[pve-devel] [PATCH v4 storage 2/7] Add prune_backups to storage API

[Fabian Ebner <f.ebner@proxmox.com>]

Implement it for generic storages supporting backups
(i.e. directory-based storages) and add a wrapper for PBS.

Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
---

Changes from v3:
    * prune_mark: merge loops
    * allow filtering by type
    * more error handling
    * allow passing along a log function
    * add more tests

 PVE/Storage.pm             |  91 +++++++++-
 PVE/Storage/PBSPlugin.pm   |  68 ++++++++
 PVE/Storage/Plugin.pm      |  65 +++++++
 test/prune_backups_test.pm | 342 +++++++++++++++++++++++++++++++++++++
 test/run_plugin_tests.pl   |   1 +
 5 files changed, 565 insertions(+), 2 deletions(-)
 create mode 100644 test/prune_backups_test.pm

diff --git a/PVE/Storage.pm b/PVE/Storage.pm
index edf9a2e..5c44386 100755
--- a/PVE/Storage.pm
+++ b/PVE/Storage.pm
@@ -40,11 +40,11 @@ use PVE::Storage::DRBDPlugin;
 use PVE::Storage::PBSPlugin;
 
 # Storage API version. Icrement it on changes in storage API interface.
-use constant APIVER => 5;
+use constant APIVER => 6;
 # Age is the number of versions we're backward compatible with.
 # This is like having 'current=APIVER' and age='APIAGE' in libtool,
 # see https://www.gnu.org/software/libtool/manual/html_node/Libtool-versioning.html
-use constant APIAGE => 4;
+use constant APIAGE => 5;
 
 # load standard plugins
 PVE::Storage::DirPlugin->register();
@@ -1540,6 +1540,93 @@ sub extract_vzdump_config {
     }
 }
 
+sub prune_backups {
+    my ($cfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc) = @_;
+
+    my $scfg = storage_config($cfg, $storeid);
+    die "storage '$storeid' does not support backups\n" if !$scfg->{content}->{backup};
+
+    if (!defined($keep)) {
+	die "no prune-backups options configured for storage '$storeid'\n"
+	    if !defined($scfg->{'prune-backups'});
+	$keep = PVE::JSONSchema::parse_property_string('prune-backups', $scfg->{'prune-backups'});
+    }
+
+    my $plugin = PVE::Storage::Plugin->lookup($scfg->{type});
+    return $plugin->prune_backups($scfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc);
+}
+
+my $prune_mark = sub {
+    my ($prune_entries, $keep_count, $id_func) = @_;
+
+    return if !$keep_count;
+
+    my $already_included = {};
+    my $newly_included = {};
+
+    foreach my $prune_entry (@{$prune_entries}) {
+	my $mark = $prune_entry->{mark};
+	my $id = $id_func->($prune_entry->{ctime});
+
+	next if $already_included->{$id};
+
+	if (defined($mark)) {
+	    $already_included->{$id} = 1 if $mark eq 'keep';
+	    next;
+	}
+
+	if (!$newly_included->{$id}) {
+	    last if scalar(keys %{$newly_included}) >= $keep_count;
+	    $newly_included->{$id} = 1;
+	    $prune_entry->{mark} = 'keep';
+	} else {
+	    $prune_entry->{mark} = 'remove';
+	}
+    }
+};
+
+sub prune_mark_backup_group {
+    my ($backup_group, $keep) = @_;
+
+    my $prune_list = [ sort { $b->{ctime} <=> $a->{ctime} } @{$backup_group} ];
+
+    $prune_mark->($prune_list, $keep->{'keep-last'}, sub {
+	my ($ctime) = @_;
+	return $ctime;
+    });
+    $prune_mark->($prune_list, $keep->{'keep-hourly'}, sub {
+	my ($ctime) = @_;
+	my (undef, undef, $hour, $day, $month, $year) = localtime($ctime);
+	return "$hour/$day/$month/$year";
+    });
+    $prune_mark->($prune_list, $keep->{'keep-daily'}, sub {
+	my ($ctime) = @_;
+	my (undef, undef, undef, $day, $month, $year) = localtime($ctime);
+	return "$day/$month/$year";
+    });
+    $prune_mark->($prune_list, $keep->{'keep-weekly'}, sub {
+	my ($ctime) = @_;
+	my ($sec, $min, $hour, $day, $month, $year) = localtime($ctime);
+	my $iso_week = int(strftime("%V", $sec, $min, $hour, $day, $month - 1, $year - 1900));
+	my $iso_week_year = int(strftime("%G", $sec, $min, $hour, $day, $month - 1, $year - 1900));
+	return "$iso_week/$iso_week_year";
+    });
+    $prune_mark->($prune_list, $keep->{'keep-monthly'}, sub {
+	my ($ctime) = @_;
+	my (undef, undef, undef, undef, $month, $year) = localtime($ctime);
+	return "$month/$year";
+    });
+    $prune_mark->($prune_list, $keep->{'keep-yearly'}, sub {
+	my ($ctime) = @_;
+	my $year = (localtime($ctime))[5];
+	return "$year";
+    });
+
+    foreach my $prune_entry (@{$prune_list}) {
+	$prune_entry->{mark} //= 'remove';
+    }
+}
+
 sub volume_export {
     my ($cfg, $fh, $volid, $format, $snapshot, $base_snapshot, $with_snapshots) = @_;
 
diff --git a/PVE/Storage/PBSPlugin.pm b/PVE/Storage/PBSPlugin.pm
index 87b09f2..12f56a9 100644
--- a/PVE/Storage/PBSPlugin.pm
+++ b/PVE/Storage/PBSPlugin.pm
@@ -258,6 +258,74 @@ sub extract_vzdump_config {
     return $config;
 }
 
+sub prune_backups {
+    my ($class, $scfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc) = @_;
+
+    $logfunc //= sub { print "$_[1]\n" };
+
+    my $backups = $class->list_volumes($storeid, $scfg, $vmid, ['backup']);
+
+    $type = 'vm' if defined($type) && $type eq 'qemu';
+    $type = 'ct' if defined($type) && $type eq 'lxc';
+
+    my $backup_groups = {};
+    foreach my $backup (@{$backups}) {
+	(my $backup_type = $backup->{format}) =~ s/^pbs-//;
+
+	next if defined($type) && $backup_type ne $type;
+
+	my $backup_group = "$backup_type/$backup->{vmid}";
+	$backup_groups->{$backup_group} = 1;
+    }
+
+    my @param;
+    foreach my $opt (keys %{$keep}) {
+	push @param, "--$opt";
+	push @param, "$keep->{$opt}";
+    }
+
+    push @param, '--dry-run' if $dryrun;
+
+    my $prune_list = [];
+    my $failed;
+
+    foreach my $backup_group (keys %{$backup_groups}) {
+	$logfunc->('info', "running 'proxmox-backup-client prune' for '$backup_group'")
+	    if !$dryrun;
+	eval {
+	    my $res = run_client_cmd($scfg, $storeid, 'prune', [ $backup_group, @param ]);
+
+	    foreach my $backup (@{$res}) {
+		die "result from proxmox-backup-client is not as expected\n"
+		    if !defined($backup->{'backup-time'})
+		    || !defined($backup->{'backup-type'})
+		    || !defined($backup->{'backup-id'})
+		    || !defined($backup->{'keep'});
+
+		my $ctime = $backup->{'backup-time'};
+		my $type = $backup->{'backup-type'};
+		my $vmid = $backup->{'backup-id'};
+		my $volid = print_volid($storeid, $type, $vmid, $ctime);
+
+		push @{$prune_list}, {
+		    ctime => $ctime,
+		    mark => $backup->{keep} ? 'keep' : 'remove',
+		    type => $type eq 'vm' ? 'qemu' : 'lxc',
+		    vmid => $vmid,
+		    volid => $volid,
+		};
+	    }
+	};
+	if (my $err = $@) {
+	    $logfunc->('err', "prune '$backup_group': $err\n");
+	    $failed = 1;
+	}
+    }
+    die "error pruning backups - check log\n" if $failed;
+
+    return $prune_list;
+}
+
 my $autogen_encryption_key = sub {
     my ($scfg, $storeid) = @_;
     my $encfile = pbs_encryption_key_file_name($scfg, $storeid);
diff --git a/PVE/Storage/Plugin.pm b/PVE/Storage/Plugin.pm
index 6e321ae..8a58ff4 100644
--- a/PVE/Storage/Plugin.pm
+++ b/PVE/Storage/Plugin.pm
@@ -1174,6 +1174,71 @@ sub check_connection {
     return 1;
 }
 
+sub prune_backups {
+    my ($class, $scfg, $storeid, $keep, $vmid, $type, $dryrun, $logfunc) = @_;
+
+    $logfunc //= sub { print "$_[1]\n" };
+
+    my $backups = $class->list_volumes($storeid, $scfg, $vmid, ['backup']);
+
+    my $backup_groups = {};
+    my $prune_list = [];
+
+    foreach my $backup (@{$backups}) {
+	my $volid = $backup->{volid};
+	my $backup_vmid = $backup->{vmid};
+	my $archive_info = eval { PVE::Storage::archive_info($volid) } // {};
+	my $backup_type = $archive_info->{type} // 'unknown';
+
+	next if defined($type) && $type ne $backup_type;
+
+	my $prune_entry = {
+	    ctime => $backup->{ctime},
+	    type => $backup_type,
+	    volid => $volid,
+	};
+
+	$prune_entry->{vmid} = $backup_vmid if defined($backup_vmid);
+
+	if ($archive_info->{is_std_name}) {
+	    $prune_entry->{ctime} = $archive_info->{ctime};
+	    my $group = "$backup_type/$backup_vmid";
+	    push @{$backup_groups->{$group}}, $prune_entry;
+	} else {
+	    # ignore backups that don't use the standard naming scheme
+	    $prune_entry->{mark} = 'protected';
+	}
+
+	push @{$prune_list}, $prune_entry;
+    }
+
+    foreach my $backup_group (values %{$backup_groups}) {
+	PVE::Storage::prune_mark_backup_group($backup_group, $keep);
+    }
+
+    my $failed;
+    if (!$dryrun) {
+	foreach my $prune_entry (@{$prune_list}) {
+	    next if $prune_entry->{mark} ne 'remove';
+
+	    my $volid = $prune_entry->{volid};
+	    $logfunc->('info', "removing backup '$volid'");
+	    eval {
+		my (undef, $volname) = parse_volume_id($volid);
+		my $archive_path = $class->filesystem_path($scfg, $volname);
+		PVE::Storage::archive_remove($archive_path);
+	    };
+	    if (my $err = $@) {
+		$logfunc->('err', "error when removing backup '$volid' - $err\n");
+		$failed = 1;
+	    }
+	}
+    }
+    die "error pruning backups - check log\n" if $failed;
+
+    return $prune_list;
+}
+
 # Import/Export interface:
 #   Any path based storage is assumed to support 'raw' and 'tar' streams, so
 #   the default implementations will return this if $scfg->{path} is set,
diff --git a/test/prune_backups_test.pm b/test/prune_backups_test.pm
new file mode 100644
index 0000000..1e6a3d1
--- /dev/null
+++ b/test/prune_backups_test.pm
@@ -0,0 +1,342 @@
+package PVE::Storage::TestPruneBackups;
+
+use strict;
+use warnings;
+
+use lib qw(..);
+
+use PVE::Storage;
+use Test::More;
+use Test::MockModule;
+
+my $storeid = 'BackTest123';
+my @vmids = (1234, 9001);
+
+# only includes the information needed for prune_backups
+my $mocked_backups_lists = {};
+
+my $basetime = 1577881101; # 2020_01_01-12_18_21 UTC
+
+foreach my $vmid (@vmids) {
+    push @{$mocked_backups_lists->{default}}, (
+	{
+	    'volid' => "$storeid:backup/vzdump-qemu-$vmid-2018_05_26-11_18_21.tar.zst",
+	    'ctime' => $basetime - 585*24*60*60 - 60*60,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-qemu-$vmid-2019_12_31-11_18_21.tar.zst",
+	    'ctime' => $basetime - 24*60*60 - 60*60,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-qemu-$vmid-2019_12_31-11_19_21.tar.zst",
+	    'ctime' => $basetime - 24*60*60 - 60*60 + 60,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-qemu-$vmid-2020_01_01-11_18_21.tar.zst",
+	    'ctime' => $basetime - 60*60,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-qemu-$vmid-2020_01_01-12_18_21.tar.zst",
+	    'ctime' => $basetime,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-lxc-$vmid-2020_01_01-12_18_21.tar.zst",
+	    'ctime' => $basetime,
+	    'vmid'  => $vmid,
+	},
+	{
+	    'volid' => "$storeid:backup/vzdump-$vmid-renamed.tar.zst",
+	    'ctime' => 1234,
+	    'vmid'  => $vmid,
+	},
+    );
+}
+push @{$mocked_backups_lists->{year1970}}, (
+    {
+	'volid' => "$storeid:backup/vzdump-lxc-321-1970_01_01-00_01_23.tar.zst",
+	'ctime' => 83,
+	'vmid'  => 321,
+    },
+    {
+	'volid' => "$storeid:backup/vzdump-lxc-321-2070_01_01-00_01_00.tar.zst",
+	'ctime' => 60*60*24 * (365*100 + 25) + 60,
+	'vmid'  => 321,
+    },
+);
+push @{$mocked_backups_lists->{novmid}}, (
+    {
+	'volid' => "$storeid:backup/vzdump-lxc-novmid.tar.gz",
+	'ctime' => 1234,
+    },
+);
+my $current_list;
+my $mock_plugin = Test::MockModule->new('PVE::Storage::Plugin');
+$mock_plugin->redefine(list_volumes => sub {
+    my ($class, $storeid, $scfg, $vmid, $content_types) = @_;
+
+    my $list = $mocked_backups_lists->{$current_list};
+
+    return $list if !defined($vmid);
+
+    return [ grep { $_->{vmid} eq $vmid } @{$list} ];
+});
+
+sub generate_expected {
+    my ($vmids, $type, $marks) = @_;
+
+    my @expected;
+    foreach my $vmid (@{$vmids}) {
+	push @expected, (
+	    {
+		'volid' => "$storeid:backup/vzdump-qemu-$vmid-2018_05_26-11_18_21.tar.zst",
+		'type'  => 'qemu',
+		'ctime' => $basetime - 585*24*60*60 - 60*60,
+		'mark'  => $marks->[0],
+		'vmid'  => $vmid,
+	    },
+	    {
+		'volid' => "$storeid:backup/vzdump-qemu-$vmid-2019_12_31-11_18_21.tar.zst",
+		'type'  => 'qemu',
+		'ctime' => $basetime - 24*60*60 - 60*60,
+		'mark'  => $marks->[1],
+		'vmid'  => $vmid,
+	    },
+	    {
+		'volid' => "$storeid:backup/vzdump-qemu-$vmid-2019_12_31-11_19_21.tar.zst",
+		'type'  => 'qemu',
+		'ctime' => $basetime - 24*60*60 - 60*60 + 60,
+		'mark'  => $marks->[2],
+		'vmid'  => $vmid,
+	    },
+	    {
+		'volid' => "$storeid:backup/vzdump-qemu-$vmid-2020_01_01-11_18_21.tar.zst",
+		'type'  => 'qemu',
+		'ctime' => $basetime - 60*60,
+		'mark'  => $marks->[3],
+		'vmid'  => $vmid,
+	    },
+	    {
+		'volid' => "$storeid:backup/vzdump-qemu-$vmid-2020_01_01-12_18_21.tar.zst",
+		'type'  => 'qemu',
+		'ctime' => $basetime,
+		'mark'  => $marks->[4],
+		'vmid'  => $vmid,
+	    },
+	) if !defined($type) || $type eq 'qemu';
+	push @expected, (
+	    {
+		'volid' => "$storeid:backup/vzdump-lxc-$vmid-2020_01_01-12_18_21.tar.zst",
+		'type'  => 'lxc',
+		'ctime' => $basetime,
+		'mark'  => $marks->[5],
+		'vmid'  => $vmid,
+	    },
+	) if !defined($type) || $type eq 'lxc';
+	push @expected, (
+	    {
+		'volid' => "$storeid:backup/vzdump-$vmid-renamed.tar.zst",
+		'type'  => 'unknown',
+		'ctime' => 1234,
+		'mark'  => 'protected',
+		'vmid'  => $vmid,
+	    },
+	) if !defined($type);
+    }
+    return [ sort { $a->{volid} cmp $b->{volid} } @expected ];
+}
+
+# an array of test cases, each test is comprised of the following keys:
+# description   => to identify a single test
+# vmid          => VMID or undef for all
+# type          => 'qemu' or 'lxc' or undef for all
+# keep          => options describing what to keep
+# list          => backups list to use. defaults to 'default'
+# expected      => what prune_backups should return
+#
+# most of them are created further below
+my $tests = [
+    {
+	description => 'last=3, multiple IDs',
+	keep => {
+	    'keep-last' => 3,
+	},
+	expected => generate_expected(\@vmids, undef, ['remove', 'remove', 'keep', 'keep', 'keep', 'keep']),
+    },
+    {
+	description => 'weekly=2, one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-weekly' => 2,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['keep', 'remove', 'remove', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'daily=weekly=monthly=1, multiple IDs',
+	keep => {
+	    'keep-hourly' => 0,
+	    'keep-daily' => 1,
+	    'keep-weekly' => 1,
+	    'keep-monthly' => 1,
+	},
+	expected => generate_expected(\@vmids, undef, ['keep', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'hourly=4, one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-hourly' => 4,
+	    'keep-daily' => 0,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['keep', 'remove', 'keep', 'keep', 'keep', 'keep']),
+    },
+    {
+	description => 'yearly=2, multiple IDs',
+	keep => {
+	    'keep-hourly' => 0,
+	    'keep-daily' => 0,
+	    'keep-weekly' => 0,
+	    'keep-monthly' => 0,
+	    'keep-yearly' => 2,
+	},
+	expected => generate_expected(\@vmids, undef, ['remove', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'last=2,hourly=2 one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-last' => 2,
+	    'keep-hourly' => 2,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['keep', 'remove', 'keep', 'keep', 'keep', 'keep']),
+    },
+    {
+	description => 'last=1,monthly=2, multiple IDs',
+	keep => {
+	    'keep-last' => 1,
+	    'keep-monthly' => 2,
+	},
+	expected => generate_expected(\@vmids, undef, ['keep', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'monthly=3, one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-monthly' => 3,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['keep', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'last=daily=weekly=1, multiple IDs',
+	keep => {
+	    'keep-last' => 1,
+	    'keep-daily' => 1,
+	    'keep-weekly' => 1,
+	},
+	expected => generate_expected(\@vmids, undef, ['keep', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'daily=2, one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-daily' => 2,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['remove', 'remove', 'keep', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'weekly=monthly=1, multiple IDs',
+	keep => {
+	    'keep-weekly' => 1,
+	    'keep-monthly' => 1,
+	},
+	expected => generate_expected(\@vmids, undef, ['keep', 'remove', 'remove', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'weekly=yearly=1, one ID',
+	vmid => $vmids[0],
+	keep => {
+	    'keep-weekly' => 1,
+	    'keep-yearly' => 1,
+	},
+	expected => generate_expected([$vmids[0]], undef, ['keep', 'remove', 'remove', 'remove', 'keep', 'keep']),
+    },
+    {
+	description => 'weekly=yearly=1, one ID, type qemu',
+	vmid => $vmids[0],
+	type => 'qemu',
+	keep => {
+	    'keep-weekly' => 1,
+	    'keep-yearly' => 1,
+	},
+	expected => generate_expected([$vmids[0]], 'qemu', ['keep', 'remove', 'remove', 'remove', 'keep', '']),
+    },
+    {
+	description => 'week=yearly=1, one ID, type lxc',
+	vmid => $vmids[0],
+	type => 'lxc',
+	keep => {
+	    'keep-last' => 1,
+	},
+	expected => generate_expected([$vmids[0]], 'lxc', ['', '', '', '', '', 'keep']),
+    },
+    {
+	description => 'yearly=1, year before 2000',
+	keep => {
+	    'keep-yearly' => 1,
+	},
+	list => 'year1970',
+	expected => [
+	    {
+		'volid' => "$storeid:backup/vzdump-lxc-321-1970_01_01-00_01_23.tar.zst",
+		'ctime' => 83,
+		'mark'  => 'remove',
+		'type'  => 'lxc',
+		'vmid'  => 321,
+	    },
+	    {
+		'volid' => "$storeid:backup/vzdump-lxc-321-2070_01_01-00_01_00.tar.zst",
+		'ctime' => 60*60*24 * (365*100 + 25) + 60,
+		'mark'  => 'keep',
+		'type'  => 'lxc',
+		'vmid'  => 321,
+	    },
+	],
+    },
+    {
+	description => 'last=1, ne ID, year before 2000',
+	keep => {
+	    'keep-last' => 1,
+	},
+	list => 'novmid',
+	expected => [
+	    {
+		'volid' => "$storeid:backup/vzdump-lxc-novmid.tar.gz",
+		'ctime' => 1234,
+		'mark'  => 'protected',
+		'type'  => 'lxc',
+	    },
+	],
+    },
+];
+
+plan tests => scalar @$tests;
+
+for my $tt (@$tests) {
+
+    my $got = eval {
+	$current_list = $tt->{list} // 'default';
+	my $res = PVE::Storage::Plugin->prune_backups($tt->{scfg}, $storeid, $tt->{keep}, $tt->{vmid}, $tt->{type}, 1);
+	return [ sort { $a->{volid} cmp $b->{volid} } @{$res} ];
+    };
+    $got = $@ if $@;
+
+    is_deeply($got, $tt->{expected}, $tt->{description}) || diag(explain($got));
+}
+
+done_testing();
+
+1;
diff --git a/test/run_plugin_tests.pl b/test/run_plugin_tests.pl
index 54322bb..d33429a 100755
--- a/test/run_plugin_tests.pl
+++ b/test/run_plugin_tests.pl
@@ -16,6 +16,7 @@ my $res = $harness->runtests(
     "path_to_volume_id_test.pm",
     "get_subdir_test.pm",
     "filesystem_path_test.pm",
+    "prune_backups_test.pm",
 );
 
 exit -1 if !$res || $res->{failed} || $res->{parse_errors};
-- 
2.20.1