X-Git-Url: https://git.proxmox.com/?p=pve-container.git;a=blobdiff_plain;f=src%2FPVE%2FLXC.pm;h=bb1cbdbd18fd507669d7b4786b9b5e7ce4f99983;hp=0742a53acfa07f60ed451930cbc250504e3889c7;hb=HEAD;hpb=a1ff8c3748674382326b9287af4607995873bc98

diff --git a/src/PVE/LXC.pm b/src/PVE/LXC.pm
index 0742a53..65d0fa8 100644
--- a/src/PVE/LXC.pm
+++ b/src/PVE/LXC.pm
@@ -3,39 +3,51 @@ package PVE::LXC;
 use strict;
 use warnings;
 
-use POSIX qw(EINTR);
-
-use Socket;
-
+use Cwd qw();
+use Errno qw(ELOOP ENOTDIR EROFS ECONNREFUSED EEXIST);
+use Fcntl qw(O_RDONLY O_WRONLY O_NOFOLLOW O_DIRECTORY :mode);
 use File::Path;
 use File::Spec;
-use Cwd qw();
-use Fcntl qw(O_RDONLY O_WRONLY O_NOFOLLOW O_DIRECTORY);
-use Errno qw(ELOOP ENOTDIR EROFS ECONNREFUSED ENOSYS EEXIST);
+use IO::Poll qw(POLLIN POLLHUP);
 use IO::Socket::UNIX;
+use POSIX qw(EINTR);
+use Socket;
+use Time::HiRes qw (gettimeofday);
 
+use PVE::AccessControl;
+use PVE::CGroup;
+use PVE::CpuSet;
 use PVE::Exception qw(raise_perm_exc);
-use PVE::Storage;
-use PVE::SafeSyslog;
+use PVE::Firewall;
+use PVE::GuestHelpers qw(check_vnet_access safe_string_ne safe_num_ne safe_boolean_ne);
 use PVE::INotify;
 use PVE::JSONSchema qw(get_standard_option);
-use PVE::Tools qw(
-    dir_glob_foreach file_get_contents file_set_contents lock_file
-    lock_file_full AT_FDCWD O_PATH $IPV4RE $IPV6RE
-);
-use PVE::CpuSet;
 use PVE::Network;
-use PVE::AccessControl;
 use PVE::ProcFSTools;
+use PVE::RESTEnvironment;
+use PVE::SafeSyslog;
+use PVE::Storage;
+use PVE::Tools qw(
+    run_command
+    dir_glob_foreach
+    file_get_contents
+    file_set_contents
+    AT_FDCWD
+    O_PATH
+    $IPV4RE
+    $IPV6RE
+);
 use PVE::Syscall qw(:fsmount);
+
+use PVE::LXC::CGroup;
 use PVE::LXC::Config;
-use PVE::GuestHelpers qw(safe_string_ne safe_num_ne safe_boolean_ne);
+use PVE::LXC::Monitor;
 use PVE::LXC::Tools;
 
-use Time::HiRes qw (gettimeofday);
 my $have_sdn;
 eval {
     require PVE::Network::SDN::Zones;
+    require PVE::Network::SDN::Vnets;
     $have_sdn = 1;
 };
 
@@ -45,6 +57,8 @@ my $nodename = PVE::INotify::nodename();
 
 my $cpuinfo= PVE::ProcFSTools::read_cpuinfo();
 
+our $NEW_DISK_RE = qr/^([^:\s]+):(\d+(\.\d+)?)$/;
+
 sub config_list {
     my $vmlist = PVE::Cluster::get_vmlist();
     my $res = {};
@@ -56,7 +70,7 @@ sub config_list {
 	my $d = $ids->{$vmid};
 	next if !$d->{node} || $d->{node} ne $nodename;
 	next if !$d->{type} || $d->{type} ne 'lxc';
-	$res->{$vmid} = { type => 'lxc', vmid => $vmid };
+	$res->{$vmid} = { type => 'lxc', vmid => int($vmid) };
     }
     return $res;
 }
@@ -106,23 +120,6 @@ sub get_container_disk_usage {
 
 my $last_proc_vmid_stat;
 
-my $parse_cpuacct_stat = sub {
-    my ($vmid, $unprivileged) = @_;
-
-    my $raw = read_cgroup_value('cpuacct', $vmid, $unprivileged, 'cpuacct.stat', 1);
-
-    my $stat = {};
-
-    if ($raw =~ m/^user (\d+)\nsystem (\d+)\n/) {
-
-	$stat->{utime} = $1;
-	$stat->{stime} = $2;
-
-    }
-
-    return $stat;
-};
-
 our $vmstatus_return_properties = {
     vmid => get_standard_option('pve-vmid'),
     status => {
@@ -179,7 +176,7 @@ our $vmstatus_return_properties = {
 sub vmstatus {
     my ($opt_vmid) = @_;
 
-    my $list = $opt_vmid ? { $opt_vmid => { type => 'lxc', vmid => $opt_vmid }} : config_list();
+    my $list = $opt_vmid ? { $opt_vmid => { type => 'lxc', vmid => int($opt_vmid) }} : config_list();
 
     my $active_hash = list_active_containers();
 
@@ -195,7 +192,7 @@ sub vmstatus {
     foreach my $vmid (keys %$list) {
 	my $d = $list->{$vmid};
 
-	eval { $d->{pid} = find_lxc_pid($vmid) if defined($active_hash->{$vmid}); };
+	eval { $d->{pid} = int(find_lxc_pid($vmid)) if defined($active_hash->{$vmid}); };
 	warn $@ if $@; # ignore errors (consider them stopped)
 
 	$d->{status} = $active_hash->{$vmid} ? 'running' : 'stopped';
@@ -211,18 +208,17 @@ sub vmstatus {
 	$d->{cpus} = $conf->{cores} || $conf->{cpulimit};
 	$d->{cpus} = $cpucount if !$d->{cpus};
 
-	$d->{lock} = $conf->{lock} || '';
 	$d->{tags} = $conf->{tags} if defined($conf->{tags});
 
 	if ($d->{pid}) {
 	    my $res = get_container_disk_usage($vmid, $d->{pid});
-	    $d->{disk} = $res->{used};
-	    $d->{maxdisk} = $res->{total};
+	    $d->{disk} = int($res->{used});
+	    $d->{maxdisk} = int($res->{total});
 	} else {
 	    $d->{disk} = 0;
 	    # use 4GB by default ??
 	    if (my $rootfs = $conf->{rootfs}) {
-		my $rootinfo = PVE::LXC::Config->parse_ct_rootfs($rootfs);
+		my $rootinfo = PVE::LXC::Config->parse_volume('rootfs', $rootfs);
 		$d->{maxdisk} = $rootinfo->{size} || (4*1024*1024*1024);
 	    } else {
 		$d->{maxdisk} = 4*1024*1024*1024;
@@ -243,7 +239,7 @@ sub vmstatus {
 	$d->{diskread} = 0;
 	$d->{diskwrite} = 0;
 
-	$d->{template} = PVE::LXC::Config->is_template($conf);
+	$d->{template} = 1 if PVE::LXC::Config->is_template($conf);
 	$d->{lock} = $conf->{lock} if $conf->{lock};
     }
 
@@ -258,55 +254,45 @@ sub vmstatus {
 
 	my $unpriv = $unprivileged->{$vmid};
 
-	if (-d '/sys/fs/cgroup/memory') {
-	    my $memory_stat = read_cgroup_list('memory', $vmid, $unpriv, 'memory.stat');
-	    my $mem_usage_in_bytes = read_cgroup_value('memory', $vmid, $unpriv, 'memory.usage_in_bytes');
+	my $cgroups = PVE::LXC::CGroup->new($vmid);
 
-	    $d->{mem} = $mem_usage_in_bytes - $memory_stat->{total_cache};
-	    $d->{swap} = read_cgroup_value('memory', $vmid, $unpriv, 'memory.memsw.usage_in_bytes') - $mem_usage_in_bytes;
+	if (defined(my $mem = $cgroups->get_memory_stat())) {
+	    $d->{mem} = int($mem->{mem});
+	    $d->{swap} = int($mem->{swap});
 	} else {
 	    $d->{mem} = 0;
 	    $d->{swap} = 0;
 	}
 
-	if (-d '/sys/fs/cgroup/blkio') {
-	    my $blkio_bytes = read_cgroup_value('blkio', $vmid, 0, 'blkio.throttle.io_service_bytes', 1); # don't check if unpriv
-	    my @bytes = split(/\n/, $blkio_bytes);
-	    foreach my $byte (@bytes) {
-		if (my ($key, $value) = $byte =~ /(Read|Write)\s+(\d+)/) {
-		    $d->{diskread} += $2 if $key eq 'Read';
-		    $d->{diskwrite} += $2 if $key eq 'Write';
-		}
-	    }
+	if (defined(my $blkio = $cgroups->get_io_stats())) {
+	    $d->{diskread} = int($blkio->{diskread});
+	    $d->{diskwrite} = int($blkio->{diskwrite});
 	} else {
 	    $d->{diskread} = 0;
 	    $d->{diskwrite} = 0;
 	}
 
-	if (-d '/sys/fs/cgroup/cpuacct') {
-	    my $pstat = $parse_cpuacct_stat->($vmid, $unpriv);
-
-	    my $used = $pstat->{utime} + $pstat->{stime};
+	if (defined(my $cpu = $cgroups->get_cpu_stat())) {
+	    # Total time (in milliseconds) used up by the cpu.
+	    my $used_ms = $cpu->{utime} + $cpu->{stime};
 
 	    my $old = $last_proc_vmid_stat->{$vmid};
 	    if (!$old) {
 		$last_proc_vmid_stat->{$vmid} = {
 		    time => $cdtime,
-		    used => $used,
+		    used => $used_ms,
 		    cpu => 0,
 		};
 		next;
 	    }
 
-	    my $dtime = ($cdtime -  $old->{time}) * $cpucount * $cpuinfo->{user_hz};
-
-	    if ($dtime > 1000) {
-		my $dutime = $used -  $old->{used};
-
-		$d->{cpu} = (($dutime/$dtime)* $cpucount) / $d->{cpus};
+	    my $delta_ms = ($cdtime - $old->{time}) * $cpucount * 1000.0;
+	    if ($delta_ms > 1000.0) {
+		my $delta_used_ms = $used_ms - $old->{used};
+		$d->{cpu} = (($delta_used_ms / $delta_ms) * $cpucount) / $d->{cpus};
 		$last_proc_vmid_stat->{$vmid} = {
 		    time => $cdtime,
-		    used => $used,
+		    used => $used_ms,
 		    cpu => $d->{cpu},
 		};
 	    } else {
@@ -334,33 +320,6 @@ sub vmstatus {
     return $list;
 }
 
-sub read_cgroup_list($$$$) {
-    my ($group, $vmid, $unprivileged, $name) = @_;
-
-    my $content = read_cgroup_value($group, $vmid, $unprivileged, $name, 1);
-
-    return { split(/\s+/, $content) };
-}
-
-sub read_cgroup_value($$$$$) {
-    my ($group, $vmid, $unprivileged, $name, $full) = @_;
-
-    my $nsdir = $unprivileged ? '' : 'ns/';
-    my $path = "/sys/fs/cgroup/$group/lxc/$vmid/${nsdir}$name";
-
-    return PVE::Tools::file_get_contents($path) if $full;
-
-    return PVE::Tools::file_read_firstline($path);
-}
-
-sub write_cgroup_value {
-   my ($group, $vmid, $name, $value) = @_;
-
-   my $path = "/sys/fs/cgroup/$group/lxc/$vmid/$name";
-   PVE::ProcFSTools::write_proc_entry($path, $value) if -e $path;
-
-}
-
 sub find_lxc_console_pids {
 
     my $res = {};
@@ -454,21 +413,6 @@ sub parse_ipv4_cidr {
     die "unable to parse ipv4 address/mask\n";
 }
 
-sub get_cgroup_subsystems {
-	my $v1 = {};
-	my $v2 = 0;
-	my $data = PVE::Tools::file_get_contents('/proc/self/cgroup');
-	while ($data =~ /^\d+:([^:\n]*):.*$/gm) {
-		my $type = $1;
-		if (length($type)) {
-			$v1->{$_} = 1 foreach split(/,/, $type);
-		} else {
-			$v2 = 1;
-		}
-	}
-	return wantarray ? ($v1, $v2) : $v1;
-}
-
 # With seccomp trap to userspace we now have the ability to optionally forward
 # certain syscalls to the "host" to handle (via our pve-lxc-syscalld daemon).
 #
@@ -496,6 +440,15 @@ sub make_seccomp_config {
 
     my $rules = {
 	keyctl => ['errno 38'],
+
+	# Disable btrfs ioctrls since they don't work particularly well in user namespaces.
+	# Particularly, without the mount option to enable rmdir removing snapshots, user
+	# namespaces can create snapshots but neither `show` or `delete` them, which is quite
+	# horrible, so for now, just disable this entirely:
+	#
+	# BTRFS_IOCTL_MAGIC 0x94, _IOC type shift is 8,
+	# so `(req & 0xFF00) == 0x9400` is a btrfs ioctl and gets an EPERM
+	ioctl  => ['errno 1 [1,0x9400,SCMP_CMP_MASKED_EQ,0xff00]'],
     };
 
     my $raw_conf = '';
@@ -633,8 +586,26 @@ sub update_lxc_config {
 	return;
     }
 
+    my ($lxc_major, $lxc_minor) = get_lxc_version();
+
     my $raw = '';
 
+    if ($lxc_major >= 4) {
+	# Explicitly don't use relative directories, which is the default, but
+	# note that we do this mostly because they are only applied for *some*
+	# cgroups. Our pve-container@.service now starts lxc-start with `-F`,
+	# so we also don't need to worry about the new monitor cgroup to
+	# confuse systemd.
+	$raw .= "lxc.cgroup.relative = 0\n";
+
+	# To make things easier, let's keep our previous cgroup layout and
+	# simply move the monitor outside:
+	$raw .= "lxc.cgroup.dir.monitor = lxc.monitor/$vmid\n";
+	# cgroup namespace separation for stronger limits:
+	$raw .= "lxc.cgroup.dir.container = lxc/$vmid\n";
+	$raw .= "lxc.cgroup.dir.container.inner = ns\n";
+    }
+
     die "missing 'arch' - internal error" if !$conf->{arch};
     $raw .= "lxc.arch = $conf->{arch}\n";
 
@@ -643,6 +614,8 @@ sub update_lxc_config {
 
     my $ostype = $conf->{ostype} || die "missing 'ostype' - internal error";
 
+    File::Path::mkpath($dir);
+
     my $cfgpath = '/usr/share/lxc/config';
     my $inc = "$cfgpath/$ostype.common.conf";
     $inc ="$cfgpath/common.conf" if !-f $inc;
@@ -662,6 +635,25 @@ sub update_lxc_config {
 	$raw .= "lxc.mount.entry = /dev/fuse dev/fuse none bind,create=file 0 0\n";
     }
 
+    if ($unprivileged && !$features->{force_rw_sys}) {
+	# unpriv. CT default to sys:rw, but that doesn't always plays well with
+	# systemd, e.g., systemd-networkd https://systemd.io/CONTAINER_INTERFACE/
+	$raw .= "lxc.mount.auto = sys:mixed\n";
+    }
+
+    PVE::LXC::Config->foreach_passthrough_device($conf, sub {
+	my ($key, $device) = @_;
+
+	die "Path is not defined for passthrough device $key\n"
+	    if !defined($device->{path});
+
+	my ($mode, $rdev) = PVE::LXC::Tools::get_device_mode_and_rdev($device->{path});
+	my $major = PVE::Tools::dev_t_major($rdev);
+	my $minor = PVE::Tools::dev_t_minor($rdev);
+	my $device_type_char = S_ISBLK($mode) ? 'b' : 'c';
+	$raw .= "lxc.cgroup2.devices.allow = $device_type_char $major:$minor rw\n";
+    });
+
     # WARNING: DO NOT REMOVE this without making sure that loop device nodes
     # cannot be exposed to the container with r/w access (cgroup perms).
     # When this is enabled mounts will still remain in the monitor's namespace
@@ -669,7 +661,7 @@ sub update_lxc_config {
     # files while the container is running!
     $raw .= "lxc.monitor.unshare = 1\n";
 
-    my $cgv1 = get_cgroup_subsystems();
+    my ($cgv1, $cgv2) = PVE::CGroup::get_cgroup_controllers();
 
     # Should we read them from /etc/subuid?
     if ($unprivileged && !$custom_idmap) {
@@ -679,7 +671,11 @@ sub update_lxc_config {
 
     if (!PVE::LXC::Config->has_dev_console($conf)) {
 	$raw .= "lxc.console.path = none\n";
-	$raw .= "lxc.cgroup.devices.deny = c 5:1 rwm\n" if $cgv1->{devices};
+	if ($cgv1->{devices}) {
+	    $raw .= "lxc.cgroup.devices.deny = c 5:1 rwm\n";
+	} elsif (defined($cgv2)) {
+	    $raw .= "lxc.cgroup2.devices.deny = c 5:1 rwm\n";
+	}
     }
 
     my $ttycount = PVE::LXC::Config->get_tty_count($conf);
@@ -687,7 +683,7 @@ sub update_lxc_config {
 
     # some init scripts expect a linux terminal (turnkey).
     $raw .= "lxc.environment = TERM=linux\n";
-    
+
     my $utsname = $conf->{hostname} || "CT$vmid";
     $raw .= "lxc.uts.name = $utsname\n";
 
@@ -700,6 +696,18 @@ sub update_lxc_config {
 
 	my $lxcswap = int(($memory + $swap)*1024*1024);
 	$raw .= "lxc.cgroup.memory.memsw.limit_in_bytes = $lxcswap\n";
+    } elsif ($cgv2->{memory}) {
+	my $memory = $conf->{memory} || 512;
+	my $swap = $conf->{swap} // 0;
+
+	# cgroup memory usage is limited by the hard 'max' limit (OOM-killer enforced) and the soft
+	# 'high' limit (cgroup processes get throttled and put under heavy reclaim pressure).
+	my ($lxc_mem_max, $lxc_mem_high) = PVE::LXC::Config::calculate_memory_constraints($memory);
+	$raw .= "lxc.cgroup2.memory.max = $lxc_mem_max\n";
+	$raw .= "lxc.cgroup2.memory.high = $lxc_mem_high\n";
+
+	my $lxcswap = int($swap*1024*1024);
+	$raw .= "lxc.cgroup2.memory.swap.max = $lxcswap\n";
     }
 
     if ($cgv1->{cpu}) {
@@ -709,14 +717,25 @@ sub update_lxc_config {
 	    $raw .= "lxc.cgroup.cpu.cfs_quota_us = $value\n";
 	}
 
-	my $shares = $conf->{cpuunits} || 1024;
+	my $shares = PVE::CGroup::clamp_cpu_shares($conf->{cpuunits});
 	$raw .= "lxc.cgroup.cpu.shares = $shares\n";
+    } elsif ($cgv2->{cpu}) {
+	# See PVE::CGroup
+	if (my $cpulimit = $conf->{cpulimit}) {
+	    my $value = int(100000*$cpulimit);
+	    $raw .= "lxc.cgroup2.cpu.max = $value 100000\n";
+	}
+
+	if (defined(my $shares = $conf->{cpuunits})) {
+	    $shares = PVE::CGroup::clamp_cpu_shares($shares);
+	    $raw .= "lxc.cgroup2.cpu.weight = $shares\n";
+	}
     }
 
     die "missing 'rootfs' configuration\n"
 	if !defined($conf->{rootfs});
 
-    my $mountpoint = PVE::LXC::Config->parse_ct_rootfs($conf->{rootfs});
+    my $mountpoint = PVE::LXC::Config->parse_volume('rootfs', $conf->{rootfs});
 
     $raw .= "lxc.rootfs.path = $dir/rootfs\n";
 
@@ -728,31 +747,47 @@ sub update_lxc_config {
 	$raw .= "lxc.net.$ind.veth.pair = veth${vmid}i${ind}\n";
 	$raw .= "lxc.net.$ind.hwaddr = $d->{hwaddr}\n" if defined($d->{hwaddr});
 	$raw .= "lxc.net.$ind.name = $d->{name}\n" if defined($d->{name});
-	$raw .= "lxc.net.$ind.mtu = $d->{mtu}\n" if defined($d->{mtu});
+
+	my $bridge_mtu = PVE::Network::read_bridge_mtu($d->{bridge});
+	my $mtu = $d->{mtu} || $bridge_mtu;
+
+	# Keep container from starting with invalid mtu configuration
+	die "$k: MTU size '$mtu' is bigger than bridge MTU '$bridge_mtu'\n"
+	    if ($mtu > $bridge_mtu);
+
+	$raw .= "lxc.net.$ind.mtu = $mtu\n";
+
+	# Starting with lxc 4.0, we do not patch lxc to execute our up-scripts.
+	if ($lxc_major >= 4) {
+	    $raw .= "lxc.net.$ind.script.up = /usr/share/lxc/lxcnetaddbr\n";
+	}
     }
 
-    if ($cgv1->{cpuset}) {
-	my $had_cpuset = 0;
-	if (my $lxcconf = $conf->{lxc}) {
-	    foreach my $entry (@$lxcconf) {
-		my ($k, $v) = @$entry;
-		$had_cpuset = 1 if $k eq 'lxc.cgroup.cpuset.cpus';
-		$raw .= "$k = $v\n";
-	    }
+    my $had_cpuset = 0;
+    if (my $lxcconf = $conf->{lxc}) {
+	foreach my $entry (@$lxcconf) {
+	    my ($k, $v) = @$entry;
+	    $had_cpuset = 1 if $k eq 'lxc.cgroup.cpuset.cpus' || $k eq 'lxc.cgroup2.cpuset.cpus';
+	    $raw .= "$k = $v\n";
 	}
+    }
 
-	my $cores = $conf->{cores};
-	if (!$had_cpuset && $cores) {
-	    my $cpuset = eval { PVE::CpuSet->new_from_cgroup('lxc', 'effective_cpus') };
-	    $cpuset = PVE::CpuSet->new_from_cgroup('', 'effective_cpus') if !$cpuset;
-	    my @members = $cpuset->members();
-	    while (scalar(@members) > $cores) {
-		my $randidx = int(rand(scalar(@members)));
-		$cpuset->delete($members[$randidx]);
-		splice(@members, $randidx, 1); # keep track of the changes
-	    }
-	    $raw .= "lxc.cgroup.cpuset.cpus = ".$cpuset->short_string()."\n";
+    my $cpuset;
+    my ($cpuset_cgroup, $cpuset_version) = eval { PVE::CGroup::cpuset_controller_path() };
+    if (defined($cpuset_cgroup)) {
+	$cpuset = eval { PVE::CpuSet->new_from_path("$cpuset_cgroup/lxc", 1) }
+	    || PVE::CpuSet->new_from_path($cpuset_cgroup, 1);
+    }
+    my $cores = $conf->{cores};
+    if (!$had_cpuset && $cores && $cpuset) {
+	my @members = $cpuset->members();
+	while (scalar(@members) > $cores) {
+	    my $randidx = int(rand(scalar(@members)));
+	    $cpuset->delete($members[$randidx]);
+	    splice(@members, $randidx, 1); # keep track of the changes
 	}
+	my $ver = $cpuset_version == 1 ? '' : '2';
+	$raw .= "lxc.cgroup$ver.cpuset.cpus = ".$cpuset->short_string()."\n";
     }
 
     File::Path::mkpath("$dir/rootfs");
@@ -766,7 +801,7 @@ sub verify_nameserver_list {
 
     my @list = ();
     foreach my $server (PVE::Tools::split_list($nameserver_list)) {
-	PVE::JSONSchema::pve_verify_ip($server);
+	PVE::LXC::Config::verify_ip_with_ll_iface($server);
 	push @list, $server;
     }
 
@@ -852,12 +887,33 @@ sub delete_mountpoint_volume {
 }
 
 sub destroy_lxc_container {
-    my ($storage_cfg, $vmid, $conf, $replacement_conf) = @_;
+    my ($storage_cfg, $vmid, $conf, $replacement_conf, $purge_unreferenced) = @_;
 
-    PVE::LXC::Config->foreach_mountpoint($conf, sub {
+    my $volids = {};
+    my $remove_volume = sub {
 	my ($ms, $mountpoint) = @_;
-	delete_mountpoint_volume($storage_cfg, $vmid, $mountpoint->{volume});
-    });
+
+	my $volume = $mountpoint->{volume};
+
+	return if $volids->{$volume};
+	$volids->{$volume} = 1;
+
+	delete_mountpoint_volume($storage_cfg, $vmid, $volume);
+    };
+    PVE::LXC::Config->foreach_volume_full($conf, {include_unused => 1}, $remove_volume);
+
+    PVE::LXC::Config->foreach_volume_full($conf->{pending}, {include_unused => 1}, $remove_volume);
+
+    if ($purge_unreferenced) { # also remove unreferenced disk
+	my $vmdisks = PVE::Storage::vdisk_list($storage_cfg, undef, $vmid, undef, 'rootdir');
+	PVE::Storage::foreach_volid($vmdisks, sub {
+	    my ($volid, $sid, $volname, $d) = @_;
+	    eval { PVE::Storage::vdisk_free($storage_cfg, $volid) };
+	    warn $@ if $@;
+	});
+    }
+
+    delete_ifaces_ipams_ips($conf, $vmid);
 
     rmdir "/var/lib/lxc/$vmid/rootfs";
     unlink "/var/lib/lxc/$vmid/config";
@@ -881,6 +937,32 @@ sub vm_stop_cleanup {
     warn $@ if $@; # avoid errors - just warn
 }
 
+sub net_tap_plug : prototype($$) {
+    my ($iface, $net) = @_;
+
+    if (defined($net->{link_down})) {
+	PVE::Tools::run_command(['/sbin/ip', 'link', 'set', 'dev', $iface, 'down']);
+	# Don't add disconnected interfaces to the bridge, otherwise e.g. applying any network
+	# change (e.g. `ifreload -a`) could (re-)activate it unintentionally.
+	return;
+    }
+
+    my ($bridge, $tag, $trunks, $rate, $hwaddr) =
+	$net->@{'bridge', 'tag', 'trunks', 'rate', 'hwaddr'};
+
+    # The nftable-based implementation from the newer proxmox-firewall does not requires FW bridges
+    my $create_firewall_bridges = $net->{firewall} && !PVE::Firewall::is_nftables();
+
+    if ($have_sdn) {
+	PVE::Network::SDN::Zones::tap_plug($iface, $bridge, $tag, $create_firewall_bridges, $trunks, $rate);
+	PVE::Network::SDN::Zones::add_bridge_fdb($iface, $hwaddr, $bridge);
+    } else {
+	PVE::Network::tap_plug($iface, $bridge, $tag, $create_firewall_bridges, $trunks, $rate, { mac => $hwaddr });
+    }
+
+    PVE::Tools::run_command(['/sbin/ip', 'link', 'set', 'dev', $iface, 'up']);
+}
+
 sub update_net {
     my ($vmid, $conf, $opt, $newnet, $netid, $rootdir) = @_;
 
@@ -899,33 +981,51 @@ sub update_net {
 	    safe_string_ne($oldnet->{name}, $newnet->{name})) {
 
 	    PVE::Network::veth_delete($veth);
+
+	    if ($have_sdn && safe_string_ne($oldnet->{hwaddr}, $newnet->{hwaddr})) {
+		eval { PVE::Network::SDN::Vnets::del_ips_from_mac($oldnet->{bridge}, $oldnet->{hwaddr}, $conf->{hostname}) };
+		warn $@ if $@;
+
+		PVE::Network::SDN::Vnets::add_next_free_cidr($newnet->{bridge}, $conf->{hostname}, $newnet->{hwaddr}, $vmid, undef, 1);
+		PVE::Network::SDN::Vnets::add_dhcp_mapping($newnet->{bridge}, $newnet->{hwaddr}, $vmid, $conf->{hostname});
+	    }
+
 	    delete $conf->{$opt};
 	    PVE::LXC::Config->write_config($vmid, $conf);
 
 	    hotplug_net($vmid, $conf, $opt, $newnet, $netid);
 
 	} else {
-	    if (safe_string_ne($oldnet->{bridge}, $newnet->{bridge}) ||
-		safe_num_ne($oldnet->{tag}, $newnet->{tag}) ||
-		safe_num_ne($oldnet->{firewall}, $newnet->{firewall})) {
+	    my $bridge_changed = safe_string_ne($oldnet->{bridge}, $newnet->{bridge});
 
+	    if ($bridge_changed ||
+		safe_num_ne($oldnet->{tag}, $newnet->{tag}) ||
+		safe_num_ne($oldnet->{firewall}, $newnet->{firewall}) ||
+		safe_boolean_ne($oldnet->{link_down}, $newnet->{link_down})
+	    ) {
 		if ($oldnet->{bridge}) {
+		    my $oldbridge = $oldnet->{bridge};
+
 		    PVE::Network::tap_unplug($veth);
 		    foreach (qw(bridge tag firewall)) {
 			delete $oldnet->{$_};
 		    }
 		    $conf->{$opt} = PVE::LXC::Config->print_lxc_network($oldnet);
 		    PVE::LXC::Config->write_config($vmid, $conf);
+
+		    if ($have_sdn && $bridge_changed) {
+			eval { PVE::Network::SDN::Vnets::del_ips_from_mac($oldbridge, $oldnet->{hwaddr}, $conf->{hostname}) };
+			warn $@ if $@;
+		    }
 		}
 
-		if ($have_sdn) {
-		    PVE::Network::SDN::Zones::tap_plug($veth, $newnet->{bridge}, $newnet->{tag}, $newnet->{firewall}, $newnet->{trunks}, $newnet->{rate});
-		} else {
-		    PVE::Network::tap_plug($veth, $newnet->{bridge}, $newnet->{tag}, $newnet->{firewall}, $newnet->{trunks}, $newnet->{rate});
+		if ($have_sdn && $bridge_changed) {
+		    PVE::Network::SDN::Vnets::add_next_free_cidr($newnet->{bridge}, $conf->{hostname}, $newnet->{hwaddr}, $vmid, undef, 1);
 		}
+		PVE::LXC::net_tap_plug($veth, $newnet);
 
 		# This includes the rate:
-		foreach (qw(bridge tag firewall rate)) {
+		foreach (qw(bridge tag firewall rate link_down)) {
 		    $oldnet->{$_} = $newnet->{$_} if $newnet->{$_};
 		}
 	    } elsif (safe_string_ne($oldnet->{rate}, $newnet->{rate})) {
@@ -938,6 +1038,11 @@ sub update_net {
 	    PVE::LXC::Config->write_config($vmid, $conf);
 	}
     } else {
+	if ($have_sdn) {
+	    PVE::Network::SDN::Vnets::add_next_free_cidr($newnet->{bridge}, $conf->{hostname}, $newnet->{hwaddr}, $vmid, undef, 1);
+	    PVE::Network::SDN::Vnets::add_dhcp_mapping($newnet->{bridge}, $newnet->{hwaddr}, $vmid, $conf->{hostname});
+	}
+
 	hotplug_net($vmid, $conf, $opt, $newnet, $netid);
     }
 
@@ -953,12 +1058,12 @@ sub hotplug_net {
 
     if ($have_sdn) {
 	PVE::Network::SDN::Zones::veth_create($veth, $vethpeer, $newnet->{bridge}, $newnet->{hwaddr});
-	PVE::Network::SDN::Zones::tap_plug($veth, $newnet->{bridge}, $newnet->{tag}, $newnet->{firewall}, $newnet->{trunks}, $newnet->{rate});
     } else {
 	PVE::Network::veth_create($veth, $vethpeer, $newnet->{bridge}, $newnet->{hwaddr});
-	PVE::Network::tap_plug($veth, $newnet->{bridge}, $newnet->{tag}, $newnet->{firewall}, $newnet->{trunks}, $newnet->{rate});
     }
 
+    PVE::LXC::net_tap_plug($veth, $newnet);
+
     # attach peer in container
     my $cmd = ['lxc-device', '-n', $vmid, 'add', $vethpeer, "$eth" ];
     PVE::Tools::run_command($cmd);
@@ -968,7 +1073,7 @@ sub hotplug_net {
     PVE::Tools::run_command($cmd);
 
     my $done = { type => 'veth' };
-    foreach (qw(bridge tag firewall hwaddr name)) {
+    foreach (qw(bridge tag firewall hwaddr name link_down)) {
 	$done->{$_} = $newnet->{$_} if $newnet->{$_};
     }
     $conf->{$opt} = PVE::LXC::Config->print_lxc_network($done);
@@ -976,6 +1081,32 @@ sub hotplug_net {
     PVE::LXC::Config->write_config($vmid, $conf);
 }
 
+sub get_interfaces {
+    my ($vmid) = @_;
+
+    my $pid = eval { find_lxc_pid($vmid); };
+    return if $@;
+
+    my $output;
+    # enters the network namespace of the container and executes 'ip a'
+    run_command(['nsenter', '-t', $pid, '--net', '--', 'ip', '--json', 'a'],
+	outfunc => sub { $output .= shift; });
+
+    my $config = JSON::decode_json($output);
+
+    my $res;
+    for my $interface ($config->@*) {
+	my $obj = { name => $interface->{ifname} };
+	for my $ip ($interface->{addr_info}->@*) {
+	    $obj->{$ip->{family}} = $ip->{local} . "/" . $ip->{prefixlen};
+	}
+	$obj->{hwaddr} = $interface->{address};
+	push @$res, $obj
+    }
+
+    return $res;
+}
+
 sub update_ipconfig {
     my ($vmid, $conf, $opt, $eth, $newnet, $rootdir) = @_;
 
@@ -1133,11 +1264,24 @@ my $do_syncfs = sub {
     my $mountdata = do { local $/ = undef; <$socket> };
     close $socket;
 
+    my %nosyncfs = (
+	cgroup => 1,
+	cgroup2 => 1,
+	devtmpfs => 1,
+	devpts => 1,
+	'fuse.lxcfs' => 1,
+	fusectl => 1,
+	mqueue => 1,
+	proc => 1,
+	sysfs => 1,
+	tmpfs => 1,
+    );
+
     # Now sync all mountpoints...
     my $mounts = PVE::ProcFSTools::parse_mounts($mountdata);
     foreach my $mp (@$mounts) {
 	my ($what, $dir, $fs) = @$mp;
-	next if $fs eq 'fuse.lxcfs';
+	next if $nosyncfs{$fs};
 	eval { PVE::Tools::sync_mountpoint($dir); };
 	warn $@ if $@;
     }
@@ -1185,7 +1329,7 @@ sub template_create {
 
     my $storecfg = PVE::Storage::config();
 
-    PVE::LXC::Config->foreach_mountpoint($conf, sub {
+    PVE::LXC::Config->foreach_volume($conf, sub {
 	my ($ms, $mountpoint) = @_;
 
 	my $volid = $mountpoint->{volume};
@@ -1194,7 +1338,7 @@ sub template_create {
 	    if !PVE::Storage::volume_has_feature($storecfg, 'template', $volid);
     });
 
-    PVE::LXC::Config->foreach_mountpoint($conf, sub {
+    PVE::LXC::Config->foreach_volume($conf, sub {
 	my ($ms, $mountpoint) = @_;
 
 	my $volid = $mountpoint->{volume};
@@ -1210,9 +1354,10 @@ sub template_create {
 }
 
 sub check_ct_modify_config_perm {
-    my ($rpcenv, $authuser, $vmid, $pool, $newconf, $delete) = @_;
+    my ($rpcenv, $authuser, $vmid, $pool, $oldconf, $newconf, $delete, $unprivileged) = @_;
 
     return 1 if $authuser eq 'root@pam';
+    my $storage_cfg = PVE::Storage::config();
 
     my $check = sub {
 	my ($opt, $delete) = @_;
@@ -1221,21 +1366,84 @@ sub check_ct_modify_config_perm {
 	} elsif ($opt eq 'rootfs' || $opt =~ /^mp\d+$/) {
 	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Config.Disk']);
 	    return if $delete;
-	    my $data = $opt eq 'rootfs' ? PVE::LXC::Config->parse_ct_rootfs($newconf->{$opt})
-					: PVE::LXC::Config->parse_ct_mountpoint($newconf->{$opt});
+	    my $data = PVE::LXC::Config->parse_volume($opt, $newconf->{$opt});
 	    raise_perm_exc("mount point type $data->{type} is only allowed for root\@pam")
 		if $data->{type} ne 'volume';
+	    my $volid = $data->{volume};
+	    if ($volid =~ $NEW_DISK_RE) {
+		my $sid = $1;
+		$rpcenv->check($authuser, "/storage/$sid", ['Datastore.AllocateSpace']);
+	    } else {
+		PVE::Storage::check_volume_access(
+		    $rpcenv,
+		    $authuser,
+		    $storage_cfg,
+		    $vmid,
+		    $volid,
+		    'rootdir',
+		);
+	    }
 	} elsif ($opt eq 'memory' || $opt eq 'swap') {
 	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Config.Memory']);
-	} elsif ($opt =~ m/^net\d+$/ || $opt eq 'nameserver' ||
-		 $opt eq 'searchdomain' || $opt eq 'hostname') {
+	} elsif ($opt =~ m/^net\d+$/) {
+	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Config.Network']);
+	    check_bridge_access($rpcenv, $authuser, $oldconf->{$opt}) if $oldconf->{$opt};
+	    check_bridge_access($rpcenv, $authuser, $newconf->{$opt}) if $newconf->{$opt};
+	} elsif ($opt =~ m/^dev\d+$/) {
+	    raise_perm_exc("configuring device passthrough is only allowed for root\@pam");
+	} elsif ($opt eq 'nameserver' || $opt eq 'searchdomain' || $opt eq 'hostname') {
 	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Config.Network']);
 	} elsif ($opt eq 'features') {
-	    # For now this is restricted to root@pam
-	    raise_perm_exc("changing feature flags is only allowed for root\@pam");
+	    raise_perm_exc("changing feature flags for privileged container is only allowed for root\@pam")
+		if !$unprivileged;
+
+	    my $nesting_changed = 0;
+	    my $other_changed = 0;
+	    if (!$delete) {
+		my $features = PVE::LXC::Config->parse_features($newconf->{$opt});
+		if (defined($oldconf) && $oldconf->{$opt}) {
+		    # existing container with features
+		    my $old_features = PVE::LXC::Config->parse_features($oldconf->{$opt});
+		    for my $feature ((keys %$old_features, keys %$features)) {
+			my $old = $old_features->{$feature} // '';
+			my $new = $features->{$feature} // '';
+			if ($old ne $new) {
+			    if ($feature eq 'nesting') {
+				$nesting_changed = 1;
+				next;
+			    } else {
+				$other_changed = 1;
+				last;
+			    }
+			}
+		    }
+		} else {
+		    # new container or no features defined
+		    if (scalar(keys %$features) == 1 && $features->{nesting}) {
+			$nesting_changed = 1;
+		    } elsif (scalar(keys %$features) > 0) {
+			$other_changed = 1;
+		    }
+		}
+	    } else {
+		my $features = PVE::LXC::Config->parse_features($oldconf->{$opt});
+		if (scalar(keys %$features) == 1 && $features->{nesting}) {
+		    $nesting_changed = 1;
+		} elsif (scalar(keys %$features) > 0) {
+		    $other_changed = 1;
+		}
+	    }
+	    raise_perm_exc("changing feature flags (except nesting) is only allowed for root\@pam")
+		if $other_changed;
+	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Allocate'])
+		if $nesting_changed;
 	} elsif ($opt eq 'hookscript') {
 	    # For now this is restricted to root@pam
 	    raise_perm_exc("changing the hookscript is only allowed for root\@pam");
+	} elsif ($opt eq 'tags') {
+	    my $old = $oldconf->{$opt};
+	    my $new = $delete ? '' : $newconf->{$opt};
+	    PVE::GuestHelpers::assert_tag_permissions($vmid, $old, $new, $rpcenv, $authuser);
 	} else {
 	    $rpcenv->check_vm_perm($authuser, $vmid, $pool, ['VM.Config.Options']);
 	}
@@ -1251,6 +1459,18 @@ sub check_ct_modify_config_perm {
     return 1;
 }
 
+sub check_bridge_access {
+    my ($rpcenv, $authuser, $raw) = @_;
+
+    return 1 if $authuser eq 'root@pam';
+
+    my $net = PVE::LXC::Config->parse_lxc_network($raw);
+    my ($bridge, $tag, $trunks) = $net->@{'bridge', 'tag', 'trunks'};
+    check_vnet_access($rpcenv, $authuser, $bridge, $tag, $trunks);
+
+    return 1;
+};
+
 sub umount_all {
     my ($vmid, $storage_cfg, $conf, $noerr) = @_;
 
@@ -1259,7 +1479,7 @@ sub umount_all {
 
     my $res = 1;
 
-    PVE::LXC::Config->foreach_mountpoint_reverse($conf, sub {
+    PVE::LXC::Config->foreach_volume_full($conf, {'reverse' => 1}, sub {
 	my ($ms, $mountpoint) = @_;
 
 	my $volid = $mountpoint->{volume};
@@ -1297,15 +1517,15 @@ sub mount_all {
     my $volid_list = PVE::LXC::Config->get_vm_volumes($conf);
     PVE::Storage::activate_volumes($storage_cfg, $volid_list);
 
-    my (undef, $rootuid, $rootgid) = parse_id_maps($conf);
+    my (undef, $root_uid, $root_gid) = parse_id_maps($conf);
 
     eval {
-	PVE::LXC::Config->foreach_mountpoint($conf, sub {
+	PVE::LXC::Config->foreach_volume($conf, sub {
 	    my ($ms, $mountpoint) = @_;
 
 	    $mountpoint->{ro} = 0 if $ignore_ro;
 
-	    mountpoint_mount($mountpoint, $rootdir, $storage_cfg, undef, $rootuid, $rootgid);
+	    mountpoint_mount($mountpoint, $rootdir, $storage_cfg, undef, $root_uid, $root_gid);
         });
     };
     if (my $err = $@) {
@@ -1381,17 +1601,17 @@ sub run_with_loopdev {
 #   * directory name of the last directory
 # So that the path $2/$3 should lead to $1 afterwards.
 sub walk_tree_nofollow($$$;$$) {
-    my ($start, $subdir, $mkdir, $rootuid, $rootgid) = @_;
+    my ($start, $subdir, $mkdir, $root_uid, $root_gid) = @_;
 
     sysopen(my $fd, $start, O_PATH | O_DIRECTORY)
 	or die "failed to open start directory $start: $!\n";
 
-    return walk_tree_nofollow_fd($start, $fd, $subdir, $mkdir, $rootuid, $rootgid);
+    return walk_tree_nofollow_fd($start, $fd, $subdir, $mkdir, $root_uid, $root_gid);
 }
 
 
 sub walk_tree_nofollow_fd($$$$;$$) {
-    my ($start_dirname, $start_fd, $subdir, $mkdir, $rootuid, $rootgid) = @_;
+    my ($start_dirname, $start_fd, $subdir, $mkdir, $root_uid, $root_gid) = @_;
 
     # splitdir() returns '' for empty components including the leading /
     my @comps = grep { length($_)>0 } File::Spec->splitdir($subdir);
@@ -1419,8 +1639,8 @@ sub walk_tree_nofollow_fd($$$$;$$) {
 	    $next = PVE::Tools::openat(fileno($fd), $component, O_NOFOLLOW | O_DIRECTORY);
 	    die "failed to create path: $dir: $!\n" if !$next;
 
-	    PVE::Tools::fchownat(fileno($next), '', $rootuid, $rootgid, PVE::Tools::AT_EMPTY_PATH)
-		if defined($rootuid) && defined($rootgid);
+	    PVE::Tools::fchownat(fileno($next), '', $root_uid, $root_gid, PVE::Tools::AT_EMPTY_PATH)
+		if defined($root_uid) && defined($root_gid);
 	}
 
 	close $second if defined($last_component) && $second != $start_fd;
@@ -1514,27 +1734,26 @@ sub bindmount {
 # from $rootdir and $mount and walk the path from $rootdir to the final
 # directory to check for symlinks.
 sub __mount_prepare_rootdir {
-    my ($rootdir, $mount, $rootuid, $rootgid) = @_;
+    my ($rootdir, $mount, $root_uid, $root_gid) = @_;
     $rootdir =~ s!/+!/!g;
     $rootdir =~ s!/+$!!;
     my $mount_path = "$rootdir/$mount";
-    my ($mpfd, $parentfd, $last_dir) = walk_tree_nofollow($rootdir, $mount, 1, $rootuid, $rootgid);
+    my ($mpfd, $parentfd, $last_dir) = walk_tree_nofollow($rootdir, $mount, 1, $root_uid, $root_gid);
     return ($rootdir, $mount_path, $mpfd, $parentfd, $last_dir);
 }
 
 # use $rootdir = undef to just return the corresponding mount path
 sub mountpoint_mount {
-    my ($mountpoint, $rootdir, $storage_cfg, $snapname, $rootuid, $rootgid) = @_;
-    return __mountpoint_mount($mountpoint, $rootdir, $storage_cfg, $snapname, $rootuid, $rootgid, undef);
+    my ($mountpoint, $rootdir, $storage_cfg, $snapname, $root_uid, $root_gid) = @_;
+    return __mountpoint_mount($mountpoint, $rootdir, $storage_cfg, $snapname, $root_uid, $root_gid, undef);
 }
 
 sub mountpoint_stage {
-    my ($mountpoint, $stage_dir, $storage_cfg, $snapname, $rootuid, $rootgid) = @_;
+    my ($mountpoint, $stage_dir, $storage_cfg, $snapname, $root_uid, $root_gid) = @_;
     my ($path, $loop, $dev) =
-	__mountpoint_mount($mountpoint, $stage_dir, $storage_cfg, $snapname, $rootuid, $rootgid, 1);
+	__mountpoint_mount($mountpoint, $stage_dir, $storage_cfg, $snapname, $root_uid, $root_gid, 1);
 
     if (!defined($path)) {
-	return undef if $! == ENOSYS;
 	die "failed to mount subvolume: $!\n";
     }
 
@@ -1549,14 +1768,14 @@ sub mountpoint_stage {
 }
 
 sub mountpoint_insert_staged {
-    my ($mount_fd, $rootdir_fd, $mp_dir, $opt, $rootuid, $rootgid) = @_;
+    my ($mount_fd, $rootdir_fd, $mp_dir, $opt, $root_uid, $root_gid) = @_;
 
     if (!defined($rootdir_fd)) {
 	sysopen($rootdir_fd, '.', O_PATH | O_DIRECTORY)
 	    or die "failed to open '.': $!\n";
     }
 
-    my $dest_fd = walk_tree_nofollow_fd('/', $rootdir_fd, $mp_dir, 1, $rootuid, $rootgid);
+    my $dest_fd = walk_tree_nofollow_fd('/', $rootdir_fd, $mp_dir, 1, $root_uid, $root_gid);
 
     PVE::Tools::move_mount(
 	fileno($mount_fd),
@@ -1569,15 +1788,8 @@ sub mountpoint_insert_staged {
 
 # Use $stage_mount, $rootdir is treated as a temporary path to "stage" the file system. The user
 #   can then open a file descriptor to it which can be used with the `move_mount` syscall.
-#   Note that if the kernel does not support the new mount API, this will not perform any action
-#   and return `undef` with $! = ENOSYS.
 sub __mountpoint_mount {
-    my ($mountpoint, $rootdir, $storage_cfg, $snapname, $rootuid, $rootgid, $stage_mount) = @_;
-
-    if (defined($stage_mount) && !PVE::LXC::Tools::can_use_new_mount_api()) {
-	$! = ENOSYS;
-	return undef;
-    }
+    my ($mountpoint, $rootdir, $storage_cfg, $snapname, $root_uid, $root_gid, $stage_mount) = @_;
 
     # When staging mount points we always mount to $rootdir directly (iow. as if `mp=/`).
     # This is required since __mount_prepare_rootdir() will return handles to the parent directory
@@ -1588,23 +1800,23 @@ sub __mountpoint_mount {
     my $type = $mountpoint->{type};
     my $quota = !$snapname && !$mountpoint->{ro} && $mountpoint->{quota};
     my $mounted_dev;
-    
+
     return if !$volid || !$mount;
 
     $mount =~ s!/+!/!g;
 
     my $mount_path;
     my ($mpfd, $parentfd, $last_dir);
-    
+
     if (defined($rootdir)) {
 	($rootdir, $mount_path, $mpfd, $parentfd, $last_dir) =
-	    __mount_prepare_rootdir($rootdir, $mount, $rootuid, $rootgid);
+	    __mount_prepare_rootdir($rootdir, $mount, $root_uid, $root_gid);
     }
 
     if (defined($stage_mount)) {
 	$mount_path = $rootdir;
     }
-    
+
     my ($storage, $volname) = PVE::Storage::parse_volume_id($volid, 1);
 
     die "unknown snapshot path for '$volid'" if !$storage && defined($snapname);
@@ -1617,8 +1829,10 @@ sub __mountpoint_mount {
     }
 
     my $acl = $mountpoint->{acl};
-    if (defined($acl)) {
-	push @$optlist, ($acl ? 'acl' : 'noacl');
+
+    if ($acl) {
+	push @$optlist, 'acl';
+	# NOTE: the else branch is handled below
     }
 
     my $optstring = join(',', @$optlist);
@@ -1631,6 +1845,7 @@ sub __mountpoint_mount {
 
 	my $scfg = PVE::Storage::storage_config($storage_cfg, $storage);
 
+	PVE::Storage::activate_volumes($storage_cfg, [$volid], $snapname);
 	my $path = PVE::Storage::map_volume($storage_cfg, $volid, $snapname);
 
 	$path = PVE::Storage::path($storage_cfg, $volid, $snapname) if !defined($path);
@@ -1638,6 +1853,12 @@ sub __mountpoint_mount {
 	my ($vtype, undef, undef, undef, undef, $isBase, $format) =
 	    PVE::Storage::parse_volname($storage_cfg, $volid);
 
+	if (defined($acl) && !$acl) {
+	    # Does having this really makes sense or should we drop it with a future major release?
+	    # Kernel 6.1 removed the noacl mount option for ext4, which is used for all raw volumes.
+	    push @$optlist, 'noacl' if $format ne 'raw';
+	}
+
 	$format = 'iso' if $vtype eq 'iso'; # allow to handle iso files
 
 	if ($format eq 'subvol') {
@@ -1685,15 +1906,16 @@ sub __mountpoint_mount {
 		}
 	    };
 	    my $use_loopdev = 0;
-	    if ($scfg->{path}) {
-		$mounted_dev = run_with_loopdev($domount, $path, $readonly);
-		$use_loopdev = 1;
-	    } elsif ($scfg->{type} eq 'drbd' || $scfg->{type} eq 'lvm' ||
-		     $scfg->{type} eq 'rbd' || $scfg->{type} eq 'lvmthin') {
-		$mounted_dev = $path;
-		&$domount($path);
+	    if ($scfg->{content}->{rootdir}) {
+		if ($scfg->{path}) {
+		    $mounted_dev = run_with_loopdev($domount, $path, $readonly);
+		    $use_loopdev = 1;
+		} else {
+		    $mounted_dev = $path;
+		    &$domount($path);
+		}
 	    } else {
-		die "unsupported storage type '$scfg->{type}'\n";
+		die "storage '$storage' does not support containers\n";
 	    }
 	    return wantarray ? ($path, $use_loopdev, $mounted_dev) : $path;
 	} else {
@@ -1712,14 +1934,14 @@ sub __mountpoint_mount {
 	warn "cannot enable quota control for bind mounts\n" if $quota;
 	return wantarray ? ($volid, 0, undef) : $volid;
     }
-    
+
     die "unsupported storage";
 }
 
-sub mountpoint_hotplug($$$) {
+sub mountpoint_hotplug :prototype($$$$$) {
     my ($vmid, $conf, $opt, $mp, $storage_cfg) = @_;
 
-    my (undef, $rootuid, $rootgid) = PVE::LXC::parse_id_maps($conf);
+    my (undef, $root_uid, $root_gid) = PVE::LXC::parse_id_maps($conf);
 
     # We do the rest in a fork with an unshared mount namespace, because:
     #  -) change our papparmor profile to that of /usr/bin/lxc-start
@@ -1755,21 +1977,22 @@ sub mountpoint_hotplug($$$) {
 	my $dir = get_staging_mount_path($opt);
 
 	# Now switch our apparmor profile before mounting:
-	my $data = 'changeprofile /usr/bin/lxc-start';
-	if (syswrite($aa_fd, $data, length($data)) != length($data)) {
+	my $data = 'changeprofile pve-container-mounthotplug';
+	my $data_written = syswrite($aa_fd, $data, length($data));
+	if (!defined($data_written) || $data_written != length($data)) {
 	    die "failed to change apparmor profile: $!\n";
 	}
 	# Check errors on close as well:
 	close($aa_fd)
 	    or die "failed to change apparmor profile (close() failed): $!\n";
 
-	my $mount_fd = mountpoint_stage($mp, $dir, $storage_cfg, undef, $rootuid, $rootgid);
+	my $mount_fd = mountpoint_stage($mp, $dir, $storage_cfg, undef, $root_uid, $root_gid);
 
 	PVE::Tools::setns(fileno($ct_mnt_ns), PVE::Tools::CLONE_NEWNS);
 	chdir('/')
 	    or die "failed to change root directory within the container's mount namespace: $!\n";
 
-	mountpoint_insert_staged($mount_fd, undef, $mp->{mp}, $opt, $rootuid, $rootgid);
+	mountpoint_insert_staged($mount_fd, undef, $mp->{mp}, $opt, $root_uid, $root_gid);
     });
 }
 
@@ -1785,7 +2008,7 @@ sub get_staging_mount_path($) {
     return $target;
 }
 
-# Mount /run/pve/mountpoints as tmpfs
+# Mount tmpfs for mount point staging and return the path.
 sub get_staging_tempfs() {
     # We choose a path in /var/lib/lxc/ here because the lxc-start apparmor profile restricts most
     # mounts to that.
@@ -1802,18 +2025,38 @@ sub get_staging_tempfs() {
 }
 
 sub mkfs {
-    my ($dev, $rootuid, $rootgid) = @_;
-
-    PVE::Tools::run_command(['mkfs.ext4', '-O', 'mmp',
-			     '-E', "root_owner=$rootuid:$rootgid",
-			     $dev]);
+    my ($dev, $root_uid, $root_gid) = @_;
+
+    run_command(
+	[
+	    'mkfs.ext4',
+	    '-O',
+	    'mmp',
+	    '-E',
+	    "root_owner=$root_uid:$root_gid",
+	    $dev,
+	],
+	outfunc => sub {
+	    my $line = shift;
+	    # a hack to print only the relevant stuff, i.e., the one which could help on repair
+	    if ($line =~ /^(Creating filesystem|Filesystem UUID|Superblock backups|\s+\d+, \d)/) {
+		print "$line\n";
+	    }
+	},
+	errfunc => sub {
+	    my $line = shift;
+	    print STDERR "$line\n" if $line && $line !~ /^mke2fs \d\.\d/;
+	}
+    );
 }
 
 sub format_disk {
-    my ($storage_cfg, $volid, $rootuid, $rootgid) = @_;
+    my ($storage_cfg, $volid, $root_uid, $root_gid) = @_;
 
     if ($volid =~ m!^/dev/.+!) {
-	mkfs($volid);
+	# FIXME: remove in Proxmox VE 9 – this code path cannot really be reached currently, using
+	# block devices needs manual preparations by the user
+	mkfs($volid, $root_uid, $root_gid);
 	return;
     }
 
@@ -1833,7 +2076,7 @@ sub format_disk {
     die "cannot format volume '$volid' (format == $format)\n"
 	if $format ne 'raw';
 
-    mkfs($path, $rootuid, $rootgid);
+    mkfs($path, $root_uid, $root_gid);
 }
 
 sub destroy_disks {
@@ -1846,7 +2089,7 @@ sub destroy_disks {
 }
 
 sub alloc_disk {
-    my ($storecfg, $vmid, $storage, $size_kb, $rootuid, $rootgid) = @_;
+    my ($storecfg, $vmid, $storage, $size_kb, $root_uid, $root_gid) = @_;
 
     my $needs_chown = 0;
     my $volid;
@@ -1856,34 +2099,24 @@ sub alloc_disk {
 
     eval {
 	my $do_format = 0;
-	if ($scfg->{type} eq 'dir' || $scfg->{type} eq 'nfs' || $scfg->{type} eq 'cifs' ) {
-	    if ($size_kb > 0) {
-		$volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'raw',
-						   undef, $size_kb);
+	if ($scfg->{content}->{rootdir} && $scfg->{path}) {
+	    if ($size_kb > 0 && !($scfg->{type} eq 'btrfs' && $scfg->{quotas})) {
+		$volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'raw', undef, $size_kb);
 		$do_format = 1;
 	    } else {
-		$volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'subvol',
-						   undef, 0);
+		$volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'subvol', undef, $size_kb);
 		$needs_chown = 1;
 	    }
 	} elsif ($scfg->{type} eq 'zfspool') {
-
-	    $volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'subvol',
-					       undef, $size_kb);
+	    $volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'subvol', undef, $size_kb);
 	    $needs_chown = 1;
-	} elsif ($scfg->{type} eq 'drbd' || $scfg->{type} eq 'lvm' || $scfg->{type} eq 'lvmthin') {
-
-	    $volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'raw', undef, $size_kb);
-	    $do_format = 1;
-
-	} elsif ($scfg->{type} eq 'rbd') {
-
+	} elsif ($scfg->{content}->{rootdir}) {
 	    $volid = PVE::Storage::vdisk_alloc($storecfg, $storage, $vmid, 'raw', undef, $size_kb);
 	    $do_format = 1;
 	} else {
-	    die "unable to create containers on storage type '$scfg->{type}'\n";
+	    die "content type 'rootdir' is not available or configured on storage '$storage'\n";
 	}
-	format_disk($storecfg, $volid, $rootuid, $rootgid) if $do_format;
+	format_disk($storecfg, $volid, $root_uid, $root_gid) if $do_format;
     };
     if (my $err = $@) {
 	# in case formatting got interrupted:
@@ -1897,17 +2130,16 @@ sub alloc_disk {
     return ($volid, $needs_chown);
 }
 
-our $NEW_DISK_RE = qr/^([^:\s]+):(\d+(\.\d+)?)$/;
 sub create_disks {
     my ($storecfg, $vmid, $settings, $conf, $pending) = @_;
 
     my $vollist = [];
 
     eval {
-	my (undef, $rootuid, $rootgid) = PVE::LXC::parse_id_maps($conf);
+	my (undef, $root_uid, $root_gid) = PVE::LXC::parse_id_maps($conf);
 	my $chown_vollist = [];
 
-	PVE::LXC::Config->foreach_mountpoint($settings, sub {
+	PVE::LXC::Config->foreach_volume($settings, sub {
 	    my ($ms, $mountpoint) = @_;
 
 	    my $volid = $mountpoint->{volume};
@@ -1921,7 +2153,7 @@ sub create_disks {
 		my $size_kb = int(${size_gb}*1024) * 1024;
 
 		my $needs_chown = 0;
-		($volid, $needs_chown) = alloc_disk($storecfg, $vmid, $storage, $size_kb, $rootuid, $rootgid);
+		($volid, $needs_chown) = alloc_disk($storecfg, $vmid, $storage, $size_kb, $root_uid, $root_gid);
 		push @$chown_vollist, $volid if $needs_chown;
 		push @$vollist, $volid;
 		$mountpoint->{volume} = $volid;
@@ -1940,7 +2172,7 @@ sub create_disks {
 	PVE::Storage::activate_volumes($storecfg, $chown_vollist, undef);
 	foreach my $volid (@$chown_vollist) {
 	    my $path = PVE::Storage::path($storecfg, $volid, undef);
-	    chown($rootuid, $rootgid, $path);
+	    chown($root_uid, $root_gid, $path);
 	}
 	PVE::Storage::deactivate_volumes($storecfg, $chown_vollist, undef);
     };
@@ -1966,12 +2198,12 @@ sub update_disksize {
 	    $changes = 1;
 	    print "$prefix updated volume size of '$mp->{volume}' in config.\n";
 	    $mp->{size} = $size;
-	    my $nomp = 1 if ($key eq 'rootfs');
-	    $conf->{$key} = PVE::LXC::Config->print_ct_mountpoint($mp, $nomp);
+	    my $no_mp = $key eq 'rootfs'; # rootfs is handled different from other mount points
+	    $conf->{$key} = PVE::LXC::Config->print_ct_mountpoint($mp, $no_mp);
 	}
     };
 
-    PVE::LXC::Config->foreach_mountpoint($conf, $update_mp);
+    PVE::LXC::Config->foreach_volume($conf, $update_mp);
 
     return $changes;
 }
@@ -2027,7 +2259,7 @@ sub update_unused {
 sub scan_volids {
     my ($cfg, $vmid) = @_;
 
-    my $info = PVE::Storage::vdisk_list($cfg, undef, $vmid);
+    my $info = PVE::Storage::vdisk_list($cfg, undef, $vmid, undef, 'rootdir');
 
     my $all_volumes = {};
     foreach my $storeid (keys %$info) {
@@ -2047,12 +2279,6 @@ sub rescan {
 
     my $cfg = PVE::Storage::config();
 
-    # FIXME: Remove once our RBD plugin can handle CT and VM on a single storage
-    # see: https://pve.proxmox.com/pipermail/pve-devel/2018-July/032900.html
-    foreach my $stor (keys %{$cfg->{ids}}) {
-	delete($cfg->{ids}->{$stor}) if !$cfg->{ids}->{$stor}->{content}->{rootdir};
-    }
-
     print "rescan volumes...\n";
     my $all_volumes = scan_volids($cfg, $vmid);
 
@@ -2160,20 +2386,21 @@ sub parse_id_maps {
     my ($conf) = @_;
 
     my $id_map = [];
-    my $rootuid = 0;
-    my $rootgid = 0;
+    my $root_uid = 0;
+    my $root_gid = 0;
 
     my $lxc = $conf->{lxc};
     foreach my $entry (@$lxc) {
 	my ($key, $value) = @$entry;
-	# FIXME: remove the 'id_map' variant when lxc-3.0 arrives
-	next if $key ne 'lxc.idmap' && $key ne 'lxc.id_map';
+
+	next if $key ne 'lxc.idmap';
+
 	if ($value =~ /^([ug])\s+(\d+)\s+(\d+)\s+(\d+)\s*$/) {
 	    my ($type, $ct, $host, $length) = ($1, $2, $3, $4);
 	    push @$id_map, [$type, $ct, $host, $length];
 	    if ($ct == 0) {
-		$rootuid = $host if $type eq 'u';
-		$rootgid = $host if $type eq 'g';
+		$root_uid = $host if $type eq 'u';
+		$root_gid = $host if $type eq 'g';
 	    }
 	} else {
 	    die "failed to parse idmap: $value\n";
@@ -2184,10 +2411,73 @@ sub parse_id_maps {
 	# Should we read them from /etc/subuid?
 	$id_map = [ ['u', '0', '100000', '65536'],
 	            ['g', '0', '100000', '65536'] ];
-	$rootuid = $rootgid = 100000;
+	$root_uid = $root_gid = 100000;
+    }
+
+    return ($id_map, $root_uid, $root_gid);
+}
+
+sub validate_id_maps {
+    my ($id_map) = @_;
+
+    # $mappings->{$type}->{$side} = [ { line => $line, start => $start, count => $count }, ... ]
+    #   $type: either "u" or "g"
+    #   $side: either "container" or "host"
+    #   $line: index of this mapping in @$id_map
+    #   $start, $count: interval of this mapping
+    my $mappings = { u => {}, g => {} };
+    for (my $i = 0; $i < scalar(@$id_map); $i++) {
+	my ($type, $ct_start, $host_start, $count) = $id_map->[$i]->@*;
+	my $sides = $mappings->{$type};
+	push $sides->{host}->@*, { line => $i, start => $host_start, count => $count };
+	push $sides->{container}->@*, { line => $i, start => $ct_start, count => $count };
+    }
+
+    # find the first conflict between two consecutive mappings when sorted by their start id
+    for my $type (qw(u g)) {
+	for my $side (qw(container host)) {
+	    my @entries = sort { $a->{start} <=> $b->{start} } $mappings->{$type}->{$side}->@*;
+	    for my $idx (1..scalar(@entries) - 1) {
+		my $previous = $entries[$idx - 1];
+		my $current = $entries[$idx];
+		if ($previous->{start} + $previous->{count} > $current->{start}) {
+		    my $conflict = $current->{start};
+		    my @previous_line = $id_map->[$previous->{line}]->@*;
+		    my @current_line = $id_map->[$current->{line}]->@*;
+		    die "invalid map entry '@current_line': $side ${type}id $conflict "
+		       ."is also mapped by entry '@previous_line'\n";
+		}
+	    }
+	}
     }
+}
+
+sub map_ct_id_to_host {
+    my ($id, $id_map, $id_type) = @_;
 
-    return ($id_map, $rootuid, $rootgid);
+    for my $mapping (@$id_map) {
+	my ($type, $ct, $host, $length) = @$mapping;
+
+	next if ($type ne $id_type);
+
+	if ($id >= $ct && $id < ($ct + $length)) {
+	    return $host - $ct + $id;
+	}
+    }
+
+    return $id;
+}
+
+sub map_ct_uid_to_host {
+    my ($uid, $id_map) = @_;
+
+    return map_ct_id_to_host($uid, $id_map, 'u');
+}
+
+sub map_ct_gid_to_host {
+    my ($gid, $id_map) = @_;
+
+    return map_ct_id_to_host($gid, $id_map, 'g');
 }
 
 sub userns_command {
@@ -2198,18 +2488,86 @@ sub userns_command {
     return [];
 }
 
+my sub print_ct_stderr_log {
+    my ($vmid) = @_;
+    my $log = eval { file_get_contents("/run/pve/ct-$vmid.stderr") };
+    return if !$log;
+
+    while ($log =~ /^\h*(lxc-start:?\s+$vmid:?\s*\S+\s*)?(.*?)\h*$/gm) {
+	my $line = $2;
+	print STDERR "$line\n";
+    }
+}
+my sub print_ct_warn_log {
+    my ($vmid) = @_;
+    my $log_fn = "/run/pve/ct-$vmid.warnings";
+    my $log = eval { file_get_contents($log_fn) };
+    return if !$log;
+
+    while ($log =~ /^\h*\s*(.*?)\h*$/gm) {
+	PVE::RESTEnvironment::log_warn($1);
+    }
+    unlink $log_fn or warn "could not unlink '$log_fn' - $!\n";
+}
+
+my sub monitor_state_change($$) {
+    my ($monitor_socket, $vmid) = @_;
+    die "no monitor socket\n" if !defined($monitor_socket);
+
+    while (1) {
+	my ($type, $name, $value) = PVE::LXC::Monitor::read_lxc_message($monitor_socket);
+
+	die "monitor socket: got EOF\n" if !defined($type);
+
+	next if $name ne "$vmid" || $type ne 'STATE';
+
+	if ($value eq PVE::LXC::Monitor::STATE_STARTING) {
+	    alarm(0); # don't timeout after seeing the starting state
+	} elsif ($value eq PVE::LXC::Monitor::STATE_ABORTING ||
+		 $value eq PVE::LXC::Monitor::STATE_STOPPING ||
+		 $value eq PVE::LXC::Monitor::STATE_STOPPED) {
+	    return 0;
+	} elsif ($value eq PVE::LXC::Monitor::STATE_RUNNING) {
+	    return 1;
+	} else {
+	    warn "unexpected message from monitor socket - " .
+		 "type: '$type' - value: '$value'\n";
+	}
+    }
+}
+my sub monitor_start($$) {
+    my ($monitor_socket, $vmid) = @_;
+
+    my $success = eval {
+	PVE::Tools::run_with_timeout(10, \&monitor_state_change, $monitor_socket, $vmid)
+    };
+    if (my $err = $@) {
+	warn "problem with monitor socket, but continuing anyway: $err\n";
+    } elsif (!$success) {
+	print_ct_stderr_log($vmid);
+	die "startup for container '$vmid' failed\n";
+    }
+}
+
 sub vm_start {
-    my ($vmid, $conf, $skiplock) = @_;
+    my ($vmid, $conf, $skiplock, $debug) = @_;
 
     # apply pending changes while starting
     if (scalar(keys %{$conf->{pending}})) {
 	my $storecfg = PVE::Storage::config();
 	PVE::LXC::Config->vmconfig_apply_pending($vmid, $conf, $storecfg);
+	PVE::LXC::Config->write_config($vmid, $conf);
 	$conf = PVE::LXC::Config->load_config($vmid); # update/reload
     }
 
     update_lxc_config($vmid, $conf);
 
+    eval {
+	my ($id_map, undef, undef) = PVE::LXC::parse_id_maps($conf);
+	PVE::LXC::validate_id_maps($id_map);
+    };
+    warn "lxc.idmap: $@" if $@;
+
     my $skiplock_flag_fn = "/run/lxc/skiplock-$vmid";
 
     if ($skiplock) {
@@ -2217,10 +2575,32 @@ sub vm_start {
 	close($fh);
     }
 
-    my $cmd = ['systemctl', 'start', "pve-container\@$vmid"];
+    my $storage_cfg = PVE::Storage::config();
+    my $vollist = PVE::LXC::Config->get_vm_volumes($conf);
+
+    PVE::Storage::activate_volumes($storage_cfg, $vollist);
+
+    my $monitor_socket = eval { PVE::LXC::Monitor::get_monitor_socket() };
+    warn $@ if $@;
+
+    unlink "/run/pve/ct-$vmid.stderr"; # systemd does not truncate log files
+
+    my $is_debug = $debug || (!defined($debug) && $conf->{debug});
+    my $base_unit = $is_debug ? 'pve-container-debug' : 'pve-container';
+
+    my $cmd = ['systemctl', 'start', "$base_unit\@$vmid"];
 
     PVE::GuestHelpers::exec_hookscript($conf, $vmid, 'pre-start', 1);
-    eval { PVE::Tools::run_command($cmd); };
+    eval {
+	run_command($cmd);
+
+	monitor_start($monitor_socket, $vmid) if defined($monitor_socket);
+
+	# if debug is requested, print the log it also when the start succeeded
+	print_ct_stderr_log($vmid) if $is_debug;
+
+	print_ct_warn_log($vmid); # always print warn log, if any
+    };
     if (my $err = $@) {
 	unlink $skiplock_flag_fn;
 	die $err;
@@ -2271,14 +2651,23 @@ sub vm_stop {
 	}
     }
 
-    eval { PVE::Tools::run_command($cmd, timeout => $shutdown_timeout) };
+    eval { run_command($cmd, timeout => $shutdown_timeout) };
+
+    # Wait until the command socket is closed.
+    # In case the lxc-stop call failed, reading from the command socket may block forever,
+    # so poll with another timeout to avoid freezing the shutdown task.
     if (my $err = $@) {
-	warn $@ if $@;
-    }
+	warn $err if $err;
 
-    my $result = <$sock>;
+	my $poll = IO::Poll->new();
+	$poll->mask($sock => POLLIN | POLLHUP); # watch for input and EOF events
+	$poll->poll($shutdown_timeout); # IO::Poll timeout is in seconds
+	return if ($poll->events($sock) & POLLHUP);
+    } else {
+	my $result = <$sock>;
+	return if !defined $result; # monitor is gone and the ct has stopped.
+    }
 
-    return if !defined $result; # monitor is gone and the ct has stopped.
     die "container did not stop\n";
 }
 
@@ -2302,13 +2691,13 @@ sub run_unshared {
 	# Unshare the mount namespace
 	die "failed to unshare mount namespace: $!\n"
 	    if !PVE::Tools::unshare(PVE::Tools::CLONE_NEWNS);
-	PVE::Tools::run_command(['mount', '--make-rslave', '/']);
+	run_command(['mount', '--make-rslave', '/']);
 	return $code->();
     });
 }
 
 my $copy_volume = sub {
-    my ($src_volid, $src, $dst_volid, $dest, $storage_cfg, $snapname, $bwlimit, $rootuid,  $rootgid) = @_;
+    my ($src_volid, $src, $dst_volid, $dest, $storage_cfg, $snapname, $bwlimit, $root_uid,  $root_gid) = @_;
 
     my $src_mp = { volume => $src_volid, mp => '/', ro => 1 };
     $src_mp->{type} = PVE::LXC::Config->classify_mountpoint($src_volid);
@@ -2320,17 +2709,28 @@ my $copy_volume = sub {
     eval {
 	# mount and copy
 	mkdir $src;
-	mountpoint_mount($src_mp, $src, $storage_cfg, $snapname, $rootuid, $rootgid);
+	mountpoint_mount($src_mp, $src, $storage_cfg, $snapname, $root_uid, $root_gid);
 	push @mounted, $src;
 	mkdir $dest;
-	mountpoint_mount($dst_mp, $dest, $storage_cfg, undef, $rootuid, $rootgid);
+	mountpoint_mount($dst_mp, $dest, $storage_cfg, undef, $root_uid, $root_gid);
 	push @mounted, $dest;
 
 	$bwlimit //= 0;
 
-	PVE::Tools::run_command(['/usr/bin/rsync', '--stats', '-X', '-A', '--numeric-ids',
-				 '-aH', '--whole-file', '--sparse', '--one-file-system',
-				 "--bwlimit=$bwlimit", "$src/", $dest]);
+	run_command([
+	    'rsync',
+	    '--stats',
+	    '-X',
+	    '-A',
+	    '--numeric-ids',
+	    '-aH',
+	    '--whole-file',
+	    '--sparse',
+	    '--one-file-system',
+	    "--bwlimit=$bwlimit",
+	    "$src/",
+	    $dest
+	]);
     };
     my $err = $@;
 
@@ -2339,7 +2739,7 @@ my $copy_volume = sub {
     while ((system {"fuser"} "fuser",  "-s", $dest) == 0) {sleep 1};
 
     foreach my $mount (reverse @mounted) {
-	eval { PVE::Tools::run_command(['/bin/umount', $mount], errfunc => sub{})};
+	eval { run_command(['/bin/umount', $mount], errfunc => sub{})};
 	warn "Can't umount $mount\n" if $@;
     }
 
@@ -2361,7 +2761,7 @@ sub copy_volume {
     my $src  = "/var/lib/lxc/$vmid/.copy-volume-2";
 
     # get id's for unprivileged container
-    my (undef, $rootuid, $rootgid) = parse_id_maps($conf);
+    my (undef, $root_uid, $root_gid) = parse_id_maps($conf);
 
     # Allocate the disk before unsharing in order to make sure zfs subvolumes
     # are visible in this namespace, otherwise the host only sees the empty
@@ -2371,15 +2771,15 @@ sub copy_volume {
 	# Make sure $mp contains a correct size.
 	$mp->{size} = PVE::Storage::volume_size_info($storage_cfg, $mp->{volume});
 	my $needs_chown;
-	($new_volid, $needs_chown) = alloc_disk($storage_cfg, $vmid, $storage, $mp->{size}/1024, $rootuid, $rootgid);
+	($new_volid, $needs_chown) = alloc_disk($storage_cfg, $vmid, $storage, $mp->{size}/1024, $root_uid, $root_gid);
 	if ($needs_chown) {
 	    PVE::Storage::activate_volumes($storage_cfg, [$new_volid], undef);
 	    my $path = PVE::Storage::path($storage_cfg, $new_volid, undef);
-	    chown($rootuid, $rootgid, $path);
+	    chown($root_uid, $root_gid, $path);
 	}
 
 	run_unshared(sub {
-	    $copy_volume->($mp->{volume}, $src, $new_volid, $dest, $storage_cfg, $snapname, $bwlimit, $rootuid, $rootgid);
+	    $copy_volume->($mp->{volume}, $src, $new_volid, $dest, $storage_cfg, $snapname, $bwlimit, $root_uid, $root_gid);
 	});
     };
     if (my $err = $@) {
@@ -2391,4 +2791,64 @@ sub copy_volume {
     return $new_volid;
 }
 
+sub get_lxc_version() {
+    my $version;
+    run_command([qw(lxc-start --version)], outfunc => sub {
+	my ($line) = @_;
+	# We only parse out major & minor version numbers.
+	if ($line =~ /^(\d+)\.(\d+)(?:\D.*)?$/) {
+	    $version = [$1, $2];
+	}
+    });
+
+    die "failed to get lxc version\n" if !defined($version);
+
+    # return as a list:
+    return $version->@*;
+}
+
+sub freeze($) {
+    my ($vmid) = @_;
+    if (PVE::CGroup::cgroup_mode() == 2) {
+	PVE::LXC::Command::freeze($vmid, 30);
+    } else {
+	PVE::LXC::CGroup->new($vmid)->freeze_thaw(1);
+    }
+}
+
+sub thaw($) {
+    my ($vmid) = @_;
+    if (PVE::CGroup::cgroup_mode() == 2) {
+	PVE::LXC::Command::unfreeze($vmid, 30);
+    } else {
+	PVE::LXC::CGroup->new($vmid)->freeze_thaw(0);
+    }
+}
+
+sub create_ifaces_ipams_ips {
+    my ($conf, $vmid) = @_;
+
+    return if !$have_sdn;
+
+    for my $opt (keys %$conf) {
+	next if $opt !~ m/^net(\d+)$/;
+	my $net = PVE::LXC::Config->parse_lxc_network($conf->{$opt});
+	next if $net->{type} ne 'veth';
+	PVE::Network::SDN::Vnets::add_next_free_cidr($net->{bridge}, $conf->{hostname}, $net->{hwaddr}, $vmid, undef, 1);
+    }
+}
+
+sub delete_ifaces_ipams_ips {
+    my ($conf, $vmid) = @_;
+
+    return if !$have_sdn;
+
+    for my $opt (keys %$conf) {
+	next if $opt !~ m/^net(\d+)$/;
+	my $net = PVE::LXC::Config->parse_lxc_network($conf->{$opt});
+	eval { PVE::Network::SDN::Vnets::del_ips_from_mac($net->{bridge}, $net->{hwaddr}, $conf->{hostname}) };
+	warn $@ if $@;
+    }
+}
+
 1;