use strict;
use warnings;
-use POSIX qw(strftime);
+
use IO::File;
use IPC::Open2;
-use PVE::Tools qw(run_command);
-use PVE::SafeSyslog;
-use PVE::INotify;
+use POSIX qw( WNOHANG );
+use Time::HiRes qw( usleep );
+
use PVE::Cluster;
+use PVE::INotify;
+use PVE::RPCEnvironment;
+use PVE::Replication;
+use PVE::ReplicationConfig;
+use PVE::ReplicationState;
use PVE::Storage;
+use PVE::Tools;
+
+use PVE::QemuConfig;
+use PVE::QemuServer::CPUConfig;
+use PVE::QemuServer::Drive;
+use PVE::QemuServer::Helpers qw(min_version);
+use PVE::QemuServer::Machine;
+use PVE::QemuServer::Monitor qw(mon_cmd);
use PVE::QemuServer;
-my $delayed_interrupt = 0;
-
-# blowfish is a fast block cipher, much faster then 3des
-my @ssh_opts = ('-c', 'blowfish', '-o', 'BatchMode=yes');
-my @ssh_cmd = ('/usr/bin/ssh', @ssh_opts);
-my @scp_cmd = ('/usr/bin/scp', @ssh_opts);
-my $qm_cmd = '/usr/sbin/qm';
-
-sub logmsg {
- my ($level, $msg) = @_;
-
- chomp $msg;
-
- return if !$msg;
-
- my $tstr = strftime("%b %d %H:%M:%S", localtime);
-
- syslog($level, $msg);
-
- foreach my $line (split (/\n/, $msg)) {
- print STDOUT "$tstr $line\n";
- }
- \*STDOUT->flush();
-}
-
-sub eval_int {
- my ($func) = @_;
-
- eval {
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = sub {
- $delayed_interrupt = 0;
- logmsg('err', "received interrupt");
- die "interrupted by signal\n";
- };
- local $SIG{PIPE} = sub {
- $delayed_interrupt = 0;
- logmsg('err', "received broken pipe interrupt");
- die "interrupted by signal\n";
- };
-
- my $di = $delayed_interrupt;
- $delayed_interrupt = 0;
-
- die "interrupted by signal\n" if $di;
-
- &$func();
- };
-}
+use PVE::AbstractMigrate;
+use base qw(PVE::AbstractMigrate);
sub fork_command_pipe {
- my ($cmd) = @_;
+ my ($self, $cmd) = @_;
my $reader = IO::File->new();
my $writer = IO::File->new();
# catch exec errors
if ($orig_pid != $$) {
- logmsg('err', "can't fork command pipe\n");
+ $self->log('err', "can't fork command pipe\n");
POSIX::_exit(1);
kill('KILL', $$);
}
}
sub finish_command_pipe {
- my $cmdpipe = shift;
+ my ($self, $cmdpipe, $timeout) = @_;
+
+ my $cpid = $cmdpipe->{pid};
+ return if !defined($cpid);
my $writer = $cmdpipe->{writer};
my $reader = $cmdpipe->{reader};
$writer->close();
$reader->close();
- my $cpid = $cmdpipe->{pid};
+ my $collect_child_process = sub {
+ my $res = waitpid($cpid, WNOHANG);
+ if (defined($res) && ($res == $cpid)) {
+ delete $cmdpipe->{cpid};
+ return 1;
+ } else {
+ return 0;
+ }
+ };
- kill(15, $cpid) if kill(0, $cpid);
+ if ($timeout) {
+ for (my $i = 0; $i < $timeout; $i++) {
+ return if &$collect_child_process();
+ sleep(1);
+ }
+ }
- waitpid($cpid, 0);
-}
+ $self->log('info', "ssh tunnel still running - terminating now with SIGTERM\n");
+ kill(15, $cpid);
-sub run_with_timeout {
- my ($timeout, $code, @param) = @_;
+ # wait again
+ for (my $i = 0; $i < 10; $i++) {
+ return if &$collect_child_process();
+ sleep(1);
+ }
- die "got timeout\n" if $timeout <= 0;
+ $self->log('info', "ssh tunnel still running - terminating now with SIGKILL\n");
+ kill 9, $cpid;
+ sleep 1;
+
+ $self->log('err', "ssh tunnel child process (PID $cpid) couldn't be collected\n")
+ if !&$collect_child_process();
+}
- my $prev_alarm;
+sub read_tunnel {
+ my ($self, $tunnel, $timeout) = @_;
- my $sigcount = 0;
+ $timeout = 60 if !defined($timeout);
- my $res;
+ my $reader = $tunnel->{reader};
+ my $output;
eval {
- local $SIG{ALRM} = sub { $sigcount++; die "got timeout\n"; };
- local $SIG{PIPE} = sub { $sigcount++; die "broken pipe\n" };
- local $SIG{__DIE__}; # see SA bug 4631
+ PVE::Tools::run_with_timeout($timeout, sub { $output = <$reader>; });
+ };
+ die "reading from tunnel failed: $@\n" if $@;
- $prev_alarm = alarm($timeout);
+ chomp $output;
- $res = &$code(@param);
+ return $output;
+}
- alarm(0); # avoid race conditions
- };
+sub write_tunnel {
+ my ($self, $tunnel, $timeout, $command) = @_;
- my $err = $@;
+ $timeout = 60 if !defined($timeout);
- alarm($prev_alarm) if defined($prev_alarm);
+ my $writer = $tunnel->{writer};
- die "unknown error" if $sigcount && !$err; # seems to happen sometimes
+ eval {
+ PVE::Tools::run_with_timeout($timeout, sub {
+ print $writer "$command\n";
+ $writer->flush();
+ });
+ };
+ die "writing to tunnel failed: $@\n" if $@;
- die $err if $err;
+ if ($tunnel->{version} && $tunnel->{version} >= 1) {
+ my $res = eval { $self->read_tunnel($tunnel, 10); };
+ die "no reply to command '$command': $@\n" if $@;
- return $res;
+ if ($res eq 'OK') {
+ return;
+ } else {
+ die "tunnel replied '$res' to command '$command'\n";
+ }
+ }
}
sub fork_tunnel {
- my ($nodeip, $lport, $rport) = @_;
+ my ($self, $tunnel_addr) = @_;
- my $cmd = [@ssh_cmd, '-o', 'BatchMode=yes',
- '-L', "$lport:localhost:$rport", $nodeip,
- 'qm', 'mtunnel' ];
+ my @localtunnelinfo = ();
+ foreach my $addr (@$tunnel_addr) {
+ push @localtunnelinfo, '-L', $addr;
+ }
- my $tunnel = fork_command_pipe($cmd);
+ my $cmd = [@{$self->{rem_ssh}}, '-o ExitOnForwardFailure=yes', @localtunnelinfo, '/usr/sbin/qm', 'mtunnel' ];
- my $reader = $tunnel->{reader};
+ my $tunnel = $self->fork_command_pipe($cmd);
- my $helo;
eval {
- run_with_timeout(60, sub { $helo = <$reader>; });
+ my $helo = $self->read_tunnel($tunnel, 60);
die "no reply\n" if !$helo;
die "no quorum on target node\n" if $helo =~ m/^no quorum$/;
die "got strange reply from mtunnel ('$helo')\n"
};
my $err = $@;
+ eval {
+ my $ver = $self->read_tunnel($tunnel, 10);
+ if ($ver =~ /^ver (\d+)$/) {
+ $tunnel->{version} = $1;
+ $self->log('info', "ssh tunnel $ver\n");
+ } else {
+ $err = "received invalid tunnel version string '$ver'\n" if !$err;
+ }
+ };
+
if ($err) {
- finish_command_pipe($tunnel);
+ $self->finish_command_pipe($tunnel);
die "can't open migration tunnel - $err";
}
return $tunnel;
}
sub finish_tunnel {
- my $tunnel = shift;
-
- my $writer = $tunnel->{writer};
+ my ($self, $tunnel) = @_;
- eval {
- run_with_timeout(30, sub {
- print $writer "quit\n";
- $writer->flush();
- });
- };
+ eval { $self->write_tunnel($tunnel, 30, 'quit'); };
my $err = $@;
- finish_command_pipe($tunnel);
+ $self->finish_command_pipe($tunnel, 30);
+
+ if ($tunnel->{sock_addr}) {
+ # ssh does not clean up on local host
+ my $cmd = ['rm', '-f', @{$tunnel->{sock_addr}}]; #
+ PVE::Tools::run_command($cmd);
+
+ # .. and just to be sure check on remote side
+ unshift @{$cmd}, @{$self->{rem_ssh}};
+ PVE::Tools::run_command($cmd);
+ }
die $err if $err;
}
-sub migrate {
- my ($node, $nodeip, $vmid, $online, $force) = @_;
+sub lock_vm {
+ my ($self, $vmid, $code, @param) = @_;
- my $starttime = time();
+ return PVE::QemuConfig->lock_config($vmid, $code, @param);
+}
- my $rem_ssh = [@ssh_cmd, "root\@$nodeip"];
+sub prepare {
+ my ($self, $vmid) = @_;
- local $SIG{INT} = $SIG{TERM} = $SIG{QUIT} = $SIG{HUP} = $SIG{PIPE} = sub {
- logmsg('err', "received interrupt - delayed");
- $delayed_interrupt = 1;
- };
+ my $online = $self->{opts}->{online};
- local $ENV{RSYNC_RSH} = join(' ', @ssh_cmd);
+ $self->{storecfg} = PVE::Storage::config();
- my $session = {
- vmid => $vmid,
- node => $node,
- nodeip => $nodeip,
- force => $force,
- storecfg => PVE::Storage::config(),
- rem_ssh => $rem_ssh,
- };
-
- my $errors;
+ # test if VM exists
+ my $conf = $self->{vmconf} = PVE::QemuConfig->load_config($vmid);
- # lock config during migration
- eval { PVE::QemuServer::lock_config($vmid, sub {
+ PVE::QemuConfig->check_lock($conf);
- eval_int(sub { prepare($session); });
- die $@ if $@;
+ my $running = 0;
+ if (my $pid = PVE::QemuServer::check_running($vmid)) {
+ die "can't migrate running VM without --online\n" if !$online;
+ $running = $pid;
- my $conf = PVE::QemuServer::load_config($vmid);
+ $self->{forcemachine} = PVE::QemuServer::Machine::qemu_machine_pxe($vmid, $conf);
- PVE::QemuServer::check_lock($conf);
+ # To support custom CPU types, we keep QEMU's "-cpu" parameter intact.
+ # Since the parameter itself contains no reference to a custom model,
+ # this makes migration independent of changes to "cpu-models.conf".
+ if ($conf->{cpu}) {
+ my $cpuconf = PVE::QemuServer::CPUConfig::parse_cpu_conf_basic($conf->{cpu});
+ if ($cpuconf && PVE::QemuServer::CPUConfig::is_custom_model($cpuconf->{cputype})) {
+ $self->{forcecpu} = PVE::QemuServer::CPUConfig::get_cpu_from_running_vm($pid);
+ }
+ }
+ }
- my $running = 0;
- if (my $pid = PVE::QemuServer::check_running($vmid)) {
- die "cant migrate running VM without --online\n" if !$online;
- $running = $pid;
+ my $loc_res = PVE::QemuServer::check_local_resources($conf, 1);
+ if (scalar @$loc_res) {
+ if ($self->{running} || !$self->{opts}->{force}) {
+ die "can't migrate VM which uses local devices: " . join(", ", @$loc_res) . "\n";
+ } else {
+ $self->log('info', "migrating VM which uses local devices");
}
+ }
- my $rhash = {};
- eval_int (sub { phase1($session, $conf, $rhash, $running); });
- my $err = $@;
+ my $vollist = PVE::QemuServer::get_vm_volumes($conf);
- if ($err) {
- if ($rhash->{clearlock}) {
- my $unset = { lock => 1 };
- eval { PVE::QemuServer::change_config_nolock($session->{vmid}, {}, $unset, 1) };
- logmsg('err', $@) if $@;
- }
- if ($rhash->{volumes}) {
- foreach my $volid (@{$rhash->{volumes}}) {
- logmsg('err', "found stale volume copy '$volid' on node '$session->{node}'");
- }
- }
- die $err;
+ my $need_activate = [];
+ foreach my $volid (@$vollist) {
+ my ($sid, $volname) = PVE::Storage::parse_volume_id($volid, 1);
+
+ # check if storage is available on both nodes
+ my $targetsid = PVE::QemuServer::map_storage($self->{opts}->{storagemap}, $sid);
+
+ my $scfg = PVE::Storage::storage_check_node($self->{storecfg}, $sid);
+ PVE::Storage::storage_check_node($self->{storecfg}, $targetsid, $self->{node});
+
+ if ($scfg->{shared}) {
+ # PVE::Storage::activate_storage checks this for non-shared storages
+ my $plugin = PVE::Storage::Plugin->lookup($scfg->{type});
+ warn "Used shared storage '$sid' is not online on source node!\n"
+ if !$plugin->check_connection($sid, $scfg);
+ } else {
+ # only activate if not shared
+ next if ($volid =~ m/vm-\d+-cloudinit/);
+ push @$need_activate, $volid;
}
+ }
- # vm is now owned by other node
- my $volids = $rhash->{volumes};
+ # activate volumes
+ PVE::Storage::activate_volumes($self->{storecfg}, $need_activate);
- if ($running) {
+ # test ssh connection
+ my $cmd = [ @{$self->{rem_ssh}}, '/bin/true' ];
+ eval { $self->cmd_quiet($cmd); };
+ die "Can't connect to destination address using public key\n" if $@;
- $rhash = {};
- eval_int(sub { phase2($session, $conf, $rhash); });
- my $err = $@;
+ return $running;
+}
- # always kill tunnel
- if ($rhash->{tunnel}) {
- eval_int(sub { finish_tunnel($rhash->{tunnel}) });
- if ($@) {
- logmsg('err', "stopping tunnel failed - $@");
- $errors = 1;
- }
- }
+sub sync_disks {
+ my ($self, $vmid) = @_;
- # fixme: ther is no config file, so this will never work
- # fixme: use kill(9, $running) to make sure it is stopped
- # always stop local VM - no interrupts possible
- eval { PVE::QemuServer::vm_stop($session->{vmid}, 1); };
- if ($@) {
- logmsg('err', "stopping vm failed - $@");
- $errors = 1;
- }
+ my $conf = $self->{vmconf};
- if ($err) {
- $errors = 1;
- logmsg('err', "online migrate failure - $err");
- }
- }
+ # local volumes which have been copied
+ # and their old_id => new_id pairs
+ $self->{volumes} = [];
+ $self->{volume_map} = {};
- # finalize -- clear migrate lock
- eval_int(sub {
- my $cmd = [ @{$session->{rem_ssh}}, $qm_cmd, 'unlock', $session->{vmid} ];
- run_command($cmd);
- });
- if ($@) {
- logmsg('err', "failed to clear migrate lock - $@");
- $errors = 1;
- }
+ my $storecfg = $self->{storecfg};
+ eval {
- # destroy local copies
- foreach my $volid (@$volids) {
- eval_int(sub { PVE::Storage::vdisk_free($session->{storecfg}, $volid); });
- my $err = $@;
+ # found local volumes and their origin
+ my $local_volumes = {};
+ my $local_volumes_errors = {};
+ my $other_errors = [];
+ my $abort = 0;
- if ($err) {
- logmsg('err', "removing local copy of '$volid' failed - $err");
- $errors = 1;
+ my $log_error = sub {
+ my ($msg, $volid) = @_;
- last if $err =~ /^interrupted by signal$/;
+ if (defined($volid)) {
+ $local_volumes_errors->{$volid} = $msg;
+ } else {
+ push @$other_errors, $msg;
}
- }
- })};
+ $abort = 1;
+ };
- my $err = $@;
+ my @sids = PVE::Storage::storage_ids($storecfg);
+ foreach my $storeid (@sids) {
+ my $scfg = PVE::Storage::storage_config($storecfg, $storeid);
+ next if $scfg->{shared};
+ next if !PVE::Storage::storage_check_enabled($storecfg, $storeid, undef, 1);
- my $delay = time() - $starttime;
- my $mins = int($delay/60);
- my $secs = $delay - $mins*60;
- my $hours = int($mins/60);
- $mins = $mins - $hours*60;
+ # get list from PVE::Storage (for unused volumes)
+ my $dl = PVE::Storage::vdisk_list($storecfg, $storeid, $vmid);
- my $duration = sprintf "%02d:%02d:%02d", $hours, $mins, $secs;
+ next if @{$dl->{$storeid}} == 0;
- if ($err) {
- my $msg = "migration aborted (duration $duration): $err\n";
- logmsg('err', $msg);
- die $msg;
- }
+ my $targetsid = PVE::QemuServer::map_storage($self->{opts}->{storagemap}, $storeid);
+ # check if storage is available on target node
+ PVE::Storage::storage_check_node($storecfg, $targetsid, $self->{node});
- if ($errors) {
- my $msg = "migration finished with problems (duration $duration)\n";
- logmsg('err', $msg);
- die $msg;
- }
+ # grandfather in existing mismatches
+ if ($targetsid ne $storeid) {
+ my $target_scfg = PVE::Storage::storage_config($storecfg, $targetsid);
+ die "content type 'images' is not available on storage '$targetsid'\n"
+ if !$target_scfg->{content}->{images};
+ }
- logmsg('info', "migration finished successfuly (duration $duration)");
-}
+ PVE::Storage::foreach_volid($dl, sub {
+ my ($volid, $sid, $volinfo) = @_;
-sub prepare {
- my ($session) = @_;
+ $local_volumes->{$volid}->{ref} = 'storage';
- my $conffile = PVE::QemuServer::config_file($session->{vmid});
- die "VM $session->{vmid} does not exist on this node\n" if ! -f $conffile;
+ # If with_snapshots is not set for storage migrate, it tries to use
+ # a raw+size stream, but on-the-fly conversion from qcow2 to raw+size
+ # back to qcow2 is currently not possible.
+ $local_volumes->{$volid}->{snapshots} = ($volinfo->{format} =~ /^(?:qcow2|vmdk)$/);
+ $local_volumes->{$volid}->{format} = $volinfo->{format};
+ });
+ }
- # test ssh connection
- my $cmd = [ @{$session->{rem_ssh}}, '/bin/true' ];
- eval { run_command($cmd); };
- die "Can't connect to destination address using public key\n" if $@;
-}
+ my $rep_cfg = PVE::ReplicationConfig->new();
+ my $replication_jobcfg = $rep_cfg->find_local_replication_job($vmid, $self->{node});
+ my $replicatable_volumes = !$replication_jobcfg ? {}
+ : PVE::QemuConfig->get_replicatable_volumes($storecfg, $vmid, $conf, 0, 1);
-sub sync_disks {
- my ($session, $conf, $rhash, $running) = @_;
+ my $test_volid = sub {
+ my ($volid, $attr) = @_;
- logmsg('info', "copying disk images");
+ if ($volid =~ m|^/|) {
+ return if $attr->{shared};
+ $local_volumes->{$volid}->{ref} = 'config';
+ die "local file/device\n";
+ }
- my $res = [];
+ my $snaprefs = $attr->{referenced_in_snapshot};
+
+ if ($attr->{cdrom}) {
+ if ($volid eq 'cdrom') {
+ my $msg = "can't migrate local cdrom drive";
+ if (defined($snaprefs) && !$attr->{referenced_in_config}) {
+ my $snapnames = join(', ', sort keys %$snaprefs);
+ $msg .= " (referenced in snapshot - $snapnames)";
+ }
+ &$log_error("$msg\n");
+ return;
+ }
+ return if $volid eq 'none';
+ }
- eval {
+ my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
- my $volhash = {};
- my $cdromhash = {};
+ my $targetsid = PVE::QemuServer::map_storage($self->{opts}->{storagemap}, $sid);
+ # check if storage is available on both nodes
+ my $scfg = PVE::Storage::storage_check_node($storecfg, $sid);
+ PVE::Storage::storage_check_node($storecfg, $targetsid, $self->{node});
- # get list from PVE::Storage (for unused volumes)
- my $dl = PVE::Storage::vdisk_list($session->{storecfg}, undef, $session->{vmid});
- PVE::Storage::foreach_volid($dl, sub {
- my ($volid, $sid, $volname) = @_;
+ return if $scfg->{shared};
- my $scfg = PVE::Storage::storage_config($session->{storecfg}, $sid);
+ $local_volumes->{$volid}->{ref} = $attr->{referenced_in_config} ? 'config' : 'snapshot';
- return if $scfg->{shared};
+ $local_volumes->{$volid}->{is_vmstate} = $attr->{is_vmstate} ? 1 : 0;
- $volhash->{$volid} = 1;
- });
+ if ($attr->{cdrom}) {
+ if ($volid =~ /vm-\d+-cloudinit/) {
+ $local_volumes->{$volid}->{ref} = 'generated';
+ return;
+ }
+ die "local cdrom image\n";
+ }
- # and add used,owned/non-shared disks (just to be sure we have all)
+ my ($path, $owner) = PVE::Storage::path($storecfg, $volid);
- my $sharedvm = 1;
- PVE::QemuServer::foreach_drive($conf, sub {
- my ($ds, $drive) = @_;
+ die "owned by other VM (owner = VM $owner)\n"
+ if !$owner || ($owner != $vmid);
- my $volid = $drive->{file};
- return if !$volid;
+ if (defined($snaprefs)) {
+ $local_volumes->{$volid}->{snapshots} = 1;
- die "cant migrate local file/device '$volid'\n" if $volid =~ m|^/|;
+ # we cannot migrate shapshots on local storage
+ # exceptions: 'zfspool' or 'qcow2' files (on directory storage)
- if (PVE::QemuServer::drive_is_cdrom($drive)) {
- die "cant migrate local cdrom drive\n" if $volid eq 'cdrom';
- return if $volid eq 'none';
- $cdromhash->{$volid} = 1;
+ die "online storage migration not possible if snapshot exists\n" if $self->{running};
+ if (!($scfg->{type} eq 'zfspool' || $local_volumes->{$volid}->{format} eq 'qcow2')) {
+ die "non-migratable snapshot exists\n";
+ }
+ }
+
+ die "referenced by linked clone(s)\n"
+ if PVE::Storage::volume_is_base_and_used($storecfg, $volid);
+ };
+
+ PVE::QemuServer::foreach_volid($conf, sub {
+ my ($volid, $attr) = @_;
+ eval { $test_volid->($volid, $attr); };
+ if (my $err = $@) {
+ &$log_error($err, $volid);
}
+ });
+
+ foreach my $vol (sort keys %$local_volumes) {
+ my $type = $replicatable_volumes->{$vol} ? 'local, replicated' : 'local';
+ my $ref = $local_volumes->{$vol}->{ref};
+ if ($ref eq 'storage') {
+ $self->log('info', "found $type disk '$vol' (via storage)\n");
+ } elsif ($ref eq 'config') {
+ &$log_error("can't live migrate attached local disks without with-local-disks option\n", $vol)
+ if $self->{running} && !$self->{opts}->{"with-local-disks"};
+ $self->log('info', "found $type disk '$vol' (in current VM config)\n");
+ } elsif ($ref eq 'snapshot') {
+ $self->log('info', "found $type disk '$vol' (referenced by snapshot(s))\n");
+ } elsif ($ref eq 'generated') {
+ $self->log('info', "found generated disk '$vol' (in current VM config)\n");
+ } else {
+ $self->log('info', "found $type disk '$vol'\n");
+ }
+ }
+ foreach my $vol (sort keys %$local_volumes_errors) {
+ $self->log('warn', "can't migrate local disk '$vol': $local_volumes_errors->{$vol}");
+ }
+ foreach my $err (@$other_errors) {
+ $self->log('warn', "$err");
+ }
+
+ if ($abort) {
+ die "can't migrate VM - check log\n";
+ }
+
+ # additional checks for local storage
+ foreach my $volid (keys %$local_volumes) {
my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
+ my $scfg = PVE::Storage::storage_config($storecfg, $sid);
- my $scfg = PVE::Storage::storage_config($session->{storecfg}, $sid);
+ my $migratable = $scfg->{type} =~ /^(?:dir|zfspool|lvmthin|lvm)$/;
- return if $scfg->{shared};
+ die "can't migrate '$volid' - storage type '$scfg->{type}' not supported\n"
+ if !$migratable;
- die "can't migrate local cdrom '$volid'\n" if $cdromhash->{$volid};
+ # image is a linked clone on local storage, se we can't migrate.
+ if (my $basename = (PVE::Storage::parse_volname($storecfg, $volid))[3]) {
+ die "can't migrate '$volid' as it's a clone of '$basename'";
+ }
+ }
+
+ if ($replication_jobcfg) {
+ if ($self->{running}) {
- $sharedvm = 0;
+ my $version = PVE::QemuServer::kvm_user_version();
+ if (!min_version($version, 4, 2)) {
+ die "can't live migrate VM with replicated volumes, pve-qemu to old (< 4.2)!\n"
+ }
- my ($path, $owner) = PVE::Storage::path($session->{storecfg}, $volid);
+ my $live_replicatable_volumes = {};
+ PVE::QemuConfig->foreach_volume($conf, sub {
+ my ($ds, $drive) = @_;
- die "can't migrate volume '$volid' - owned by other VM (owner = VM $owner)\n"
- if !$owner || ($owner != $session->{vmid});
+ my $volid = $drive->{file};
+ $live_replicatable_volumes->{$ds} = $volid
+ if defined($replicatable_volumes->{$volid});
+ });
+ foreach my $drive (keys %$live_replicatable_volumes) {
+ my $volid = $live_replicatable_volumes->{$drive};
- $volhash->{$volid} = 1;
- });
+ my $bitmap = "repl_$drive";
+
+ # start tracking before replication to get full delta + a few duplicates
+ $self->log('info', "$drive: start tracking writes using block-dirty-bitmap '$bitmap'");
+ mon_cmd($vmid, 'block-dirty-bitmap-add', node => "drive-$drive", name => $bitmap);
- if ($running && !$sharedvm) {
- die "can't do online migration - VM uses local disks\n";
+ # other info comes from target node in phase 2
+ $self->{target_drive}->{$drive}->{bitmap} = $bitmap;
+ }
+ }
+ $self->log('info', "replicating disk images");
+
+ my $start_time = time();
+ my $logfunc = sub { $self->log('info', shift) };
+ $self->{replicated_volumes} = PVE::Replication::run_replication(
+ 'PVE::QemuConfig', $replication_jobcfg, $start_time, $start_time, $logfunc);
}
- # do some checks first
- foreach my $volid (keys %$volhash) {
- my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
- my $scfg = PVE::Storage::storage_config($session->{storecfg}, $sid);
+ # sizes in config have to be accurate for remote node to correctly
+ # allocate disks, rescan to be sure
+ my $volid_hash = PVE::QemuServer::scan_volids($storecfg, $vmid);
+ PVE::QemuConfig->foreach_volume($conf, sub {
+ my ($key, $drive) = @_;
+ my ($updated, $old_size, $new_size) = PVE::QemuServer::Drive::update_disksize($drive, $volid_hash);
+ if (defined($updated)) {
+ $conf->{$key} = PVE::QemuServer::print_drive($updated);
+ $self->log('info', "size of disk '$updated->{file}' ($key) updated from $old_size to $new_size\n");
+ }
+ });
- die "can't migrate '$volid' - storagy type '$scfg->{type}' not supported\n"
- if $scfg->{type} ne 'dir';
+ # we want to set the efidisk size in the config to the size of the
+ # real OVMF_VARS.fd image, else we can create a too big image, which does not work
+ if (defined($conf->{efidisk0})) {
+ PVE::QemuServer::update_efidisk_size($conf);
}
- foreach my $volid (keys %$volhash) {
+ $self->log('info', "copying local disk images") if scalar(%$local_volumes);
+
+ foreach my $volid (keys %$local_volumes) {
my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
- push @{$rhash->{volumes}}, $volid;
- PVE::Storage::storage_migrate($session->{storecfg}, $volid, $session->{nodeip}, $sid);
+ my $targetsid = PVE::QemuServer::map_storage($self->{opts}->{storagemap}, $sid);
+ my $ref = $local_volumes->{$volid}->{ref};
+ if ($self->{running} && $ref eq 'config') {
+ push @{$self->{online_local_volumes}}, $volid;
+ } elsif ($ref eq 'generated') {
+ die "can't live migrate VM with local cloudinit disk. use a shared storage instead\n" if $self->{running};
+ # skip all generated volumes but queue them for deletion in phase3_cleanup
+ push @{$self->{volumes}}, $volid;
+ next;
+ } else {
+ next if $self->{replicated_volumes}->{$volid};
+ push @{$self->{volumes}}, $volid;
+ my $opts = $self->{opts};
+ # use 'migrate' limit for transfer to other node
+ my $bwlimit = PVE::Storage::get_bandwidth_limit('migration', [$targetsid, $sid], $opts->{bwlimit});
+ # JSONSchema and get_bandwidth_limit use kbps - storage_migrate bps
+ $bwlimit = $bwlimit * 1024 if defined($bwlimit);
+
+ my $storage_migrate_opts = {
+ 'bwlimit' => $bwlimit,
+ 'insecure' => $opts->{migration_type} eq 'insecure',
+ 'with_snapshots' => $local_volumes->{$volid}->{snapshots},
+ 'allow_rename' => !$local_volumes->{$volid}->{is_vmstate},
+ };
+
+ my $logfunc = sub { $self->log('info', $_[0]); };
+ my $new_volid = eval {
+ PVE::Storage::storage_migrate($storecfg, $volid, $self->{ssh_info},
+ $targetsid, $storage_migrate_opts, $logfunc);
+ };
+ if (my $err = $@) {
+ die "storage migration for '$volid' to storage '$targetsid' failed - $err\n";
+ }
+
+ $self->{volume_map}->{$volid} = $new_volid;
+ $self->log('info', "volume '$volid' is '$new_volid' on the target\n");
+ }
}
};
die "Failed to sync data - $@" if $@;
}
-sub phase1 {
- my ($session, $conf, $rhash, $running) = @_;
+sub cleanup_remotedisks {
+ my ($self) = @_;
- logmsg('info', "starting migration of VM $session->{vmid} to node '$session->{node}' ($session->{nodeip})");
+ foreach my $target_drive (keys %{$self->{target_drive}}) {
+ my $drivestr = $self->{target_drive}->{$target_drive}->{drivestr};
+ next if !defined($drivestr);
- if (my $loc_res = PVE::QemuServer::check_local_resources($conf, 1)) {
- if ($running || !$session->{force}) {
- die "can't migrate VM which uses local devices\n";
- } else {
- logmsg('info', "migrating VM which uses local devices");
+ my $drive = PVE::QemuServer::parse_drive($target_drive, $drivestr);
+
+ # don't clean up replicated disks!
+ next if defined($self->{replicated_volumes}->{$drive->{file}});
+
+ my ($storeid, $volname) = PVE::Storage::parse_volume_id($drive->{file});
+
+ my $cmd = [@{$self->{rem_ssh}}, 'pvesm', 'free', "$storeid:$volname"];
+
+ eval{ PVE::Tools::run_command($cmd, outfunc => sub {}, errfunc => sub {}) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
}
}
+}
- # set migrate lock in config file
- $rhash->{clearlock} = 1;
+sub cleanup_bitmaps {
+ my ($self) = @_;
+ foreach my $drive (keys %{$self->{target_drive}}) {
+ my $bitmap = $self->{target_drive}->{$drive}->{bitmap};
+ next if !$bitmap;
+ $self->log('info', "$drive: removing block-dirty-bitmap '$bitmap'");
+ mon_cmd($self->{vmid}, 'block-dirty-bitmap-remove', node => "drive-$drive", name => $bitmap);
+ }
+}
- PVE::QemuServer::change_config_nolock($session->{vmid}, { lock => 'migrate' }, {}, 1);
+sub phase1 {
+ my ($self, $vmid) = @_;
- sync_disks($session, $conf, $rhash, $running);
+ $self->log('info', "starting migration of VM $vmid to node '$self->{node}' ($self->{nodeip})");
- # move config to remote node
- my $conffile = PVE::QemuServer::config_file($session->{vmid});
- my $newconffile = PVE::QemuServer::config_file($session->{vmid}, $session->{node});
+ my $conf = $self->{vmconf};
- die "Failed to move config to node '$session->{node}' - rename failed: $!\n"
- if !rename($conffile, $newconffile);
+ # set migrate lock in config file
+ $conf->{lock} = 'migrate';
+ PVE::QemuConfig->write_config($vmid, $conf);
+
+ sync_disks($self, $vmid);
+
+ # sync_disks fixes disk sizes to match their actual size, write changes so
+ # target allocates correct volumes
+ PVE::QemuConfig->write_config($vmid, $conf);
};
+sub phase1_cleanup {
+ my ($self, $vmid, $err) = @_;
+
+ $self->log('info', "aborting phase 1 - cleanup resources");
+
+ my $conf = $self->{vmconf};
+ delete $conf->{lock};
+ eval { PVE::QemuConfig->write_config($vmid, $conf) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ }
+
+ if ($self->{volumes}) {
+ foreach my $volid (@{$self->{volumes}}) {
+ $self->log('err', "found stale volume copy '$volid' on node '$self->{node}'");
+ # fixme: try to remove ?
+ }
+ }
+
+ eval { $self->cleanup_bitmaps() };
+ if (my $err =$@) {
+ $self->log('err', $err);
+ }
+
+}
+
sub phase2 {
- my ($session, $conf, $rhash) = shift;
+ my ($self, $vmid) = @_;
- logmsg('info', "starting VM on remote node '$session->{node}'");
+ my $conf = $self->{vmconf};
+ $self->log('info', "starting VM $vmid on remote node '$self->{node}'");
+
+ my $raddr;
my $rport;
+ my $ruri; # the whole migration dst. URI (protocol:address[:port])
+ my $nodename = PVE::INotify::nodename();
## start on remote node
- my $cmd = [@{$session->{rem_ssh}}, $qm_cmd, 'start',
- $session->{vmid}, '--stateuri', 'tcp', '--skiplock'];
+ my $cmd = [@{$self->{rem_ssh}}];
+
+ my $spice_ticket;
+ if (PVE::QemuServer::vga_conf_has_spice($conf->{vga})) {
+ my $res = mon_cmd($vmid, 'query-spice');
+ $spice_ticket = $res->{ticket};
+ }
+
+ push @$cmd , 'qm', 'start', $vmid, '--skiplock', '--migratedfrom', $nodename;
+
+ my $migration_type = $self->{opts}->{migration_type};
+
+ push @$cmd, '--migration_type', $migration_type;
+
+ push @$cmd, '--migration_network', $self->{opts}->{migration_network}
+ if $self->{opts}->{migration_network};
+
+ if ($migration_type eq 'insecure') {
+ push @$cmd, '--stateuri', 'tcp';
+ } else {
+ push @$cmd, '--stateuri', 'unix';
+ }
+
+ if ($self->{forcemachine}) {
+ push @$cmd, '--machine', $self->{forcemachine};
+ }
- run_command($cmd, outfunc => sub {
+ if ($self->{forcecpu}) {
+ push @$cmd, '--force-cpu', $self->{forcecpu};
+ }
+
+ if ($self->{online_local_volumes}) {
+ push @$cmd, '--targetstorage', ($self->{opts}->{targetstorage} // '1');
+ }
+
+ my $spice_port;
+ my $tunnel_addr = [];
+ my $sock_addr = [];
+ # version > 0 for unix socket support
+ my $nbd_protocol_version = 1;
+ # TODO change to 'spice_ticket: <ticket>\n' in 7.0
+ my $input = $spice_ticket ? "$spice_ticket\n" : "\n";
+ $input .= "nbd_protocol_version: $nbd_protocol_version\n";
+ foreach my $volid (keys %{$self->{replicated_volumes}}) {
+ $input .= "replicated_volume: $volid\n";
+ }
+
+ my $target_replicated_volumes = {};
+
+ # Note: We try to keep $spice_ticket secret (do not pass via command line parameter)
+ # instead we pipe it through STDIN
+ my $exitcode = PVE::Tools::run_command($cmd, input => $input, outfunc => sub {
my $line = shift;
- if ($line =~ m/^migration listens on port (\d+)$/) {
- $rport = $1;
+ if ($line =~ m/^migration listens on tcp:(localhost|[\d\.]+|\[[\d\.:a-fA-F]+\]):(\d+)$/) {
+ $raddr = $1;
+ $rport = int($2);
+ $ruri = "tcp:$raddr:$rport";
+ }
+ elsif ($line =~ m!^migration listens on unix:(/run/qemu-server/(\d+)\.migrate)$!) {
+ $raddr = $1;
+ die "Destination UNIX sockets VMID does not match source VMID" if $vmid ne $2;
+ $ruri = "unix:$raddr";
+ }
+ elsif ($line =~ m/^migration listens on port (\d+)$/) {
+ $raddr = "localhost";
+ $rport = int($1);
+ $ruri = "tcp:$raddr:$rport";
}
- });
+ elsif ($line =~ m/^spice listens on port (\d+)$/) {
+ $spice_port = int($1);
+ }
+ elsif ($line =~ m/^storage migration listens on nbd:(localhost|[\d\.]+|\[[\d\.:a-fA-F]+\]):(\d+):exportname=(\S+) volume:(\S+)$/) {
+ my $drivestr = $4;
+ my $nbd_uri = "nbd:$1:$2:exportname=$3";
+ my $targetdrive = $3;
+ $targetdrive =~ s/drive-//g;
+
+ $self->{target_drive}->{$targetdrive}->{drivestr} = $drivestr;
+ $self->{target_drive}->{$targetdrive}->{nbd_uri} = $nbd_uri;
+ } elsif ($line =~ m!^storage migration listens on nbd:unix:(/run/qemu-server/(\d+)_nbd\.migrate):exportname=(\S+) volume:(\S+)$!) {
+ my $drivestr = $4;
+ die "Destination UNIX socket's VMID does not match source VMID" if $vmid ne $2;
+ my $nbd_unix_addr = $1;
+ my $nbd_uri = "nbd:unix:$nbd_unix_addr:exportname=$3";
+ my $targetdrive = $3;
+ $targetdrive =~ s/drive-//g;
+
+ $self->{target_drive}->{$targetdrive}->{drivestr} = $drivestr;
+ $self->{target_drive}->{$targetdrive}->{nbd_uri} = $nbd_uri;
+ push @$tunnel_addr, "$nbd_unix_addr:$nbd_unix_addr";
+ push @$sock_addr, $nbd_unix_addr;
+ } elsif ($line =~ m/^re-using replicated volume: (\S+) - (.*)$/) {
+ my $drive = $1;
+ my $volid = $2;
+ $target_replicated_volumes->{$volid} = $drive;
+ } elsif ($line =~ m/^QEMU: (.*)$/) {
+ $self->log('info', "[$self->{node}] $1\n");
+ }
+ }, errfunc => sub {
+ my $line = shift;
+ $self->log('info', "[$self->{node}] $line");
+ }, noerr => 1);
- die "unable to detect remote migration port\n" if !$rport;
+ die "remote command failed with exit code $exitcode\n" if $exitcode;
- logmsg('info', "starting migration tunnel");
+ die "unable to detect remote migration address\n" if !$raddr;
+
+ if (scalar(keys %$target_replicated_volumes) != scalar(keys %{$self->{replicated_volumes}})) {
+ die "number of replicated disks on source and target node do not match - target node too old?\n"
+ }
- ## create tunnel to remote port
- my $lport = PVE::QemuServer::next_migrate_port();
- $rhash->{tunnel} = fork_tunnel($session->{nodeip}, $lport, $rport);
+ $self->log('info', "start remote tunnel");
- logmsg('info', "starting online/live migration");
- # start migration
+ if ($migration_type eq 'secure') {
+
+ if ($ruri =~ /^unix:/) {
+ unlink $raddr;
+ push @$tunnel_addr, "$raddr:$raddr";
+ $self->{tunnel} = $self->fork_tunnel($tunnel_addr);
+ push @$sock_addr, $raddr;
+
+ my $unix_socket_try = 0; # wait for the socket to become ready
+ while ($unix_socket_try <= 100) {
+ $unix_socket_try++;
+ my $available = 0;
+ foreach my $sock (@$sock_addr) {
+ if (-S $sock) {
+ $available++;
+ }
+ }
+
+ if ($available == @$sock_addr) {
+ last;
+ }
+
+ usleep(50000);
+ }
+ if ($unix_socket_try > 100) {
+ $self->{errors} = 1;
+ $self->finish_tunnel($self->{tunnel});
+ die "Timeout, migration socket $ruri did not get ready";
+ }
+
+ } elsif ($ruri =~ /^tcp:/) {
+ my $tunnel_addr;
+ if ($raddr eq "localhost") {
+ # for backwards compatibility with older qemu-server versions
+ my $pfamily = PVE::Tools::get_host_address_family($nodename);
+ my $lport = PVE::Tools::next_migrate_port($pfamily);
+ $tunnel_addr = "$lport:localhost:$rport";
+ }
+
+ $self->{tunnel} = $self->fork_tunnel($tunnel_addr);
+
+ } else {
+ die "unsupported protocol in migration URI: $ruri\n";
+ }
+ } else {
+ #fork tunnel for insecure migration, to send faster commands like resume
+ $self->{tunnel} = $self->fork_tunnel();
+ }
+ $self->{tunnel}->{sock_addr} = $sock_addr if (@$sock_addr);
my $start = time();
- PVE::QemuServer::vm_monitor_command($session->{vmid}, "migrate -d \"tcp:localhost:$lport\"");
+ my $opt_bwlimit = $self->{opts}->{bwlimit};
+
+ if (defined($self->{online_local_volumes})) {
+ $self->{storage_migration} = 1;
+ $self->{storage_migration_jobs} = {};
+ $self->log('info', "starting storage migration");
+
+ die "The number of local disks does not match between the source and the destination.\n"
+ if (scalar(keys %{$self->{target_drive}}) != scalar @{$self->{online_local_volumes}});
+ foreach my $drive (keys %{$self->{target_drive}}){
+ my $target = $self->{target_drive}->{$drive};
+ my $nbd_uri = $target->{nbd_uri};
+
+ my $source_drive = PVE::QemuServer::parse_drive($drive, $conf->{$drive});
+ my $target_drive = PVE::QemuServer::parse_drive($drive, $target->{drivestr});
+
+ my $source_volid = $source_drive->{file};
+ my $target_volid = $target_drive->{file};
+
+ my $source_sid = PVE::Storage::Plugin::parse_volume_id($source_volid);
+ my $target_sid = PVE::Storage::Plugin::parse_volume_id($target_volid);
+
+ my $bwlimit = PVE::Storage::get_bandwidth_limit('migration', [$source_sid, $target_sid], $opt_bwlimit);
+ my $bitmap = $target->{bitmap};
+
+ $self->log('info', "$drive: start migration to $nbd_uri");
+ PVE::QemuServer::qemu_drive_mirror($vmid, $drive, $nbd_uri, $vmid, undef, $self->{storage_migration_jobs}, 'skip', undef, $bwlimit, $bitmap);
- my $lstat = '';
+ $self->{volume_map}->{$source_volid} = $target_volid;
+ $self->log('info', "volume '$source_volid' is '$target_volid' on the target\n");
+ }
+ }
+
+ $self->log('info', "starting online/live migration on $ruri");
+ $self->{livemigration} = 1;
+
+ # load_defaults
+ my $defaults = PVE::QemuServer::load_defaults();
+
+ $self->log('info', "set migration_caps");
+ eval {
+ PVE::QemuServer::set_migration_caps($vmid);
+ };
+ warn $@ if $@;
+
+ my $qemu_migrate_params = {};
+
+ # migrate speed can be set via bwlimit (datacenter.cfg and API) and via the
+ # migrate_speed parameter in qm.conf - take the lower of the two.
+ my $bwlimit = PVE::Storage::get_bandwidth_limit('migration', undef, $opt_bwlimit) // 0;
+ my $migrate_speed = $conf->{migrate_speed} // $bwlimit;
+ # migrate_speed is in MB/s, bwlimit in KB/s
+ $migrate_speed *= 1024;
+
+ $migrate_speed = ($bwlimit < $migrate_speed) ? $bwlimit : $migrate_speed;
+
+ # always set migrate speed (overwrite kvm default of 32m) we set a very high
+ # default of 8192m which is basically unlimited
+ $migrate_speed ||= ($defaults->{migrate_speed} || 8192) * 1024;
+
+ # qmp takes migrate_speed in B/s.
+ $migrate_speed *= 1024;
+ $self->log('info', "migration speed limit: $migrate_speed B/s");
+ $qemu_migrate_params->{'max-bandwidth'} = int($migrate_speed);
+
+ my $migrate_downtime = $defaults->{migrate_downtime};
+ $migrate_downtime = $conf->{migrate_downtime} if defined($conf->{migrate_downtime});
+ # migrate-set-parameters expects limit in ms
+ $migrate_downtime *= 1000;
+ $self->log('info', "migration downtime limit: $migrate_downtime ms");
+ $qemu_migrate_params->{'downtime-limit'} = int($migrate_downtime);
+
+ # set cachesize to 10% of the total memory
+ my $memory = $conf->{memory} || $defaults->{memory};
+ my $cachesize = int($memory * 1048576 / 10);
+ $cachesize = round_powerof2($cachesize);
+
+ $self->log('info', "migration cachesize: $cachesize B");
+ $qemu_migrate_params->{'xbzrle-cache-size'} = int($cachesize);
+
+ $self->log('info', "set migration parameters");
+ eval {
+ mon_cmd($vmid, "migrate-set-parameters", %{$qemu_migrate_params});
+ };
+ $self->log('info', "migrate-set-parameters error: $@") if $@;
+
+ if (PVE::QemuServer::vga_conf_has_spice($conf->{vga})) {
+ my $rpcenv = PVE::RPCEnvironment::get();
+ my $authuser = $rpcenv->get_user();
+
+ my (undef, $proxyticket) = PVE::AccessControl::assemble_spice_ticket($authuser, $vmid, $self->{node});
+
+ my $filename = "/etc/pve/nodes/$self->{node}/pve-ssl.pem";
+ my $subject = PVE::AccessControl::read_x509_subject_spice($filename);
+
+ $self->log('info', "spice client_migrate_info");
+
+ eval {
+ mon_cmd($vmid, "client_migrate_info", protocol => 'spice',
+ hostname => $proxyticket, 'port' => 0, 'tls-port' => $spice_port,
+ 'cert-subject' => $subject);
+ };
+ $self->log('info', "client_migrate_info error: $@") if $@;
+
+ }
+
+ $self->log('info', "start migrate command to $ruri");
+ eval {
+ mon_cmd($vmid, "migrate", uri => $ruri);
+ };
+ my $merr = $@;
+ $self->log('info', "migrate uri => $ruri failed: $merr") if $merr;
+
+ my $lstat = 0;
+ my $usleep = 1000000;
+ my $i = 0;
+ my $err_count = 0;
+ my $lastrem = undef;
+ my $downtimecounter = 0;
while (1) {
- sleep (2);
- my $stat = PVE::QemuServer::vm_monitor_command($session->{vmid}, "info migrate", 1);
- if ($stat =~ m/^Migration status: (active|completed|failed|cancelled)$/im) {
- my $ms = $1;
-
- if ($stat ne $lstat) {
- if ($ms eq 'active') {
- my ($trans, $rem, $total) = (0, 0, 0);
- $trans = $1 if $stat =~ m/^transferred ram: (\d+) kbytes$/im;
- $rem = $1 if $stat =~ m/^remaining ram: (\d+) kbytes$/im;
- $total = $1 if $stat =~ m/^total ram: (\d+) kbytes$/im;
-
- logmsg('info', "migration status: $ms (transferred ${trans}KB, " .
- "remaining ${rem}KB), total ${total}KB)");
- } else {
- logmsg('info', "migration status: $ms");
- }
+ $i++;
+ my $avglstat = $lstat/$i if $lstat;
+
+ usleep($usleep);
+ my $stat;
+ eval {
+ $stat = mon_cmd($vmid, "query-migrate");
+ };
+ if (my $err = $@) {
+ $err_count++;
+ warn "query migrate failed: $err\n";
+ $self->log('info', "query migrate failed: $err");
+ if ($err_count <= 5) {
+ usleep(1000000);
+ next;
}
+ die "too many query migrate failures - aborting\n";
+ }
+
+ if (defined($stat->{status}) && $stat->{status} =~ m/^(setup)$/im) {
+ sleep(1);
+ next;
+ }
- if ($ms eq 'completed') {
+ if (defined($stat->{status}) && $stat->{status} =~ m/^(active|completed|failed|cancelled)$/im) {
+ $merr = undef;
+ $err_count = 0;
+ if ($stat->{status} eq 'completed') {
my $delay = time() - $start;
if ($delay > 0) {
- my $mbps = sprintf "%.2f", $conf->{memory}/$delay;
- logmsg('info', "migration speed: $mbps MB/s");
+ my $mbps = sprintf "%.2f", $memory / $delay;
+ my $downtime = $stat->{downtime} || 0;
+ $self->log('info', "migration speed: $mbps MB/s - downtime $downtime ms");
}
}
- if ($ms eq 'failed' || $ms eq 'cancelled') {
+ if ($stat->{status} eq 'failed' || $stat->{status} eq 'cancelled') {
+ $self->log('info', "migration status error: $stat->{status}");
die "aborting\n"
}
- last if $ms ne 'active';
+ if ($stat->{status} ne 'active') {
+ $self->log('info', "migration status: $stat->{status}");
+ last;
+ }
+
+ if ($stat->{ram}->{transferred} ne $lstat) {
+ my $trans = $stat->{ram}->{transferred} || 0;
+ my $rem = $stat->{ram}->{remaining} || 0;
+ my $total = $stat->{ram}->{total} || 0;
+ my $xbzrlecachesize = $stat->{"xbzrle-cache"}->{"cache-size"} || 0;
+ my $xbzrlebytes = $stat->{"xbzrle-cache"}->{"bytes"} || 0;
+ my $xbzrlepages = $stat->{"xbzrle-cache"}->{"pages"} || 0;
+ my $xbzrlecachemiss = $stat->{"xbzrle-cache"}->{"cache-miss"} || 0;
+ my $xbzrleoverflow = $stat->{"xbzrle-cache"}->{"overflow"} || 0;
+ # reduce sleep if remainig memory is lower than the average transfer speed
+ $usleep = 100000 if $avglstat && $rem < $avglstat;
+
+ $self->log('info', "migration status: $stat->{status} (transferred ${trans}, " .
+ "remaining ${rem}), total ${total})");
+
+ if (${xbzrlecachesize}) {
+ $self->log('info', "migration xbzrle cachesize: ${xbzrlecachesize} transferred ${xbzrlebytes} pages ${xbzrlepages} cachemiss ${xbzrlecachemiss} overflow ${xbzrleoverflow}");
+ }
+
+ if (($lastrem && $rem > $lastrem ) || ($rem == 0)) {
+ $downtimecounter++;
+ }
+ $lastrem = $rem;
+
+ if ($downtimecounter > 5) {
+ $downtimecounter = 0;
+ $migrate_downtime *= 2;
+ $self->log('info', "auto-increased downtime to continue migration: $migrate_downtime ms");
+ eval {
+ # migrate-set-parameters does not touch values not
+ # specified, so this only changes downtime-limit
+ mon_cmd($vmid, "migrate-set-parameters", 'downtime-limit' => int($migrate_downtime));
+ };
+ $self->log('info', "migrate-set-parameters error: $@") if $@;
+ }
+
+ }
+
+
+ $lstat = $stat->{ram}->{transferred};
+
} else {
- die "unable to parse migration status '$stat' - aborting\n";
+ die $merr if $merr;
+ die "unable to parse migration status '$stat->{status}' - aborting\n";
}
- $lstat = $stat;
+ }
+}
+
+sub phase2_cleanup {
+ my ($self, $vmid, $err) = @_;
+
+ return if !$self->{errors};
+ $self->{phase2errors} = 1;
+
+ $self->log('info', "aborting phase 2 - cleanup resources");
+
+ $self->log('info', "migrate_cancel");
+ eval {
+ mon_cmd($vmid, "migrate_cancel");
};
+ $self->log('info', "migrate_cancel error: $@") if $@;
+
+ my $conf = $self->{vmconf};
+ delete $conf->{lock};
+ eval { PVE::QemuConfig->write_config($vmid, $conf) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ }
+
+ # cleanup ressources on target host
+ if ($self->{storage_migration}) {
+ eval { PVE::QemuServer::qemu_blockjobs_cancel($vmid, $self->{storage_migration_jobs}) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ }
+ }
+
+ eval { $self->cleanup_bitmaps() };
+ if (my $err =$@) {
+ $self->log('err', $err);
+ }
+
+ my $nodename = PVE::INotify::nodename();
+
+ my $cmd = [@{$self->{rem_ssh}}, 'qm', 'stop', $vmid, '--skiplock', '--migratedfrom', $nodename];
+ eval{ PVE::Tools::run_command($cmd, outfunc => sub {}, errfunc => sub {}) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+
+ # cleanup after stopping, otherwise disks might be in-use by target VM!
+ eval { PVE::QemuMigrate::cleanup_remotedisks($self) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ }
+
+
+ if ($self->{tunnel}) {
+ eval { finish_tunnel($self, $self->{tunnel}); };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+ }
}
+
+sub phase3 {
+ my ($self, $vmid) = @_;
+
+ my $volids = $self->{volumes};
+ return if $self->{phase2errors};
+
+ # destroy local copies
+ foreach my $volid (@$volids) {
+ eval { PVE::Storage::vdisk_free($self->{storecfg}, $volid); };
+ if (my $err = $@) {
+ $self->log('err', "removing local copy of '$volid' failed - $err");
+ $self->{errors} = 1;
+ last if $err =~ /^interrupted by signal$/;
+ }
+ }
+}
+
+sub phase3_cleanup {
+ my ($self, $vmid, $err) = @_;
+
+ my $conf = $self->{vmconf};
+ return if $self->{phase2errors};
+
+ my $tunnel = $self->{tunnel};
+
+ if ($self->{storage_migration}) {
+ # finish block-job with block-job-cancel, to disconnect source VM from NBD
+ # to avoid it trying to re-establish it. We are in blockjob ready state,
+ # thus, this command changes to it to blockjob complete (see qapi docs)
+ eval { PVE::QemuServer::qemu_drive_mirror_monitor($vmid, undef, $self->{storage_migration_jobs}, 'cancel'); };
+
+ if (my $err = $@) {
+ eval { PVE::QemuServer::qemu_blockjobs_cancel($vmid, $self->{storage_migration_jobs}) };
+ eval { PVE::QemuMigrate::cleanup_remotedisks($self) };
+ die "Failed to complete storage migration: $err\n";
+ }
+ }
+
+ if ($self->{volume_map}) {
+ my $target_drives = $self->{target_drive};
+
+ # FIXME: for NBD storage migration we now only update the volid, and
+ # not the full drivestr from the target node. Workaround that until we
+ # got some real rescan, to avoid things like wrong format in the drive
+ delete $conf->{$_} for keys %$target_drives;
+ PVE::QemuConfig->update_volume_ids($conf, $self->{volume_map});
+
+ for my $drive (keys %$target_drives) {
+ $conf->{$drive} = $target_drives->{$drive}->{drivestr};
+ }
+ PVE::QemuConfig->write_config($vmid, $conf);
+ }
+
+ # transfer replication state before move config
+ $self->transfer_replication_state() if $self->{replicated_volumes};
+
+ # move config to remote node
+ my $conffile = PVE::QemuConfig->config_file($vmid);
+ my $newconffile = PVE::QemuConfig->config_file($vmid, $self->{node});
+
+ die "Failed to move config to node '$self->{node}' - rename failed: $!\n"
+ if !rename($conffile, $newconffile);
+
+ $self->switch_replication_job_target() if $self->{replicated_volumes};
+
+ if ($self->{livemigration}) {
+ if ($self->{storage_migration}) {
+ # stop nbd server on remote vm - requirement for resume since 2.9
+ my $cmd = [@{$self->{rem_ssh}}, 'qm', 'nbdstop', $vmid];
+
+ eval{ PVE::Tools::run_command($cmd, outfunc => sub {}, errfunc => sub {}) };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+ }
+
+ # config moved and nbd server stopped - now we can resume vm on target
+ if ($tunnel && $tunnel->{version} && $tunnel->{version} >= 1) {
+ eval {
+ $self->write_tunnel($tunnel, 30, "resume $vmid");
+ };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+ } else {
+ my $cmd = [@{$self->{rem_ssh}}, 'qm', 'resume', $vmid, '--skiplock', '--nocheck'];
+ my $logf = sub {
+ my $line = shift;
+ $self->log('err', $line);
+ };
+ eval { PVE::Tools::run_command($cmd, outfunc => sub {}, errfunc => $logf); };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+ }
+
+ if ($self->{storage_migration} && PVE::QemuServer::parse_guest_agent($conf)->{fstrim_cloned_disks} && $self->{running}) {
+ my $cmd = [@{$self->{rem_ssh}}, 'qm', 'guest', 'cmd', $vmid, 'fstrim'];
+ eval{ PVE::Tools::run_command($cmd, outfunc => sub {}, errfunc => sub {}) };
+ }
+ }
+
+ # close tunnel on successful migration, on error phase2_cleanup closed it
+ if ($tunnel) {
+ eval { finish_tunnel($self, $tunnel); };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+ }
+
+ eval {
+ my $timer = 0;
+ if (PVE::QemuServer::vga_conf_has_spice($conf->{vga}) && $self->{running}) {
+ $self->log('info', "Waiting for spice server migration");
+ while (1) {
+ my $res = mon_cmd($vmid, 'query-spice');
+ last if int($res->{'migrated'}) == 1;
+ last if $timer > 50;
+ $timer ++;
+ usleep(200000);
+ }
+ }
+ };
+
+ # always stop local VM
+ eval { PVE::QemuServer::vm_stop($self->{storecfg}, $vmid, 1, 1); };
+ if (my $err = $@) {
+ $self->log('err', "stopping vm failed - $err");
+ $self->{errors} = 1;
+ }
+
+ # always deactivate volumes - avoid lvm LVs to be active on several nodes
+ eval {
+ my $vollist = PVE::QemuServer::get_vm_volumes($conf);
+ PVE::Storage::deactivate_volumes($self->{storecfg}, $vollist);
+ };
+ if (my $err = $@) {
+ $self->log('err', $err);
+ $self->{errors} = 1;
+ }
+
+ if($self->{storage_migration}) {
+ # destroy local copies
+ my $volids = $self->{online_local_volumes};
+
+ foreach my $volid (@$volids) {
+ # keep replicated volumes!
+ next if $self->{replicated_volumes}->{$volid};
+
+ eval { PVE::Storage::vdisk_free($self->{storecfg}, $volid); };
+ if (my $err = $@) {
+ $self->log('err', "removing local copy of '$volid' failed - $err");
+ $self->{errors} = 1;
+ last if $err =~ /^interrupted by signal$/;
+ }
+ }
+
+ }
+
+ # clear migrate lock
+ my $cmd = [ @{$self->{rem_ssh}}, 'qm', 'unlock', $vmid ];
+ $self->cmd_logerr($cmd, errmsg => "failed to clear migrate lock");
+}
+
+sub final_cleanup {
+ my ($self, $vmid) = @_;
+
+ # nothing to do
+}
+
+sub round_powerof2 {
+ return 1 if $_[0] < 2;
+ return 2 << int(log($_[0]-1)/log(2));
+}
+
+1;