab or release memory pages from the host.
451 footnote:[A good explanation of the inner workings of the balloon driver can be found here https://rwmj.wordpress.com/2010/07/17/virtio-balloon/]
453 When multiple VMs use the autoallocate facility, it is possible to set a
454 *Shares* coefficient which indicates the relative amount of the free host memory
455 that each VM should take. Suppose for instance you have four VMs, three of them
456 running a HTTP server and the last one is a database server. To cache more
457 database blocks in the database server RAM, you would like to prioritize the
458 database VM when spare RAM is available. For this you assign a Shares property
459 of 3000 to the database VM, leaving the other VMs to the Shares default setting
460 of 1000. The host server has 32GB of RAM, and is currently using 16GB, leaving 32
461 * 80/100 - 16 = 9GB RAM to be allocated to the VMs. The database VM will get 9 *
462 3000 / (3000 + 1000 + 1000 + 1000) = 4.5 GB extra RAM and each HTTP server will
465 All Linux distributions released after 2010 have the balloon kernel driver
466 included. For Windows OSes, the balloon driver needs to be added manually and can
467 incur a slowdown of the guest, so we don't recommend using it on critical
469 // see https://forum.proxmox.com/threads/solved-hyper-threading-vs-no-hyper-threading-fixed-vs-variable-memory.20265/
471 When allocating RAM to your VMs, a good rule of thumb is always to leave 1GB
472 of RAM available to the host.
475 [[qm_network_device]]
479 [thumbnail="gui-create-vm-network.png"]
481 Each VM can have many _Network interface controllers_ (NIC), of four different
484 * *Intel E1000* is the default, and emulates an Intel Gigabit network card.
485 * the *VirtIO* paravirtualized NIC should be used if you aim for maximum
486 performance. Like all VirtIO devices, the guest OS should have the proper driver
488 * the *Realtek 8139* emulates an older 100 MB/s network card, and should
489 only be used when emulating older operating systems ( released before 2002 )
490 * the *vmxnet3* is another paravirtualized device, which should only be used
491 when importing a VM from another hypervisor.
493 {pve} will generate for each NIC a random *MAC address*, so that your VM is
494 addressable on Ethernet networks.
496 The NIC you added to the VM can follow one of two different models:
498 * in the default *Bridged mode* each virtual NIC is backed on the host by a
499 _tap device_, ( a software loopback device simulating an Ethernet NIC ). This
500 tap device is added to a bridge, by default vmbr0 in {pve}. In this mode, VMs
501 have direct access to the Ethernet LAN on which the host is located.
502 * in the alternative *NAT mode*, each virtual NIC will only communicate with
503 the Qemu user networking stack, where a built-in router and DHCP server can
504 provide network access. This built-in DHCP will serve addresses in the private
505 10.0.2.0/24 range. The NAT mode is much slower than the bridged mode, and
506 should only be used for testing. This mode is only available via CLI or the API,
507 but not via the WebUI.
509 You can also skip adding a network device when creating a VM by selecting *No
513 If you are using the VirtIO driver, you can optionally activate the
514 *Multiqueue* option. This option allows the guest OS to process networking
515 packets using multiple virtual CPUs, providing an increase in the total number
516 of packets transferred.
518 //http://blog.vmsplice.net/2011/09/qemu-internals-vhost-architecture.html
519 When using the VirtIO driver with {pve}, each NIC network queue is passed to the
520 host kernel, where the queue will be processed by a kernel thread spawn by the
521 vhost driver. With this option activated, it is possible to pass _multiple_
522 network queues to the host kernel for each NIC.
524 //https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Tuning_and_Optimization_Guide/sect-Virtualization_Tuning_Optimization_Guide-Networking-Techniques.html#sect-Virtualization_Tuning_Optimization_Guide-Networking-Multi-queue_virtio-net
525 When using Multiqueue, it is recommended to set it to a value equal
526 to the number of Total Cores of your guest. You also need to set in
527 the VM the number of multi-purpose channels on each VirtIO NIC with the ethtool
530 `ethtool -L ens1 combined X`
532 where X is the number of the number of vcpus of the VM.
534 You should note that setting the Multiqueue parameter to a value greater
535 than one will increase the CPU load on the host and guest systems as the
536 traffic increases. We recommend to set this option only when the VM has to
537 process a great number of incoming connections, such as when the VM is running
538 as a router, reverse proxy or a busy HTTP server doing long polling.
541 [[qm_usb_passthrough]]
545 There are two different types of USB passthrough devices:
547 * Host USB passthrough
548 * SPICE USB passthrough
550 Host USB passthrough works by giving a VM a USB device of the host.
551 This can either be done via the vendor- and product-id, or
552 via the host bus and port.
554 The vendor/product-id looks like this: *0123:abcd*,
555 where *0123* is the id of the vendor, and *abcd* is the id
556 of the product, meaning two pieces of the same usb device
559 The bus/port looks like this: *1-2.3.4*, where *1* is the bus
560 and *2.3.4* is the port path. This represents the physical
561 ports of your host (depending of the internal order of the
564 If a device is present in a VM configuration when the VM starts up,
565 but the device is not present in the host, the VM can boot without problems.
566 As soon as the device/port is available in the host, it gets passed through.
568 WARNING: Using this kind of USB passthrough means that you cannot move
569 a VM online to another host, since the hardware is only available
570 on the host the VM is currently residing.
572 The second type of passthrough is SPICE USB passthrough. This is useful
573 if you use a SPICE client which supports it. If you add a SPICE USB port
574 to your VM, you can passthrough a USB device from where your SPICE client is,
575 directly to the VM (for example an input device or hardware dongle).
582 In order to properly emulate a computer, QEMU needs to use a firmware.
583 By default QEMU uses *SeaBIOS* for this, which is an open-source, x86 BIOS
584 implementation. SeaBIOS is a good choice for most standard setups.
586 There are, however, some scenarios in which a BIOS is not a good firmware
587 to boot from, e.g. if you want to do VGA passthrough. footnote:[Alex Williamson has a very good blog entry about this.
588 http://vfio.blogspot.co.at/2014/08/primary-graphics-assignment-without-vga.html]
589 In such cases, you should rather use *OVMF*, which is an open-source UEFI implementation. footnote:[See the OVMF Project http://www.tianocore.org/ovmf/]
591 If you want to use OVMF, there are several things to consider:
593 In order to save things like the *boot order*, there needs to be an EFI Disk.
594 This disk will be included in backups and snapshots, and there can only be one.
596 You can create such a disk with the following command:
598 qm set <vmid> -efidisk0 <storage>:1,format=<format>
600 Where *<storage>* is the storage where you want to have the disk, and
601 *<format>* is a format which the storage supports. Alternatively, you can
602 create such a disk through the web interface with 'Add' -> 'EFI Disk' in the
603 hardware section of a VM.
605 When using OVMF with a virtual display (without VGA passthrough),
606 you need to set the client resolution in the OVMF menu(which you can reach
607 with a press of the ESC button during boot), or you have to choose
608 SPICE as the display type.
610 [[qm_startup_and_shutdown]]
611 Automatic Start and Shutdown of Virtual Machines
612 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
614 After creating your VMs, you probably want them to start automatically
615 when the host system boots. For this you need to select the option 'Start at
616 boot' from the 'Options' Tab of your VM in the web interface, or set it with
617 the following command:
619 qm set <vmid> -onboot 1
621 .Start and Shutdown Order
623 [thumbnail="gui-qemu-edit-start-order.png"]
625 In some case you want to be able to fine tune the boot order of your
626 VMs, for instance if one of your VM is providing firewalling or DHCP
627 to other guest systems. For this you can use the following
630 * *Start/Shutdown order*: Defines the start order priority. E.g. set it to 1 if
631 you want the VM to be the first to be started. (We use the reverse startup
632 order for shutdown, so a machine with a start order of 1 would be the last to
633 be shut down). If multiple VMs have the same order defined on a host, they will
634 additionally be ordered by 'VMID' in ascending order.
635 * *Startup delay*: Defines the interval between this VM start and subsequent
636 VMs starts . E.g. set it to 240 if you want to wait 240 seconds before starting
638 * *Shutdown timeout*: Defines the duration in seconds {pve} should wait
639 for the VM to be offline after issuing a shutdown command.
640 By default this value is set to 180, which means that {pve} will issue a
641 shutdown request and wait 180 seconds for the machine to be offline. If
642 the machine is still online after the timeout it will be stopped forcefully.
644 NOTE: VMs managed by the HA stack do not follow the 'start on boot' and
645 'boot order' options currently. Those VMs will be skipped by the startup and
646 shutdown algorithm as the HA manager itself ensures that VMs get started and
649 Please note that machines without a Start/Shutdown order parameter will always
650 start after those where the parameter is set. Further, this parameter can only
651 be enforced between virtual machines running on the same host, not
659 [thumbnail="gui-qemu-migrate.png"]
661 If you have a cluster, you can migrate your VM to another host with
663 qm migrate <vmid> <target>
665 There are generally two mechanisms for this
667 * Online Migration (aka Live Migration)
673 When your VM is running and it has no local resources defined (such as disks
674 on local storage, passed through devices, etc.) you can initiate a live
675 migration with the -online flag.
680 This starts a Qemu Process on the target host with the 'incoming' flag, which
681 means that the process starts and waits for the memory data and device states
682 from the source Virtual Machine (since all other resources, e.g. disks,
683 are shared, the memory content and device state are the only things left
686 Once this connection is established, the source begins to send the memory
687 content asynchronously to the target. If the memory on the source changes,
688 those sections are marked dirty and there will be another pass of sending data.
689 This happens until the amount of data to send is so small that it can
690 pause the VM on the source, send the remaining data to the target and start
691 the VM on the target in under a second.
696 For Live Migration to work, there are some things required:
698 * The VM has no local resources (e.g. passed through devices, local disks, etc.)
699 * The hosts are in the same {pve} cluster.
700 * The hosts have a working (and reliable) network connection.
701 * The target host must have the same or higher versions of the
702 {pve} packages. (It *might* work the other way, but this is never guaranteed)
707 If you have local resources, you can still offline migrate your VMs,
708 as long as all disk are on storages, which are defined on both hosts.
709 Then the migration will copy the disk over the network to the target host.
711 [[qm_copy_and_clone]]
715 [thumbnail="gui-qemu-full-clone.png"]
717 VM installation is usually done using an installation media (CD-ROM)
718 from the operation system vendor. Depending on the OS, this can be a
719 time consuming task one might want to avoid.
721 An easy way to deploy many VMs of the same type is to copy an existing
722 VM. We use the term 'clone' for such copies, and distinguish between
723 'linked' and 'full' clones.
727 The result of such copy is an independent VM. The
728 new VM does not share any storage resources with the original.
731 It is possible to select a *Target Storage*, so one can use this to
732 migrate a VM to a totally different storage. You can also change the
733 disk image *Format* if the storage driver supports several formats.
736 NOTE: A full clone need to read and copy all VM image data. This is
737 usually much slower than creating a linked clone.
740 Some storage types allows to copy a specific *Snapshot*, which
741 defaults to the 'current' VM data. This also means that the final copy
742 never includes any additional snapshots from the original VM.
747 Modern storage drivers supports a way to generate fast linked
748 clones. Such a clone is a writable copy whose initial contents are the
749 same as the original data. Creating a linked clone is nearly
750 instantaneous, and initially consumes no additional space.
753 They are called 'linked' because the new image still refers to the
754 original. Unmodified data blocks are read from the original image, but
755 modification are written (and afterwards read) from a new
756 location. This technique is called 'Copy-on-write'.
759 This requires that the original volume is read-only. With {pve} one
760 can convert any VM into a read-only <<qm_templates, Template>>). Such
761 templates can later be used to create linked clones efficiently.
764 NOTE: You cannot delete the original template while linked clones
768 It is not possible to change the *Target storage* for linked clones,
769 because this is a storage internal feature.
772 The *Target node* option allows you to create the new VM on a
773 different node. The only restriction is that the VM is on shared
774 storage, and that storage is also available on the target node.
776 To avoid resource conflicts, all network interface MAC addresses gets
777 randomized, and we generate a new 'UUID' for the VM BIOS (smbios1)
782 Virtual Machine Templates
783 -------------------------
785 One can convert a VM into a Template. Such templates are read-only,
786 and you can use them to create linked clones.
788 NOTE: It is not possible to start templates, because this would modify
789 the disk images. If you want to change the template, create a linked
790 clone and modify that.
792 Importing Virtual Machines and disk images
793 ------------------------------------------
795 A VM export from a foreign hypervisor takes usually the form of one or more disk
796 images, with a configuration file describing the settings of the VM (RAM,
798 The disk images can be in the vmdk format, if the disks come from
799 VMware or VirtualBox, or qcow2 if the disks come from a KVM hypervisor.
800 The most popular configuration format for VM exports is the OVF standard, but in
801 practice interoperation is limited because many settings are not implemented in
802 the standard itself, and hypervisors export the supplementary information
803 in non-standard extensions.
805 Besides the problem of format, importing disk images from other hypervisors
806 may fail if the emulated hardware changes too much from one hypervisor to
807 another. Windows VMs are particularly concerned by this, as the OS is very
808 picky about any changes of hardware. This problem may be solved by
809 installing the MergeIDE.zip utility available from the Internet before exporting
810 and choosing a hard disk type of *IDE* before booting the imported Windows VM.
812 Finally there is the question of paravirtualized drivers, which improve the
813 speed of the emulated system and are specific to the hypervisor.
814 GNU/Linux and other free Unix OSes have all the necessary drivers installed by
815 default and you can switch to the paravirtualized drivers right after importing
816 the VM. For Windows VMs, you need to install the Windows paravirtualized
819 GNU/Linux and other free Unix can usually be imported without hassle. Note
820 that we cannot guarantee a successful import/export of Windows VMs in all
821 cases due to the problems above.
823 Step-by-step example of a Windows OVF import
824 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
827 https://developer.microsoft.com/en-us/windows/downloads/virtual-machines/[Virtual Machines downloads]
828 to get started with Windows development.We are going to use one of these
829 to demonstrate the OVF import feature.
831 Download the Virtual Machine zip
832 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
834 After getting informed about the user agreement, choose the _Windows 10
835 Enterprise (Evaluation - Build)_ for the VMware platform, and download the zip.
837 Extract the disk image from the zip
838 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
840 Using the `unzip` utility or any archiver of your choice, unpack the zip,
841 and copy via ssh/scp the ovf and vmdk files to your {pve} host.
843 Import the Virtual Machine
844 ^^^^^^^^^^^^^^^^^^^^^^^^^^
846 This will create a new virtual machine, using cores, memory and
847 VM name as read from the OVF manifest, and import the disks to the +local-lvm+
848 storage. You have to configure the network manually.
850 qm importovf 999 WinDev1709Eval.ovf local-lvm
852 The VM is ready to be started.
854 Adding an external disk image to a Virtual Machine
855 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
857 You can also add an existing disk image to a VM, either coming from a
858 foreign hypervisor, or one that you created yourself.
860 Suppose you created a Debian/Ubuntu disk image with the 'vmdebootstrap' tool:
862 vmdebootstrap --verbose \
863 --size 10GiB --serial-console \
864 --grub --no-extlinux \
865 --package openssh-server \
866 --package avahi-daemon \
867 --package qemu-guest-agent \
868 --hostname vm600 --enable-dhcp \
869 --customize=./copy_pub_ssh.sh \
870 --sparse --image vm600.raw
872 You can now create a new target VM for this image.
874 qm create 600 --net0 virtio,bridge=vmbr0 --name vm600 --serial0 socket \
875 --bootdisk scsi0 --scsihw virtio-scsi-pci --ostype l26
877 Add the disk image as +unused0+ to the VM, using the storage +pvedir+:
879 qm importdisk 600 vm600.raw pvedir
881 Finally attach the unused disk to the SCSI controller of the VM:
883 qm set 600 --scsi0 pvedir:600/vm-600-disk-1.raw
885 The VM is ready to be started.
888 include::qm-cloud-init.adoc[]
890 Managing Virtual Machines with `qm`
891 ------------------------------------
893 qm is the tool to manage Qemu/Kvm virtual machines on {pve}. You can
894 create and destroy virtual machines, and control execution
895 (start/stop/suspend/resume). Besides that, you can use qm to set
896 parameters in the associated config file. It is also possible to
897 create and delete virtual disks.
902 Using an iso file uploaded on the 'local' storage, create a VM
903 with a 4 GB IDE disk on the 'local-lvm' storage
905 qm create 300 -ide0 local-lvm:4 -net0 e1000 -cdrom local:iso/proxmox-mailgateway_2.1.iso
911 Send a shutdown request, then wait until the VM is stopped.
913 qm shutdown 300 && qm wait 300
915 Same as above, but only wait for 40 seconds.
917 qm shutdown 300 && qm wait 300 -timeout 40
924 VM configuration files are stored inside the Proxmox cluster file
925 system, and can be accessed at `/etc/pve/qemu-server/<VMID>.conf`.
926 Like other files stored inside `/etc/pve/`, they get automatically
927 replicated to all other cluster nodes.
929 NOTE: VMIDs < 100 are reserved for internal purposes, and VMIDs need to be
932 .Example VM Configuration
940 net0: e1000=EE:D2:28:5F:B6:3E,bridge=vmbr0
941 virtio0: local:vm-100-disk-1,size=32G
944 Those configuration files are simple text files, and you can edit them
945 using a normal text editor (`vi`, `nano`, ...). This is sometimes
946 useful to do small corrections, but keep in mind that you need to
947 restart the VM to apply such changes.
949 For that reason, it is usually better to use the `qm` command to
950 generate and modify those files, or do the whole thing using the GUI.
951 Our toolkit is smart enough to instantaneously apply most changes to
952 running VM. This feature is called "hot plug", and there is no
953 need to restart the VM in that case.
959 VM configuration files use a simple colon separated key/value
960 format. Each line has the following format:
967 Blank lines in those files are ignored, and lines starting with a `#`
968 character are treated as comments and are also ignored.
975 When you create a snapshot, `qm` stores the configuration at snapshot
976 time into a separate snapshot section within the same configuration
977 file. For example, after creating a snapshot called ``testsnapshot'',
978 your configuration file will look like this:
980 .VM configuration with snapshot
994 There are a few snapshot related properties like `parent` and
995 `snaptime`. The `parent` property is used to store the parent/child
996 relationship between snapshots. `snaptime` is the snapshot creation
997 time stamp (Unix epoch).
1004 include::qm.conf.5-opts.adoc[]
1010 Online migrations, snapshots and backups (`vzdump`) set a lock to
1011 prevent incompatible concurrent actions on the affected VMs. Sometimes
1012 you need to remove such a lock manually (e.g., after a power failure).
1016 CAUTION: Only do that if you are sure the action which set the lock is
1025 `/etc/pve/qemu-server/<VMID>.conf`::
1027 Configuration file for the VM '<VMID>'.
1030 include::pve-copyright.adoc[]