X-Git-Url: https://git.proxmox.com/?p=qemu-server.git;a=blobdiff_plain;f=PVE%2FQemuMigrate.pm;h=07d57ed3b70f2836be69ef111064112d5dc3efd1;hp=bf0bb44cb64a01ce279fcdd68e7bbcbb28b85a72;hb=c7789f54ad61e4e4658377259bfbecde141a6ee4;hpb=5009a8c755fb52d6e9a85bc10996c1c2bfeda950

diff --git a/PVE/QemuMigrate.pm b/PVE/QemuMigrate.pm
index bf0bb44..07d57ed 100644
--- a/PVE/QemuMigrate.pm
+++ b/PVE/QemuMigrate.pm
@@ -13,6 +13,9 @@ use PVE::Storage;
 use PVE::QemuServer;
 use Time::HiRes qw( usleep );
 use PVE::RPCEnvironment;
+use PVE::ReplicationConfig;
+use PVE::ReplicationState;
+use PVE::Replication;
 
 use base qw(PVE::AbstractMigrate);
 
@@ -88,20 +91,62 @@ sub finish_command_pipe {
 	if !&$collect_child_process();
 }
 
+sub read_tunnel {
+    my ($self, $tunnel, $timeout) = @_;
+
+    $timeout = 60 if !defined($timeout);
+
+    my $reader = $tunnel->{reader};
+
+    my $output;
+    eval {
+	PVE::Tools::run_with_timeout($timeout, sub { $output = <$reader>; });
+    };
+    die "reading from tunnel failed: $@\n" if $@;
+
+    chomp $output;
+
+    return $output;
+}
+
+sub write_tunnel {
+    my ($self, $tunnel, $timeout, $command) = @_;
+
+    $timeout = 60 if !defined($timeout);
+
+    my $writer = $tunnel->{writer};
+
+    eval {
+	PVE::Tools::run_with_timeout($timeout, sub {
+	    print $writer "$command\n";
+	    $writer->flush();
+	});
+    };
+    die "writing to tunnel failed: $@\n" if $@;
+
+    if ($tunnel->{version} && $tunnel->{version} >= 1) {
+	my $res = eval { $self->read_tunnel($tunnel, 10); };
+	die "no reply to command '$command': $@\n" if $@;
+
+	if ($res eq 'OK') {
+	    return;
+	} else {
+	    die "tunnel replied '$res' to command '$command'\n";
+	}
+    }
+}
+
 sub fork_tunnel {
     my ($self, $tunnel_addr) = @_;
 
     my @localtunnelinfo = defined($tunnel_addr) ? ('-L' , $tunnel_addr ) : ();
 
-    my $cmd = [@{$self->{rem_ssh}}, '-o ExitOnForwardFailure=yes', @localtunnelinfo, 'qm', 'mtunnel' ];
+    my $cmd = [@{$self->{rem_ssh}}, '-o ExitOnForwardFailure=yes', @localtunnelinfo, '/usr/sbin/qm', 'mtunnel' ];
 
     my $tunnel = $self->fork_command_pipe($cmd);
 
-    my $reader = $tunnel->{reader};
-
-    my $helo;
     eval {
-	PVE::Tools::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"
@@ -109,6 +154,16 @@ sub fork_tunnel {
     };
     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) {
 	$self->finish_command_pipe($tunnel);
 	die "can't open migration tunnel - $err";
@@ -119,14 +174,7 @@ sub fork_tunnel {
 sub finish_tunnel {
     my ($self, $tunnel) = @_;
 
-    my $writer = $tunnel->{writer};
-
-    eval {
-	PVE::Tools::run_with_timeout(30, sub {
-	    print $writer "quit\n";
-	    $writer->flush();
-	});
-    };
+    eval { $self->write_tunnel($tunnel, 30, 'quit'); };
     my $err = $@;
 
     $self->finish_command_pipe($tunnel, 30);
@@ -186,7 +234,7 @@ sub prepare {
 	my ($sid, $volname) = PVE::Storage::parse_volume_id($volid, 1);
 
 	# check if storage is available on both nodes
-	my $targetsid = $self->{opts}->{targetstorage} ? $self->{opts}->{targetstorage} : $sid;
+	my $targetsid = $self->{opts}->{targetstorage} // $sid;
 
 	my $scfg = PVE::Storage::storage_check_node($self->{storecfg}, $sid);
 	PVE::Storage::storage_check_node($self->{storecfg}, $targetsid, $self->{node});
@@ -221,7 +269,7 @@ sub sync_disks {
     # local volumes which have been copied
     $self->{volumes} = [];
 
-    my $res = [];
+    my $override_targetsid = $self->{opts}->{targetstorage};
 
     eval {
 
@@ -255,7 +303,7 @@ sub sync_disks {
 
 	    next if @{$dl->{$storeid}} == 0;
 
-	    my $targetsid = $self->{opts}->{targetstorage} ? $self->{opts}->{targetstorage} : $storeid;
+	    my $targetsid = $override_targetsid // $storeid;
 
 	    # check if storage is available on target node
 	    PVE::Storage::storage_check_node($self->{storecfg}, $targetsid, $self->{node});
@@ -264,7 +312,7 @@ sub sync_disks {
 	    PVE::Storage::foreach_volid($dl, sub {
 		my ($volid, $sid, $volname) = @_;
 
-		$local_volumes->{$volid} = 'storage';
+		$local_volumes->{$volid}->{ref} = 'storage';
 	    });
 	}
 
@@ -272,7 +320,8 @@ sub sync_disks {
 	    my ($volid, $attr) = @_;
 
 	    if ($volid =~ m|^/|) {
-		$local_volumes->{$volid} = 'config';
+		return if $attr->{shared};
+		$local_volumes->{$volid}->{ref} = 'config';
 		die "local file/device\n";
 	    }
 
@@ -293,7 +342,7 @@ sub sync_disks {
 
 	    my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
 
-	    my $targetsid = $self->{opts}->{targetstorage} ? $self->{opts}->{targetstorage} : $sid;
+	    my $targetsid = $override_targetsid // $sid;
 	    # check if storage is available on both nodes
 	    my $scfg = PVE::Storage::storage_check_node($self->{storecfg}, $sid);
 	    PVE::Storage::storage_check_node($self->{storecfg}, $targetsid, $self->{node});
@@ -302,7 +351,7 @@ sub sync_disks {
 
 	    $sharedvm = 0;
 
-	    $local_volumes->{$volid} = $attr->{referenced_in_config} ? 'config' : 'snapshot';
+	    $local_volumes->{$volid}->{ref} = $attr->{referenced_in_config} ? 'config' : 'snapshot';
 
 	    die "local cdrom image\n" if $attr->{cdrom};
 
@@ -311,11 +360,12 @@ sub sync_disks {
 	    die "owned by other VM (owner = VM $owner)\n"
 		if !$owner || ($owner != $self->{vmid});
 
+	    my $format = PVE::QemuServer::qemu_img_format($scfg, $volname);
+	    $local_volumes->{$volid}->{snapshots} = defined($snaprefs) || ($format =~ /^(?:qcow2|vmdk)$/);
 	    if (defined($snaprefs)) {
 		# we cannot migrate shapshots on local storage
 		# exceptions: 'zfspool' or 'qcow2' files (on directory storage)
 
-		my $format = PVE::QemuServer::qemu_img_format($scfg, $volname);
 		die "online storage migration not possible if snapshot exists\n" if $self->{running};
 		if (!($scfg->{type} eq 'zfspool' || $format eq 'qcow2')) {
 		    die "non-migratable snapshot exists\n";
@@ -335,12 +385,14 @@ sub sync_disks {
         });
 
 	foreach my $vol (sort keys %$local_volumes) {
-	    if ($local_volumes->{$vol} eq 'storage') {
+	    my $ref = $local_volumes->{$vol}->{ref};
+	    if ($ref eq 'storage') {
 		$self->log('info', "found local disk '$vol' (via storage)\n");
-	    } elsif ($local_volumes->{$vol} eq 'config') {
-		die "can't live migrate attached local disks without with-local-disks option\n" if $self->{running} && !$self->{opts}->{"with-local-disks"};
+	    } 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"};
 		$self->log('info', "found local disk '$vol' (in current VM config)\n");
-	    } elsif ($local_volumes->{$vol} eq 'snapshot') {
+	    } elsif ($ref eq 'snapshot') {
 		$self->log('info', "found local disk '$vol' (referenced by snapshot(s))\n");
 	    } else {
 		$self->log('info', "found local disk '$vol'\n");
@@ -354,10 +406,6 @@ sub sync_disks {
 	    $self->log('warn', "$err");
 	}
 
-	if ($self->{running} && !$sharedvm && !$self->{opts}->{targetstorage}) {
-	    $self->{opts}->{targetstorage} = 1; #use same sid for remote local
-	}
-
 	if ($abort) {
 	    die "can't migrate VM - check log\n";
 	}
@@ -379,16 +427,33 @@ sub sync_disks {
 	    }
 	}
 
+	my $rep_volumes;
+
 	$self->log('info', "copying disk images");
 
+	my $rep_cfg = PVE::ReplicationConfig->new();
+
+	if (my $jobcfg = $rep_cfg->find_local_replication_job($vmid, $self->{node})) {
+	    die "can't live migrate VM with replicated volumes\n" if $self->{running};
+	    my $start_time = time();
+	    my $logfunc = sub { my ($msg) = @_;  $self->log('info', $msg); };
+	    $rep_volumes = PVE::Replication::run_replication(
+	       'PVE::QemuConfig', $jobcfg, $start_time, $start_time, $logfunc);
+	    $self->{replicated_volumes} = $rep_volumes;
+	}
+
 	foreach my $volid (keys %$local_volumes) {
 	    my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
-	    if ($self->{running} && $self->{opts}->{targetstorage} && $local_volumes->{$volid} eq 'config') {
+	    my $targetsid = $override_targetsid // $sid;
+	    if ($self->{running} && $local_volumes->{$volid}->{ref} eq 'config') {
 		push @{$self->{online_local_volumes}}, $volid;
 	    } else {
+		next if $rep_volumes->{$volid};
 		push @{$self->{volumes}}, $volid;
 		my $insecure = $self->{opts}->{migration_type} eq 'insecure';
-		PVE::Storage::storage_migrate($self->{storecfg}, $volid, $self->{ssh_info}, $sid, undef, undef, undef, undef, $insecure);
+		my $with_snapshots = $local_volumes->{$volid}->{snapshots};
+		PVE::Storage::storage_migrate($self->{storecfg}, $volid, $self->{ssh_info}, $targetsid,
+					      undef, undef, undef, undef, $insecure, $with_snapshots);
 	    }
 	}
     };
@@ -488,8 +553,8 @@ sub phase2 {
 	push @$cmd, '--machine', $self->{forcemachine};
     }
 
-    if ($self->{opts}->{targetstorage}) {
-	push @$cmd, '--targetstorage', $self->{opts}->{targetstorage};
+    if ($self->{online_local_volumes}) {
+	push @$cmd, '--targetstorage', ($self->{opts}->{targetstorage} // '1');
     }
 
     my $spice_port;
@@ -534,8 +599,9 @@ sub phase2 {
 
     die "unable to detect remote migration address\n" if !$raddr;
 
+    $self->log('info', "start remote tunnel");
+
     if ($migration_type eq 'secure') {
-	$self->log('info', "start remote tunnel");
 
 	if ($ruri =~ /^unix:/) {
 	    unlink $raddr;
@@ -568,11 +634,14 @@ sub phase2 {
 	} 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();
     }
 
     my $start = time();
 
-    if ($self->{opts}->{targetstorage} && defined($self->{online_local_volumes})) {
+    if (defined($self->{online_local_volumes})) {
 	$self->{storage_migration} = 1;
 	$self->{storage_migration_jobs} = {};
 	$self->log('info', "starting storage migration");
@@ -581,7 +650,7 @@ sub phase2 {
 	    if (scalar(keys %{$self->{target_drive}}) != scalar @{$self->{online_local_volumes}});
 	foreach my $drive (keys %{$self->{target_drive}}){
 	    my $nbd_uri = $self->{target_drive}->{$drive}->{nbd_uri};
-	    $self->log('info', "$drive: start migration to to $nbd_uri");
+	    $self->log('info', "$drive: start migration to $nbd_uri");
 	    PVE::QemuServer::qemu_drive_mirror($vmid, $drive, $nbd_uri, $vmid, undef, $self->{storage_migration_jobs}, 1);
 	}
     }
@@ -619,8 +688,11 @@ sub phase2 {
     };
     warn $@ if $@;
 
-    #set cachesize 10% of the total memory
-    my $cachesize = int($conf->{memory}*1048576/10);
+    # 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', "set cachesize: $cachesize");
     eval {
 	PVE::QemuServer::vm_mon_cmd_nocheck($vmid, "migrate-set-cache-size", value => int($cachesize));
@@ -655,7 +727,7 @@ sub phase2 {
     $self->log('info', "migrate uri => $ruri failed: $merr") if $merr;
 
     my $lstat = 0;
-    my $usleep = 2000000;
+    my $usleep = 1000000;
     my $i = 0;
     my $err_count = 0;
     my $lastrem = undef;
@@ -691,7 +763,7 @@ sub phase2 {
 	    if ($stat->{status} eq 'completed') {
 		my $delay = time() - $start;
 		if ($delay > 0) {
-		    my $mbps = sprintf "%.2f", $conf->{memory}/$delay;
+		    my $mbps = sprintf "%.2f", $memory / $delay;
 		    my $downtime = $stat->{downtime} || 0;
 		    $self->log('info', "migration speed: $mbps MB/s - downtime $downtime ms");
 		}
@@ -716,8 +788,8 @@ sub phase2 {
 		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 if lower than the everage transfert 
-		$usleep = 300000 if $avglstat && $rem < $avglstat;
+		# 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 ${total})");
@@ -751,16 +823,6 @@ sub phase2 {
 	    die "unable to parse migration status '$stat->{status}' - aborting\n";
 	}
     }
-
-    # just to be sure that the tunnel gets closed on successful migration, on error
-    # phase2_cleanup closes it *after* stopping the remote waiting VM
-    if (!$self->{errors} && $self->{tunnel}) {
-	eval { finish_tunnel($self, $self->{tunnel});  };
-	if (my $err = $@) {
-	    $self->log('err', $err);
-	    $self->{errors} = 1;
-	}
-    }
 }
 
 sub phase2_cleanup {
@@ -839,6 +901,8 @@ sub phase3_cleanup {
     my $conf = $self->{vmconf};
     return if $self->{phase2errors};
 
+    my $tunnel = $self->{tunnel};
+
     if ($self->{storage_migration}) {
 	# finish block-job
 	eval { PVE::QemuServer::qemu_drive_mirror_monitor($vmid, undef, $self->{storage_migration_jobs}); };
@@ -856,6 +920,9 @@ sub phase3_cleanup {
 	}
     }
 
+    # 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, $self->{node});
@@ -863,13 +930,10 @@ sub phase3_cleanup {
     die "Failed to move config to node '$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}) {
-	    # remove drives referencing the nbd server from source
-	    # otherwise vm_stop might hang later on
-	    foreach my $drive (keys %{$self->{target_drive}}){
-		PVE::QemuServer::vm_mon_cmd_nocheck($vmid, "device_del", id => $drive);
-	    }
 	    # stop nbd server on remote vm - requirement for resume since 2.9
 	    my $cmd = [@{$self->{rem_ssh}}, 'qm', 'nbdstop', $vmid];
 
@@ -879,14 +943,38 @@ sub phase3_cleanup {
 		$self->{errors} = 1;
 	    }
 	}
+
 	# config moved and nbd server stopped - now we can resume vm on target
-	my $cmd = [@{$self->{rem_ssh}}, 'qm', 'resume', $vmid, '--skiplock', '--nocheck'];
-	eval{ PVE::Tools::run_command($cmd, outfunc => sub {}, 
-		errfunc => sub {
-		    my $line = shift;
-        	    $self->log('err', $line);
-		});
-	};
+	if ($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;
@@ -950,4 +1038,9 @@ sub final_cleanup {
     # nothing to do
 }
 
+sub round_powerof2 {
+    return 1 if $_[0] < 2;
+    return 2 << int(log($_[0]-1)/log(2));
+}
+
 1;