upvm aims to make it crazy easy to spin up new virtual machines using pre-installed OS-images provided by existing virt-builder repos. Once an image is cached locally, the process of image creation + and guest instantiation should take less than a minute. Take a look:
upvm supervises virt-builder and virt-install commands that connect to your local libvirt daemon. Virtual machine images will be saved (by default) to /var/lib/upvm and virtual machines will be started on your system.
upvm might not be for you if:
- you've already invested the time to get vagrant to work for you
- you already have all the images you need in a RHEV environment
- you want to spin up new machines in an OpenStack cloud
I used to use LVM for all my VMs too ... so upvm can use virt-builder to generate an image and then write that image out to a block device you specify with the --primary-blockdev option. This optional feature of course requires some sudo access (to run dd-helper.py as root) which upvm will walk you through the first time you do it.
Note: these instructions are most likely outdated and install an older, Python-2 version of upvm
command -v dnf && dnf=dnf || dnf=yum
sudo $dnf install http://people.redhat.com/rsawhill/rpms/latest-rsawaroha-release.rpm
sudo $dnf install upvm
sudo /usr/share/upvm/initial-setup
upvm -h
- Install libguestfs-tools (
virt-buildercommand) - Make sure you have the
virshandvirt-installcommands as well git clone https://github.com/godlike64/upvm.gitcd upvmsudo ./initial-setup./upvm.py -h
Regardless of install method, you will be missing 2 HIGHLY-recommended but optional python modules. One provides bash-tab-completion and the other provides config-file support. upvm will print warnings about this to stderr but if you want to get them now:
- If RHEL7, ensure access to EPEL
sudo $dnf install pipsudo pip install argcompletesudo activate-global-python-argcompletesudo pip install configargparse
Are you annoyed with always having to open the virt-manager GUI to do stuff with you VMs?
- Get valine.
This project was originally created from upvm, which has been unmaintained for a few years now. The most urgent need behind this project was to have Python 3 support, which should work (albeit not everything was tested). All contributions are welcome.
usage: upvm [--loglevel {debug,info,error}] [--build-image-only] [--nocolor]
[--noconsole] [--cachedir VBCACHEDIR] [-h] [--help] [-l]
[--arch {x86_64,i386,i686,ia64,armv71,ppc,ppc64,ppc64le,aarch64,sparc64,s390,s390x}]
[--info] [-n VMNAME] [--os-variant OS] [--root-password SELECTOR]
[--dnsdomain DOMAIN] [--hostname HOSTNAME | --hostname-prompt]
[--no-dhcphostname] [--img-dir DIR] [--img-size SIZE]
[--img-format {auto,raw,qcow2}] [--timezone TIMEZONE]
[--ssh-pubkey FILE] [--upload FILE:DEST] [--run SCRIPT]
[--run-command CMD+ARGS] [--firstboot SCRIPT]
[--firstboot-command CMD+ARGS] [--install PKG,PKG,@GROUP...]
[--firstboot-install PKG,PKG,@GROUP...] [--selinux-relabel]
[--vbuilder-arg ARG] [--primary-blockdev BLOCKDEV] [-m OPTIONS]
[-c OPTIONS] [-d OPTIONS] [-w OPTIONS] [--boot OPTIONS | --uefi]
[--vinstall-arg ARG]
[TEMPLATE]
Leverage virt-builder & virt-install to spin up new VMs with ease
SIMPLE OPTIONS:
Tweak runtime behavior of upvm.
--loglevel {debug,info,error}
Control verbosity during operation; with 'debug', all
external command-calls are logged, including full
virt-builder & virt-install cmdlines; with 'info',
tidbits of status messages are printing along the way
--build-image-only Quit after virt-builder finishes making the image file
--nocolor Disable all color terminal enhancements
--noconsole Disable post-install auto-execution of 'virsh console
VMNAME' (automatically disabled when running with no
tty)
--cachedir VBCACHEDIR
Set the directory for virt-builder cached templates
(default: '/var/cache/upvm' which is only writable by
members of the 'libvirt' group; note that this
contrasts with virt-builder's default more secure &
more wasteful behavior of saving per-user cache in
homedirs)
-h Show short usage summary and exit
--help Show this help message and exit
SELECTING INITIAL IMAGE TO BUILD:
Choose template to build with virt-builder.
TEMPLATE Specify the template to install from
-l, --list List available templates
--arch {x86_64,i386,i686,ia64,armv71,ppc,ppc64,ppc64le,aarch64,sparc64,s390,s390x}
Specify architecture in case there are multiple
choices (defaults to same architecture as upvm)
--info Print virt-builder metadata for chosen template
IMPORTANT HARDWARE-LEVEL (VM) OPTIONS:
Guest hardware configured by virt-install (after image is built).
-n, --name VMNAME Guest name will be VMNAME (defaults to value of
TEMPLATE)
--os-variant OS Optionally override auto-detected os-type to disable
virt-io (default is 'auto', unlike virt-install
itself; see virt-install --os-variant)
TOTALLY OPTIONAL OS-LEVEL OPTIONS:
Disk image modifications done by virt-builder.
--root-password SELECTOR
If this option is not used (either on cmdline or via
config file), upvm will prompt for the root password
at runtime and optionally save it to per-user config-
file for future use; syntax: use 'password:PASSWORD'
or 'file:FILENAME' or 'random' (see virt-builder
--root-password for more)
--dnsdomain DOMAIN Set DNS domain name to be appended to auto-generated
HOSTNAME (default: 'example.com'; keep in mind however
that this option has no effect if --hostname is
specified)
--hostname HOSTNAME Set system hostname inside the OS (default:
automatically uses VM name; specify '!' to prevent
changing the hostname, which implies --no-dhcphostname
option)
--hostname-prompt Configure upvm to always prompt for system hostname
when not specified (default: automatically use VM name
+ value of DOMAIN with no prompt)
--no-dhcphostname Disable the default behavior of adding
'DHCP_HOSTNAME=<HOSTNAME>' to ifcfg-eth0
--img-dir DIR Set directory where image file will be saved to and
launched from (default: '/var/lib/upvm' which is a
simple directory-backed libvirt storage pool named
'upvm')
--img-size SIZE Set output disk image size (default: native template
size; example: '7G')
--img-format {auto,raw,qcow2}
Set output disk image format (default: 'auto', to
auto-detect based on input image format; however,
until this is implemented in virt-builder, it will
fallback to qcow2)
--timezone TIMEZONE Set system timezone inside the OS (use traditional
syntax, i.e., paths rooted in /usr/share/zoneinfo,
e.g.: 'Europe/London' or 'America/Los_Angeles' or
'Asia/Calcutta')
--ssh-pubkey FILE Inject specific ssh pubkeys (option may be used more
than once) into the /root/.ssh/authorized_keys file of
the guest (default: contents of local
'~/.ssh/authorized_keys' and all present pubkeys with
standard filenames -- see man page explanation for ssh
-i option -- will be automatically injected)
--upload FILE:DEST Upload local FILE to DEST in disk image; warning:
directories in DEST must exist and file uid/gid &
perms are preserved (may be used more than once)
--run SCRIPT Launch disk image inside small container, copy local
SCRIPT into disk image and execute it chrooted into
guest filesystem (may be used more than once)
--run-command CMD+ARGS
Launch disk image inside small container, and execute
command chrooted into guest filesystem (may be used
more than once)
--firstboot SCRIPT Install SCRIPT inside disk image such that it will run
on first boot (may be used more than once)
--firstboot-command CMD+ARGS
Configure disk image such that when guest first boots
it will execute command (may be used more than once)
--install PKG,PKG,@GROUP...
Launch disk image inside small container and execute
'yum/apt-get install' against named package(s),
chrooted into guest filesystem (using this auto-
triggers SELinux relabel; may be used more than once)
--firstboot-install PKG,PKG,@GROUP...
Configure disk image such that when guest first boots
it will install named package(s) (may be used more
than once)
--selinux-relabel Trigger an SELinux relabel on first boot (critical if
any important files are changed)
--vbuilder-arg, -B ARG
Add ARG as an extra option/argument to the virt-
builder command which creates the disk image (may be
used more than once; NOTE: to pass options that start
with a dash, use '--vbuilder-arg=--option' or '-B=-o',
for example: '-B=--verbose -B=--update -B=--copy-
in=/localpath:/mnt/remote')
TOTALLY OPTIONAL HARDWARE-LEVEL (VM) OPTIONS:
Guest hardware configured by virt-install (after image is built). Most options
correspond with virt-install options of same name. More configurability is
possible than described. See virt-install man page.
--primary-blockdev BLOCKDEV
To use this option, specify a BLOCKDEV like
'/dev/sdb2' or '/dev/vg/logvol' and then in that case,
the image generated by virt-builder [using the above
options] will be written to BLOCKDEV and then deleted;
the first time you try this as a non-root user, upvm
will exit with a warning explaining how to add the
appropriate sudo access to the dd helper-wrapper
-m, --memory OPTIONS If this option is omitted, the guest will be defined
with a RAM allocation of 1024 MiB; in its simplest
form OPTIONS can be the amount of RAM in MiB; more
complicated example: '-m 512,maxmemory=4096' (which
would set the default to 512 MiB but allow raising up
to 4 GiB without powering off)
-c, --vcpus OPTIONS If this option is omitted, the guest will be allocated
a single virtual CPU; in its simplest form OPTIONS can
be the number of CPUs; more complicated example: '-c
1,maxvcpus=4' (which would set the default 1 CPU but
allow raising to 4 without powering off)
-d, --disk OPTIONS In its simplest form OPTIONS can be 'size=N' to add a
new disk of size N GiB to the guest from the default
storage pool (which will likely result in a new auto-
named sparse qcow2 image file in
/var/lib/libvirt/images); more examples: '-d
device=cdrom' or '-d /path/iso,device=cdrom' or '-d
/path/existing/file' or '-d pool=storagepool,size=5'
-w, --network OPTIONS
If this option is omitted, a single NIC will be
created in the guest and connected to a bridge (if one
exists) or the 'default' virtual network; if this
option is used once it will modify the default NIC;
this option can be specified multiple times to setup
more than one NIC; in its simplest form OPTIONS can be
'bridge=BRIDGE' (where BRIDGE is a bridge device name,
e.g., 'br0') or 'network=NAME' (where NAME is a
virtual network, e.g., 'default'); more complicated
example: '-w network=default -w bridge=br0' (where the
1st NIC would be connected to the default private
network and the 2nd would be connected to the
[presumably public] bridge br0)
--boot OPTIONS If this option is omitted, its value will default to
'cdrom,hd,network,menu=on,useserial=on'
--uefi This is option does not directly correspond to a virt-
install option of the same name; this option can only
be used if the above '--boot' is omitted; in that case
it will result in adding 'uefi' to the above-mentioned
default boot opts
--vinstall-arg, -I ARG
Add ARG as an extra option/argument to the virt-
install command which creates a guest from the vb-
created disk image (may be used more than once; NOTE:
to pass options that start with a dash, use
'--vinstall-arg=--option' or '-I=-o', for example:
'-I=--cpu=core2duo -I=--video=cirrus
-I=--graphics=vnc,password=mypass')
ABOUT CONFIG FILES:
All of the above options can also be set in config files /etc/upvm.conf or
~/.config/upvm.conf (e.g., 'loglevel = debug') or specified via environment
variables capitalized and prefixed with 'UPVM_' (e.g., UPVM_LOGLEVEL=debug).
See /usr/share/upvm/example.conf for examples of proper config-file syntax
and keep in mind that cmdline values override environment variables which
override config file values which override defaults.
VERSION:
upvm v0.18 last mod Jun 17, 2021
See <http://github.com/godlike64/upvm> to report bugs or RFEs