use PVE::Cluster qw(cfs_register_file cfs_read_file cfs_write_file);
use PVE::CGroup;
+use PVE::CpuSet;
use PVE::DataCenterConfig;
use PVE::Exception qw(raise raise_param_exc);
use PVE::Format qw(render_duration render_bytes);
use PVE::JSONSchema qw(get_standard_option parse_property_string);
use PVE::ProcFSTools;
use PVE::PBSClient;
+use PVE::RESTEnvironment qw(log_warn);
use PVE::RPCEnvironment;
use PVE::Storage;
use PVE::SysFSTools;
use PVE::QMPClient;
use PVE::QemuConfig;
-use PVE::QemuServer::Helpers qw(min_version config_aware_timeout);
+use PVE::QemuServer::Helpers qw(min_version config_aware_timeout windows_version);
use PVE::QemuServer::Cloudinit;
use PVE::QemuServer::CGroup;
use PVE::QemuServer::CPUConfig qw(print_cpu_device get_cpu_options);
optional => 1,
});
-PVE::JSONSchema::register_standard_option('pve-targetstorage', {
- description => "Mapping from source to target storages. Providing only a single storage ID maps all source storages to that storage. Providing the special value '1' will map each source storage to itself.",
- type => 'string',
- format => 'storage-pair-list',
- optional => 1,
-});
-
#no warnings 'redefine';
my $nodename_cache;
default => 'std',
optional => 1,
default_key => 1,
- enum => [qw(cirrus qxl qxl2 qxl3 qxl4 none serial0 serial1 serial2 serial3 std virtio vmware)],
+ enum => [qw(cirrus qxl qxl2 qxl3 qxl4 none serial0 serial1 serial2 serial3 std virtio virtio-gl vmware)],
},
memory => {
description => "Sets the VGA memory (in MiB). Has no effect with serial display.",
optional => 1,
type => 'string', format => 'pve-hotplug-features',
description => "Selectively enable hotplug features. This is a comma separated list of"
- ." hotplug features: 'network', 'disk', 'cpu', 'memory' and 'usb'. Use '0' to disable"
+ ." hotplug features: 'network', 'disk', 'cpu', 'memory', 'usb' and 'cloudinit'. Use '0' to disable"
." hotplug completely. Using '1' as value is an alias for the default `network,disk,usb`.",
default => 'network,disk,usb',
},
keyboard => {
optional => 1,
type => 'string',
- description => "Keyboard layout for VNC server. The default is read from the"
- ."'/etc/pve/datacenter.cfg' configuration file. It should not be necessary to set it.",
+ description => "Keyboard layout for VNC server. This option is generally not required and"
+ ." is often better handled from within the guest OS.",
enum => PVE::Tools::kvmkeymaplist(),
default => undef,
},
description => "Some (read-only) meta-information about this guest.",
optional => 1,
},
+ affinity => {
+ type => 'string', format => 'pve-cpuset',
+ description => "List of host cores used to execute guest processes.",
+ optional => 1,
+ },
};
my $cicustom_fmt = {
searchdomain => {
optional => 1,
type => 'string',
- description => "cloud-init: Sets DNS search domains for a container. Create will'
+ description => 'cloud-init: Sets DNS search domains for a container. Create will'
.' automatically use the setting from the host if neither searchdomain nor nameserver'
- .' are set.",
+ .' are set.',
},
nameserver => {
optional => 1,
type => 'string', format => 'address-list',
- description => "cloud-init: Sets DNS server IP address for a container. Create will'
+ description => 'cloud-init: Sets DNS server IP address for a container. Create will'
.' automatically use the setting from the host if neither searchdomain nor nameserver'
- .' are set.",
+ .' are set.',
},
sshkeys => {
optional => 1,
$confdesc->{$key} = $confdesc_cloudinit->{$key};
}
+PVE::JSONSchema::register_format('pve-cpuset', \&pve_verify_cpuset);
+sub pve_verify_cpuset {
+ my ($set_text, $noerr) = @_;
+
+ my ($count, $members) = eval { PVE::CpuSet::parse_cpuset($set_text) };
+
+ if ($@) {
+ return if $noerr;
+ die "unable to parse cpuset option\n";
+ }
+
+ return PVE::CpuSet->new($members)->short_string();
+}
+
PVE::JSONSchema::register_format('pve-volume-id-or-qm-path', \&verify_volume_id_or_qm_path);
sub verify_volume_id_or_qm_path {
my ($volid, $noerr) = @_;
- if ($volid eq 'none' || $volid eq 'cdrom' || $volid =~ m|^/|) {
- return $volid;
- }
+ return $volid if $volid eq 'none' || $volid eq 'cdrom';
+
+ return verify_volume_id_or_absolute_path($volid, $noerr);
+}
+
+PVE::JSONSchema::register_format('pve-volume-id-or-absolute-path', \&verify_volume_id_or_absolute_path);
+sub verify_volume_id_or_absolute_path {
+ my ($volid, $noerr) = @_;
+
+ return $volid if $volid =~ m|^/|;
- # if its neither 'none' nor 'cdrom' nor a path, check if its a volume-id
$volid = eval { PVE::JSONSchema::check_format('pve-volume-id', $volid, '') };
if ($@) {
return if $noerr;
usb3 => {
optional => 1,
type => 'boolean',
- description => "Specifies whether if given host option is a USB3 device or port.",
+ description => "Specifies whether if given host option is a USB3 device or port."
+ ." For modern guests (machine version >= 7.1 and ostype l26 and windows > 7), this flag"
+ ." is irrelevant (all devices are plugged into a xhci controller).",
default => 0,
},
};
$data = $confdesc->{hotplug}->{default} if $data eq '1';
foreach my $feature (PVE::Tools::split_list($data)) {
- if ($feature =~ m/^(network|disk|cpu|memory|usb)$/) {
+ if ($feature =~ m/^(network|disk|cpu|memory|usb|cloudinit)$/) {
$res->{$1} = 1;
} else {
die "invalid hotplug feature '$feature'\n";
# we use uhci for old VMs because tablet driver was buggy in older qemu
my $usbbus;
- if (PVE::QemuServer::Machine::machine_type_is_q35($conf) || $arch eq 'aarch64') {
+ if ($q35 || $arch eq 'aarch64') {
$usbbus = 'ehci';
} else {
$usbbus = 'uhci';
# sometimes, just plain disable...
my $lvm_no_io_uring = $scfg && $scfg->{type} eq 'lvm';
+ # io_uring causes problems when used with CIFS since kernel 5.15
+ # Some discussion: https://www.spinics.net/lists/linux-cifs/msg26734.html
+ my $cifs_no_io_uring = $scfg && $scfg->{type} eq 'cifs';
+
if (!$drive->{aio}) {
- if ($io_uring && !$rbd_no_io_uring && !$lvm_no_io_uring) {
+ if ($io_uring && !$rbd_no_io_uring && !$lvm_no_io_uring && !$cifs_no_io_uring) {
# io_uring supports all cache modes
$opts .= ",aio=io_uring";
} else {
my ($pbs_conf, $pbs_name) = @_;
my $blockdev = "driver=pbs,node-name=$pbs_name,read-only=on";
$blockdev .= ",repository=$pbs_conf->{repository}";
+ $blockdev .= ",namespace=$pbs_conf->{namespace}" if $pbs_conf->{namespace};
$blockdev .= ",snapshot=$pbs_conf->{snapshot}";
$blockdev .= ",archive=$pbs_conf->{archive}";
$blockdev .= ",keyfile=$pbs_conf->{keyfile}" if $pbs_conf->{keyfile};
'std' => 'VGA',
'vmware' => 'vmware-svga',
'virtio' => 'virtio-vga',
+ 'virtio-gl' => 'virtio-vga-gl',
};
sub print_vga_device {
my $memory = "";
if ($vgamem_mb) {
- if ($vga->{type} eq 'virtio') {
+ if ($vga->{type} =~ /^virtio/) {
my $bytes = PVE::Tools::convert_size($vgamem_mb, "mb" => "b");
$memory = ",max_hostmem=$bytes";
} elsif ($qxlnum) {
$pciaddr = print_pci_addr($vgaid, $bridges, $arch, $machine);
}
+ if ($vga->{type} eq 'virtio-gl') {
+ my $base = '/usr/lib/x86_64-linux-gnu/lib';
+ die "missing libraries for '$vga->{type}' detected! Please install 'libgl1' and 'libegl1'\n"
+ if !-e "${base}EGL.so.1" || !-e "${base}GL.so.1";
+
+ die "no DRM render node detected (/dev/dri/renderD*), no GPU? - needed for '$vga->{type}' display\n"
+ if !PVE::Tools::dir_glob_regex('/dev/dri/', "renderD.*");
+ }
+
return "$type,id=${vgaid}${memory}${max_outputs}${pciaddr}${edidoff}";
}
if (drive_is_cloudinit($drive)) {
eval { PVE::Storage::vdisk_free($storecfg, $drive->{file}) };
warn $@ if $@;
+ delete $conf->{cloudinit};
} elsif (!drive_is_cdrom($drive)) {
my $volid = $drive->{file};
if (vm_is_volid_owner($storecfg, $vmid, $volid)) {
# add JSON properties for create and set function
sub json_config_properties {
- my $prop = shift;
+ my ($prop, $with_disk_alloc) = @_;
my $skip_json_config_opts = {
parent => 1,
foreach my $opt (keys %$confdesc) {
next if $skip_json_config_opts->{$opt};
- $prop->{$opt} = $confdesc->{$opt};
+
+ if ($with_disk_alloc && is_valid_drivename($opt)) {
+ $prop->{$opt} = $PVE::QemuServer::Drive::drivedesc_hash_with_alloc->{$opt};
+ } else {
+ $prop->{$opt} = $confdesc->{$opt};
+ }
}
return $prop;
}
+# Properties that we can read from an OVF file
+sub json_ovf_properties {
+ my $prop = {};
+
+ for my $device (PVE::QemuServer::Drive::valid_drive_names()) {
+ $prop->{$device} = {
+ type => 'string',
+ format => 'pve-volume-id-or-absolute-path',
+ description => "Disk image that gets imported to $device",
+ optional => 1,
+ };
+ }
+
+ $prop->{cores} = {
+ type => 'integer',
+ description => "The number of CPU cores.",
+ optional => 1,
+ };
+ $prop->{memory} = {
+ type => 'integer',
+ description => "Amount of RAM for the VM in MB.",
+ optional => 1,
+ };
+ $prop->{name} = {
+ type => 'string',
+ description => "Name of the VM.",
+ optional => 1,
+ };
+
+ return $prop;
+}
+
# return copy of $confdesc_cloudinit to generate documentation
sub cloudinit_config_properties {
digest => Digest::SHA::sha1_hex($raw),
snapshots => {},
pending => {},
+ cloudinit => {},
};
my $handle_error = sub {
$descr = undef;
$conf = $res->{$section} = {};
next;
+ } elsif ($line =~ m/^\[special:cloudinit\]\s*$/i) {
+ $section = 'cloudinit';
+ $descr = undef;
+ $conf = $res->{$section} = {};
+ next;
} elsif ($line =~ m/^\[([a-z][a-z0-9_\-]+)\]\s*$/i) {
$section = $1;
next;
}
- if ($line =~ m/^\#(.*)\s*$/) {
+ if ($line =~ m/^\#(.*)$/) {
$descr = '' if !defined($descr);
$descr .= PVE::Tools::decode_text($1) . "\n";
next;
foreach my $key (keys %$cref) {
next if $key eq 'digest' || $key eq 'description' || $key eq 'snapshots' ||
- $key eq 'snapstate' || $key eq 'pending';
+ $key eq 'snapstate' || $key eq 'pending' || $key eq 'cloudinit';
my $value = $cref->{$key};
if ($key eq 'delete') {
die "propertry 'delete' is only allowed in [PENDING]\n"
&$cleanup_config($conf->{pending}, 1);
+ &$cleanup_config($conf->{cloudinit});
+
foreach my $snapname (keys %{$conf->{snapshots}}) {
die "internal error: snapshot name '$snapname' is forbidden" if lc($snapname) eq 'pending';
&$cleanup_config($conf->{snapshots}->{$snapname}, undef, $snapname);
}
foreach my $key (sort keys %$conf) {
- next if $key =~ /^(digest|description|pending|snapshots)$/;
+ next if $key =~ /^(digest|description|pending|cloudinit|snapshots)$/;
$raw .= "$key: $conf->{$key}\n";
}
return $raw;
$raw .= &$generate_raw_config($conf->{pending}, 1);
}
+ if (scalar(keys %{$conf->{cloudinit}})){
+ $raw .= "\n[special:cloudinit]\n";
+ $raw .= &$generate_raw_config($conf->{cloudinit});
+ }
+
foreach my $snapname (sort keys %{$conf->{snapshots}}) {
$raw .= "\n[$snapname]\n";
$raw .= &$generate_raw_config($conf->{snapshots}->{$snapname});
return \@flags;
}
-my sub get_cpuunits {
- my ($conf) = @_;
- my $is_cgroupv2 = PVE::CGroup::cgroup_mode() == 2;
-
- my $cpuunits = $conf->{cpuunits};
- return $is_cgroupv2 ? 100 : 1024 if !defined($cpuunits);
-
- if ($is_cgroupv2) {
- $cpuunits = 10000 if $cpuunits >= 10000; # v1 can be higher, so clamp v2 there
- } else {
- $cpuunits = 2 if $cpuunits < 2; # v2 can be lower, so clamp v1 there
- }
- return $cpuunits;
-}
-
# Since commit 277d33454f77ec1d1e0bc04e37621e4dd2424b67 in pve-qemu, smm is not off by default
# anymore. But smm=off seems to be required when using SeaBIOS and serial display.
my sub should_disable_smm {
my $use_old_bios_files = undef;
($use_old_bios_files, $machine_type) = qemu_use_old_bios_files($machine_type);
+ if ($conf->{affinity}) {
+ push @$cmd, "/usr/bin/taskset";
+ push @$cmd, "--cpu-list";
+ push @$cmd, "--all-tasks";
+ push @$cmd, $conf->{affinity};
+ }
+
push @$cmd, $kvm_binary;
push @$cmd, '-id', $vmid;
my $vmname = $conf->{name} || "vm$vmid";
- push @$cmd, '-name', $vmname;
+ push @$cmd, '-name', "$vmname,debug-threads=on";
push @$cmd, '-no-shutdown';
$read_only_str = ',readonly=on' if drive_is_read_only($conf, $d);
} else {
- warn "no efidisk configured! Using temporary efivars disk.\n";
+ log_warn("no efidisk configured! Using temporary efivars disk.");
$path = "/tmp/$vmid-ovmf.fd";
PVE::Tools::file_copy($ovmf_vars, $path, -s $ovmf_vars);
$format = 'raw';
# add usb controllers
my @usbcontrollers = PVE::QemuServer::USB::get_usb_controllers(
- $conf, $bridges, $arch, $machine_type, $usbdesc->{format}, $MAX_USB_DEVICES);
+ $conf, $bridges, $arch, $machine_type, $usbdesc->{format}, $MAX_USB_DEVICES, $machine_version);
push @$devices, @usbcontrollers if @usbcontrollers;
my $vga = parse_vga($conf->{vga});
$usb_dev_features->{spice_usb3} = 1 if min_version($machine_version, 4, 0);
my @usbdevices = PVE::QemuServer::USB::get_usb_devices(
- $conf, $usbdesc->{format}, $MAX_USB_DEVICES, $usb_dev_features, $bootorder);
+ $conf, $usbdesc->{format}, $MAX_USB_DEVICES, $usb_dev_features, $bootorder, $machine_version);
push @$devices, @usbdevices if @usbdevices;
# serial devices
if ($vga->{type} && $vga->{type} !~ m/^serial\d+$/ && $vga->{type} ne 'none'){
push @$devices, '-device', print_vga_device(
$conf, $vga, $arch, $machine_version, $machine_type, undef, $qxlnum, $bridges);
+
+ push @$cmd, '-display', 'egl-headless,gl=core' if $vga->{type} eq 'virtio-gl'; # VIRGL
+
my $socket = PVE::QemuServer::Helpers::vnc_socket($vmid);
push @$cmd, '-vnc', "unix:$socket,password=on";
} else {
my $spice_port;
- if ($qxlnum) {
+ if ($qxlnum || $vga->{type} =~ /^virtio/) {
if ($qxlnum > 1) {
if ($winversion){
for (my $i = 1; $i < $qxlnum; $i++){
# enable balloon by default, unless explicitly disabled
if (!defined($conf->{balloon}) || $conf->{balloon}) {
my $pciaddr = print_pci_addr("balloon0", $bridges, $arch, $machine_type);
- push @$devices, '-device', "virtio-balloon-pci,id=balloon0$pciaddr";
+ my $ballooncmd = "virtio-balloon-pci,id=balloon0$pciaddr";
+ $ballooncmd .= ",free-page-reporting=on" if min_version($machine_version, 6, 2);
+ push @$devices, '-device', $ballooncmd;
}
if ($conf->{watchdog}) {
$iothread .= ",iothread=iothread-$controller_prefix$controller";
push @$cmd, '-object', "iothread,id=iothread-$controller_prefix$controller";
} elsif ($drive->{iothread}) {
- warn "iothread is only valid with virtio disk or virtio-scsi-single controller, ignoring\n";
+ log_warn(
+ "iothread is only valid with virtio disk or virtio-scsi-single controller, ignoring\n"
+ );
}
my $queues = '';
'tags' => 1,
};
+for my $opt (keys %$confdesc_cloudinit) {
+ $fast_plug_option->{$opt} = 1;
+};
+
# hotplug changes in [PENDING]
# $selection hash can be used to only apply specified options, for
# example: { cores => 1 } (only apply changed 'cores')
die "skip\n" if !$hotplug_features->{memory};
PVE::QemuServer::Memory::qemu_memory_hotplug($vmid, $conf, $defaults, $opt);
} elsif ($opt eq 'cpuunits') {
- $cgroup->change_cpu_shares(undef, 1024);
+ $cgroup->change_cpu_shares(undef);
} elsif ($opt eq 'cpulimit') {
$cgroup->change_cpu_quota(undef, undef); # reset, cgroup module can better decide values
} else {
}
}
- my ($apply_pending_cloudinit, $apply_pending_cloudinit_done);
- $apply_pending_cloudinit = sub {
- return if $apply_pending_cloudinit_done; # once is enough
- $apply_pending_cloudinit_done = 1; # once is enough
-
- my ($key, $value) = @_;
-
- my @cloudinit_opts = keys %$confdesc_cloudinit;
- foreach my $opt (keys %{$conf->{pending}}) {
- next if !grep { $_ eq $opt } @cloudinit_opts;
- $conf->{$opt} = delete $conf->{pending}->{$opt};
- }
-
- my $pending_delete_hash = PVE::QemuConfig->parse_pending_delete($conf->{pending}->{delete});
- foreach my $opt (sort keys %$pending_delete_hash) {
- next if !grep { $_ eq $opt } @cloudinit_opts;
- PVE::QemuConfig->remove_from_pending_delete($conf, $opt);
- delete $conf->{$opt};
- }
-
- my $new_conf = { %$conf };
- $new_conf->{$key} = $value;
- PVE::QemuServer::Cloudinit::generate_cloudinitconfig($new_conf, $vmid);
- };
-
foreach my $opt (keys %{$conf->{pending}}) {
next if $selection && !$selection->{$opt};
my $value = $conf->{pending}->{$opt};
# some changes can be done without hotplug
my $drive = parse_drive($opt, $value);
if (drive_is_cloudinit($drive)) {
- &$apply_pending_cloudinit($opt, $value);
+ PVE::QemuServer::Cloudinit::generate_cloudinitconfig($conf, $vmid);
}
vmconfig_update_disk($storecfg, $conf, $hotplug_features->{disk},
$vmid, $opt, $value, $arch, $machine_type);
die "skip\n" if !$hotplug_features->{memory};
$value = PVE::QemuServer::Memory::qemu_memory_hotplug($vmid, $conf, $defaults, $opt, $value);
} elsif ($opt eq 'cpuunits') {
- my $new_cpuunits = get_cpuunits({ $opt => $conf->{pending}->{$opt} }); # to clamp
- $cgroup->change_cpu_shares($new_cpuunits, 1024);
+ my $new_cpuunits = PVE::CGroup::clamp_cpu_shares($conf->{pending}->{$opt}); #clamp
+ $cgroup->change_cpu_shares($new_cpuunits);
} elsif ($opt eq 'cpulimit') {
my $cpulimit = $conf->{pending}->{$opt} == 0 ? -1 : int($conf->{pending}->{$opt} * 100000);
$cgroup->change_cpu_quota($cpulimit, 100000);
delete $conf->{pending}->{$opt};
}
}
-
PVE::QemuConfig->write_config($vmid, $conf);
+
+ if($hotplug_features->{cloudinit}) {
+ my $pending = PVE::QemuServer::Cloudinit::get_pending_config($conf, $vmid);
+ my $regenerate = undef;
+ for my $item (@$pending) {
+ $regenerate = 1 if defined($item->{delete}) or defined($item->{pending});
+ }
+ PVE::QemuServer::vmconfig_update_cloudinit_drive($storecfg, $conf, $vmid) if $regenerate;
+ }
}
sub try_deallocate_drive {
PVE::QemuConfig->cleanup_pending($conf);
+ my $generate_cloudnit = undef;
+
foreach my $opt (keys %{$conf->{pending}}) { # add/change
next if $opt eq 'delete'; # just to be sure
eval {
if (my $err = $@) {
$add_apply_error->($opt, $err);
} else {
+
+ if (is_valid_drivename($opt)) {
+ my $drive = parse_drive($opt, $conf->{pending}->{$opt});
+ $generate_cloudnit = 1 if drive_is_cloudinit($drive);
+ }
+
$conf->{$opt} = delete $conf->{pending}->{$opt};
}
}
# write all changes at once to avoid unnecessary i/o
PVE::QemuConfig->write_config($vmid, $conf);
+ PVE::QemuServer::Cloudinit::generate_cloudinitconfig($conf, $vmid) if $generate_cloudnit;
}
sub vmconfig_update_net {
vm_deviceplug($storecfg, $conf, $vmid, $opt, $drive, $arch, $machine_type);
}
+sub vmconfig_update_cloudinit_drive {
+ my ($storecfg, $conf, $vmid) = @_;
+
+ my $cloudinit_ds = undef;
+ my $cloudinit_drive = undef;
+
+ PVE::QemuConfig->foreach_volume($conf, sub {
+ my ($ds, $drive) = @_;
+ if (PVE::QemuServer::drive_is_cloudinit($drive)) {
+ $cloudinit_ds = $ds;
+ $cloudinit_drive = $drive;
+ }
+ });
+
+ return if !$cloudinit_drive;
+
+ PVE::QemuServer::Cloudinit::generate_cloudinitconfig($conf, $vmid);
+ my $running = PVE::QemuServer::check_running($vmid);
+
+ if ($running) {
+ my $path = PVE::Storage::path($storecfg, $cloudinit_drive->{file});
+ if ($path) {
+ mon_cmd($vmid, "eject", force => JSON::true, id => "$cloudinit_ds");
+ mon_cmd($vmid, "blockdev-change-medium", id => "$cloudinit_ds", filename => "$path");
+ }
+ }
+}
+
# called in locked context by incoming migration
sub vm_migrate_get_nbd_disks {
my ($storecfg, $conf, $replicated_volumes) = @_;
# type => secure/insecure - tunnel over encrypted connection or plain-text
# nbd_proto_version => int, 0 for TCP, 1 for UNIX
# replicated_volumes => which volids should be re-used with bitmaps for nbd migration
-# tpmstate_vol => new volid of tpmstate0, not yet contained in config
+# offline_volumes => new volids of offline migrated disks like tpmstate and cloudinit, not yet
+# contained in config
sub vm_start_nolock {
my ($storecfg, $vmid, $conf, $params, $migrate_opts) = @_;
# this way we can reuse the old ISO with the correct config
PVE::QemuServer::Cloudinit::generate_cloudinitconfig($conf, $vmid) if !$migratedfrom;
- # override TPM state vol if migrated, conf is out of date still
- if (my $tpmvol = $migrate_opts->{tpmstate_vol}) {
- my $parsed = parse_drive("tpmstate0", $conf->{tpmstate0});
- $parsed->{file} = $tpmvol;
- $conf->{tpmstate0} = print_drive($parsed);
+ # override offline migrated volumes, conf is out of date still
+ if (my $offline_volumes = $migrate_opts->{offline_volumes}) {
+ for my $key (sort keys $offline_volumes->%*) {
+ my $parsed = parse_drive($key, $conf->{$key});
+ $parsed->{file} = $offline_volumes->{$key};
+ $conf->{$key} = print_drive($parsed);
+ }
}
my $defaults = load_defaults();
PVE::QemuServer::PCI::reserve_pci_usage($pci_id_list, $vmid, $start_timeout);
eval {
+ my $uuid;
for my $id (sort keys %$pci_devices) {
my $d = $pci_devices->{$id};
for my $dev ($d->{pciid}->@*) {
- PVE::QemuServer::PCI::prepare_pci_device($vmid, $dev->{id}, $id, $d->{mdev});
+ my $info = PVE::QemuServer::PCI::prepare_pci_device($vmid, $dev->{id}, $id, $d->{mdev});
+
+ # nvidia grid needs the uuid of the mdev as qemu parameter
+ if ($d->{mdev} && !defined($uuid) && $info->{vendor} eq '10de') {
+ $uuid = PVE::QemuServer::PCI::generate_mdev_uuid($vmid, $id);
+ }
}
}
+ push @$cmd, '-uuid', $uuid if defined($uuid);
};
if (my $err = $@) {
- eval { PVE::QemuServer::PCI::remove_pci_reservation($pci_id_list) };
+ eval { cleanup_pci_devices($vmid, $conf) };
warn $@ if $@;
die $err;
}
# timeout should be more than enough here...
PVE::Systemd::wait_for_unit_removed("$vmid.scope", 20);
- my $cpuunits = get_cpuunits($conf);
+ my $cpuunits = PVE::CGroup::clamp_cpu_shares($conf->{cpuunits});
my %run_params = (
timeout => $statefile ? undef : $start_timeout,
if (my $err = $@) {
# deactivate volumes if start fails
eval { PVE::Storage::deactivate_volumes($storecfg, $vollist); };
- eval { PVE::QemuServer::PCI::remove_pci_reservation($pci_id_list) };
+ warn $@ if $@;
+ eval { cleanup_pci_devices($vmid, $conf) };
+ warn $@ if $@;
die "start failed: $err";
}
return $vollist;
}
+sub cleanup_pci_devices {
+ my ($vmid, $conf) = @_;
+
+ foreach my $key (keys %$conf) {
+ next if $key !~ m/^hostpci(\d+)$/;
+ my $hostpciindex = $1;
+ my $uuid = PVE::SysFSTools::generate_mdev_uuid($vmid, $hostpciindex);
+ my $d = parse_hostpci($conf->{$key});
+ if ($d->{mdev}) {
+ # NOTE: avoid PVE::SysFSTools::pci_cleanup_mdev_device as it requires PCI ID and we
+ # don't want to break ABI just for this two liner
+ my $dev_sysfs_dir = "/sys/bus/mdev/devices/$uuid";
+ PVE::SysFSTools::file_write("$dev_sysfs_dir/remove", "1") if -e $dev_sysfs_dir;
+ }
+ }
+ PVE::QemuServer::PCI::remove_pci_reservation($vmid);
+}
+
sub vm_stop_cleanup {
my ($storecfg, $vmid, $conf, $keepActive, $apply_pending_changes) = @_;
unlink '/dev/shm/pve-shm-' . ($ivshmem->{name} // $vmid);
}
- my $ids = [];
- foreach my $key (keys %$conf) {
- next if $key !~ m/^hostpci(\d+)$/;
- my $hostpciindex = $1;
- my $d = parse_hostpci($conf->{$key});
- my $uuid = PVE::SysFSTools::generate_mdev_uuid($vmid, $hostpciindex);
-
- foreach my $pci (@{$d->{pciid}}) {
- my $pciid = $pci->{id};
- push @$ids, $pci->{id};
- PVE::SysFSTools::pci_cleanup_mdev_device($pciid, $uuid);
- }
- }
- PVE::QemuServer::PCI::remove_pci_reservation($ids);
+ cleanup_pci_devices($vmid, $conf);
vmconfig_apply_pending($vmid, $conf, $storecfg) if $apply_pending_changes;
};
my $restore_cleanup_oldconf = sub {
my ($storecfg, $vmid, $oldconf, $virtdev_hash) = @_;
+ my $kept_disks = {};
+
PVE::QemuConfig->foreach_volume($oldconf, sub {
my ($ds, $drive) = @_;
if (my $err = $@) {
warn $err;
}
+ } else {
+ $kept_disks->{$volid} = 1;
}
});
- # delete vmstate files, after the restore we have no snapshots anymore
- foreach my $snapname (keys %{$oldconf->{snapshots}}) {
+ # after the restore we have no snapshots anymore
+ for my $snapname (keys $oldconf->{snapshots}->%*) {
my $snap = $oldconf->{snapshots}->{$snapname};
if ($snap->{vmstate}) {
eval { PVE::Storage::vdisk_free($storecfg, $snap->{vmstate}); };
warn $err;
}
}
+
+ for my $volid (keys $kept_disks->%*) {
+ eval { PVE::Storage::volume_snapshot_delete($storecfg, $volid, $snapname); };
+ warn $@ if $@;
+ }
}
};
}
my $restore_deactivate_volumes = sub {
- my ($storecfg, $devinfo) = @_;
+ my ($storecfg, $virtdev_hash) = @_;
my $vollist = [];
- foreach my $devname (keys %$devinfo) {
- my $volid = $devinfo->{$devname}->{volid};
- push @$vollist, $volid if $volid;
+ for my $dev (values $virtdev_hash->%*) {
+ push $vollist->@*, $dev->{volid} if $dev->{volid};
}
- PVE::Storage::deactivate_volumes($storecfg, $vollist);
+ eval { PVE::Storage::deactivate_volumes($storecfg, $vollist); };
+ print STDERR $@ if $@;
};
my $restore_destroy_volumes = sub {
- my ($storecfg, $devinfo) = @_;
+ my ($storecfg, $virtdev_hash) = @_;
- foreach my $devname (keys %$devinfo) {
- my $volid = $devinfo->{$devname}->{volid};
- next if !$volid;
+ for my $dev (values $virtdev_hash->%*) {
+ my $volid = $dev->{volid} or next;
eval {
- if ($volid =~ m|^/|) {
- unlink $volid || die 'unlink failed\n';
- } else {
- PVE::Storage::vdisk_free($storecfg, $volid);
- }
+ PVE::Storage::vdisk_free($storecfg, $volid);
print STDERR "temporary volume '$volid' sucessfuly removed\n";
};
print STDERR "unable to cleanup '$volid' - $@" if $@;
}
};
+my $restore_merge_config = sub {
+ my ($filename, $backup_conf_raw, $override_conf) = @_;
+
+ my $backup_conf = parse_vm_config($filename, $backup_conf_raw);
+ for my $key (keys $override_conf->%*) {
+ $backup_conf->{$key} = $override_conf->{$key};
+ }
+
+ return $backup_conf;
+};
+
sub scan_volids {
my ($cfg, $vmid) = @_;
my $keyfile = PVE::Storage::PBSPlugin::pbs_encryption_key_file_name($storecfg, $storeid);
my $repo = PVE::PBSClient::get_repository($scfg);
+ my $namespace = $scfg->{namespace};
# This is only used for `pbs-restore` and the QEMU PBS driver (live-restore)
my $password = PVE::Storage::PBSPlugin::pbs_get_password($scfg, $storeid);
my $new_conf_raw = '';
my $rpcenv = PVE::RPCEnvironment::get();
- my $devinfo = {};
+ my $devinfo = {}; # info about drives included in backup
+ my $virtdev_hash = {}; # info about allocated drives
eval {
# enable interrupts
my $index = PVE::Tools::file_get_contents($index_fn);
$index = decode_json($index);
- # print Dumper($index);
foreach my $info (@{$index->{files}}) {
if ($info->{filename} =~ m/^(drive-\S+).img.fidx$/) {
my $devname = $1;
my $fh = IO::File->new($cfgfn, "r") ||
die "unable to read qemu-server.conf - $!\n";
- my $virtdev_hash = $parse_backup_hints->($rpcenv, $user, $storecfg, $fh, $devinfo, $options);
+ $virtdev_hash = $parse_backup_hints->($rpcenv, $user, $storecfg, $fh, $devinfo, $options);
# fixme: rate limit?
# for live-restore we only want to preload the efidisk and TPM state
next if $options->{live} && $virtdev ne 'efidisk0' && $virtdev ne 'tpmstate0';
+ my @ns_arg;
+ if (defined(my $ns = $scfg->{namespace})) {
+ @ns_arg = ('--ns', $ns);
+ }
+
my $pbs_restore_cmd = [
'/usr/bin/pbs-restore',
'--repository', $repo,
+ @ns_arg,
$pbs_backup_name,
"$d->{devname}.img.fidx",
$path,
my $err = $@;
if ($err || !$options->{live}) {
- $restore_deactivate_volumes->($storecfg, $devinfo);
+ $restore_deactivate_volumes->($storecfg, $virtdev_hash);
}
rmtree $tmpdir;
if ($err) {
- $restore_destroy_volumes->($storecfg, $devinfo);
+ $restore_destroy_volumes->($storecfg, $virtdev_hash);
die $err;
}
$new_conf_raw .= "\nlock: create";
}
- PVE::Tools::file_set_contents($conffile, $new_conf_raw);
-
- PVE::Cluster::cfs_update(); # make sure we read new file
+ my $new_conf = $restore_merge_config->($conffile, $new_conf_raw, $options->{override_conf});
+ PVE::QemuConfig->write_config($vmid, $new_conf);
eval { rescan($vmid, 1); };
warn $@ if $@;
# these special drives are already restored before start
delete $devinfo->{'drive-efidisk0'};
delete $devinfo->{'drive-tpmstate0-backup'};
- pbs_live_restore($vmid, $conf, $storecfg, $devinfo, $repo, $keyfile, $pbs_backup_name);
+
+ my $pbs_opts = {
+ repo => $repo,
+ keyfile => $keyfile,
+ snapshot => $pbs_backup_name,
+ namespace => $namespace,
+ };
+ pbs_live_restore($vmid, $conf, $storecfg, $devinfo, $pbs_opts);
PVE::QemuConfig->remove_lock($vmid, "create");
}
}
sub pbs_live_restore {
- my ($vmid, $conf, $storecfg, $restored_disks, $repo, $keyfile, $snap) = @_;
+ my ($vmid, $conf, $storecfg, $restored_disks, $opts) = @_;
print "starting VM for live-restore\n";
- print "repository: '$repo', snapshot: '$snap'\n";
+ print "repository: '$opts->{repo}', snapshot: '$opts->{snapshot}'\n";
my $pbs_backing = {};
for my $ds (keys %$restored_disks) {
$ds =~ m/^drive-(.*)$/;
my $confname = $1;
$pbs_backing->{$confname} = {
- repository => $repo,
- snapshot => $snap,
+ repository => $opts->{repo},
+ snapshot => $opts->{snapshot},
archive => "$ds.img.fidx",
};
- $pbs_backing->{$confname}->{keyfile} = $keyfile if -e $keyfile;
+ $pbs_backing->{$confname}->{keyfile} = $opts->{keyfile} if -e $opts->{keyfile};
+ $pbs_backing->{$confname}->{namespace} = $opts->{namespace} if defined($opts->{namespace});
my $drive = parse_drive($confname, $conf->{$confname});
print "restoring '$ds' to '$drive->{file}'\n";
my $err = $@;
if ($err) {
- warn "An error occured during live-restore: $err\n";
+ warn "An error occurred during live-restore: $err\n";
_do_vm_stop($storecfg, $vmid, 1, 1, 10, 0, 1);
die "live-restore failed\n";
}
my $oldtimeout;
my $timeout = 5;
- my $devinfo = {};
+ my $devinfo = {}; # info about drives included in backup
+ my $virtdev_hash = {}; # info about allocated drives
my $rpcenv = PVE::RPCEnvironment::get();
PVE::Tools::file_copy($fwcfgfn, "${pve_firewall_dir}/$vmid.fw");
}
- my $virtdev_hash = $parse_backup_hints->($rpcenv, $user, $cfg, $fh, $devinfo, $opts);
+ $virtdev_hash = $parse_backup_hints->($rpcenv, $user, $cfg, $fh, $devinfo, $opts);
foreach my $info (values %{$virtdev_hash}) {
my $storeid = $info->{storeid};
alarm($oldtimeout) if $oldtimeout;
- $restore_deactivate_volumes->($cfg, $devinfo);
+ $restore_deactivate_volumes->($cfg, $virtdev_hash);
close($fifofh) if $fifofh;
unlink $mapfifo;
rmtree $tmpdir;
if ($err) {
- $restore_destroy_volumes->($cfg, $devinfo);
+ $restore_destroy_volumes->($cfg, $virtdev_hash);
die $err;
}
- PVE::Tools::file_set_contents($conffile, $new_conf_raw);
-
- PVE::Cluster::cfs_update(); # make sure we read new file
+ my $new_conf = $restore_merge_config->($conffile, $new_conf_raw, $opts->{override_conf});
+ PVE::QemuConfig->write_config($vmid, $new_conf);
eval { rescan($vmid, 1); };
warn $@ if $@;
sub restore_tar_archive {
my ($archive, $vmid, $user, $opts) = @_;
+ if (scalar(keys $opts->{override_conf}->%*) > 0) {
+ my $keystring = join(' ', keys $opts->{override_conf}->%*);
+ die "cannot pass along options ($keystring) when restoring from tar archive\n";
+ }
+
if ($archive ne '-') {
my $firstfile = tar_archive_read_firstfile($archive);
die "ERROR: file '$archive' does not look like a QemuServer vzdump backup\n"
$src_path = PVE::Storage::path($storecfg, $src_volid, $snapname);
$src_is_iscsi = ($src_path =~ m|^iscsi://|);
$cachemode = 'none' if $src_scfg->{type} eq 'zfspool';
- } elsif (-f $src_volid) {
+ } elsif (-f $src_volid || -b $src_volid) {
$src_path = $src_volid;
if ($src_path =~ m/\.($PVE::QemuServer::Drive::QEMU_FORMAT_RE)$/) {
$src_format = $1;
my ($storecfg, $source, $dest, $full, $newvollist, $jobs, $completion, $qga, $bwlimit) = @_;
my ($vmid, $running) = $source->@{qw(vmid running)};
- my ($drivename, $drive, $snapname) = $source->@{qw(drivename drive snapname)};
+ my ($src_drivename, $drive, $snapname) = $source->@{qw(drivename drive snapname)};
- my ($newvmid, $efisize) = $dest->@{qw(vmid efisize)};
+ my ($newvmid, $dst_drivename, $efisize) = $dest->@{qw(vmid drivename efisize)};
my ($storage, $format) = $dest->@{qw(storage format)};
+ my $use_drive_mirror = $full && $running && $src_drivename && !$snapname;
+
+ if ($src_drivename && $dst_drivename && $src_drivename ne $dst_drivename) {
+ die "cloning from/to EFI disk requires EFI disk\n"
+ if $src_drivename eq 'efidisk0' || $dst_drivename eq 'efidisk0';
+ die "cloning from/to TPM state requires TPM state\n"
+ if $src_drivename eq 'tpmstate0' || $dst_drivename eq 'tpmstate0';
+
+ # This would lead to two device nodes in QEMU pointing to the same backing image!
+ die "cannot change drive name when cloning disk from/to the same VM\n"
+ if $use_drive_mirror && $vmid == $newvmid;
+ }
+
+ die "cannot move TPM state while VM is running\n"
+ if $use_drive_mirror && $src_drivename eq 'tpmstate0';
+
my $newvolid;
+ print "create " . ($full ? 'full' : 'linked') . " clone of drive ";
+ print "$src_drivename " if $src_drivename;
+ print "($drive->{file})\n";
+
if (!$full) {
- print "create linked clone of drive $drivename ($drive->{file})\n";
$newvolid = PVE::Storage::vdisk_clone($storecfg, $drive->{file}, $newvmid, $snapname);
push @$newvollist, $newvolid;
} else {
my $dst_format = resolve_dst_disk_format($storecfg, $storeid, $volname, $format);
- print "create full clone of drive $drivename ($drive->{file})\n";
my $name = undef;
my $size = undef;
if (drive_is_cloudinit($drive)) {
}
$snapname = undef;
$size = PVE::QemuServer::Cloudinit::CLOUDINIT_DISK_SIZE;
- } elsif ($drivename eq 'efidisk0') {
+ } elsif ($dst_drivename eq 'efidisk0') {
$size = $efisize or die "internal error - need to specify EFI disk size\n";
- } elsif ($drivename eq 'tpmstate0') {
+ } elsif ($dst_drivename eq 'tpmstate0') {
$dst_format = 'raw';
$size = PVE::QemuServer::Drive::TPMSTATE_DISK_SIZE;
} else {
}
my $sparseinit = PVE::Storage::volume_has_feature($storecfg, 'sparseinit', $newvolid);
- if (!$running || $snapname) {
+ if ($use_drive_mirror) {
+ qemu_drive_mirror($vmid, $src_drivename, $newvolid, $newvmid, $sparseinit, $jobs,
+ $completion, $qga, $bwlimit);
+ } else {
# TODO: handle bwlimits
- if ($drivename eq 'efidisk0') {
+ if ($dst_drivename eq 'efidisk0') {
# the relevant data on the efidisk may be smaller than the source
# e.g. on RBD/ZFS, so we use dd to copy only the amount
# that is given by the OVMF_VARS.fd
- my $src_path = PVE::Storage::path($storecfg, $drive->{file});
+ my $src_path = PVE::Storage::path($storecfg, $drive->{file}, $snapname);
my $dst_path = PVE::Storage::path($storecfg, $newvolid);
+ my $src_format = (PVE::Storage::parse_volname($storecfg, $drive->{file}))[6];
+
# better for Ceph if block size is not too small, see bug #3324
my $bs = 1024*1024;
- run_command(['qemu-img', 'dd', '-n', '-O', $dst_format, "bs=$bs", "osize=$size",
- "if=$src_path", "of=$dst_path"]);
+ my $cmd = ['qemu-img', 'dd', '-n', '-O', $dst_format];
+
+ if ($src_format eq 'qcow2' && $snapname) {
+ die "cannot clone qcow2 EFI disk snapshot - requires QEMU >= 6.2\n"
+ if !min_version(kvm_user_version(), 6, 2);
+ push $cmd->@*, '-l', $snapname;
+ }
+ push $cmd->@*, "bs=$bs", "osize=$size", "if=$src_path", "of=$dst_path";
+ run_command($cmd);
} else {
qemu_img_convert($drive->{file}, $newvolid, $size, $snapname, $sparseinit);
}
- } else {
- die "cannot move TPM state while VM is running\n" if $drivename eq 'tpmstate0';
-
- qemu_drive_mirror($vmid, $drivename, $newvolid, $newvmid, $sparseinit, $jobs,
- $completion, $qga, $bwlimit);
}
}
}
sub get_efivars_size {
- my ($conf) = @_;
+ my ($conf, $efidisk) = @_;
+
my $arch = get_vm_arch($conf);
- my $efidisk = $conf->{efidisk0} ? parse_drive('efidisk0', $conf->{efidisk0}) : undef;
+ $efidisk //= $conf->{efidisk0} ? parse_drive('efidisk0', $conf->{efidisk0}) : undef;
my $smm = PVE::QemuServer::Machine::machine_type_is_q35($conf);
my (undef, $ovmf_vars) = get_ovmf_files($arch, $efidisk, $smm);
die "uefi vars image '$ovmf_vars' not found\n" if ! -f $ovmf_vars;
return ($maxdev, $controller, $controller_prefix);
}
-sub windows_version {
- my ($ostype) = @_;
-
- return 0 if !$ostype;
-
- my $winversion = 0;
-
- if($ostype eq 'wxp' || $ostype eq 'w2k3' || $ostype eq 'w2k') {
- $winversion = 5;
- } elsif($ostype eq 'w2k8' || $ostype eq 'wvista') {
- $winversion = 6;
- } elsif ($ostype =~ m/^win(\d+)$/) {
- $winversion = $1;
- }
-
- return $winversion;
-}
-
sub resolve_dst_disk_format {
my ($storecfg, $storeid, $src_volname, $format) = @_;
my ($defFormat, $validFormats) = PVE::Storage::storage_default_format($storecfg, $storeid);
projects / qemu-server.git / blobdiff