]>
Commit | Line | Data |
---|---|---|
c62eb5b2 FG |
1 | package PVE::CLI::pve5to6; |
2 | ||
3 | use strict; | |
4 | use warnings; | |
5 | ||
6 | use PVE::API2::APT; | |
7 | use PVE::API2::Ceph; | |
8 | use PVE::API2::LXC; | |
9 | use PVE::API2::Qemu; | |
70effe27 | 10 | use PVE::API2::Certificates; |
c62eb5b2 FG |
11 | |
12 | use PVE::Ceph::Tools; | |
13 | use PVE::Cluster; | |
14 | use PVE::Corosync; | |
15 | use PVE::INotify; | |
16 | use PVE::JSONSchema; | |
17 | use PVE::RPCEnvironment; | |
18 | use PVE::Storage; | |
e9ad9ab4 | 19 | use PVE::Tools qw(run_command $IPV4RE $IPV6RE); |
eb7fda99 | 20 | use PVE::QemuServer; |
a0841973 | 21 | use PVE::QemuConfig; |
c62eb5b2 | 22 | |
aa517822 | 23 | use AptPkg::Cache; |
e9ad9ab4 | 24 | use Socket qw(AF_INET AF_INET6 inet_ntop); |
23bba5a4 FG |
25 | use Term::ANSIColor; |
26 | ||
c62eb5b2 FG |
27 | use PVE::CLIHandler; |
28 | ||
29 | use base qw(PVE::CLIHandler); | |
30 | ||
31 | my $nodename = PVE::INotify::nodename(); | |
32 | ||
33 | sub setup_environment { | |
34 | PVE::RPCEnvironment->setup_default_cli_env(); | |
35 | } | |
36 | ||
6afec5da TL |
37 | my $min_pve_major = 5; |
38 | my $min_pve_minor = 4; | |
c62eb5b2 FG |
39 | my $min_pve_pkgrel = 2; |
40 | ||
41 | my $counters = { | |
42 | pass => 0, | |
43 | skip => 0, | |
44 | warn => 0, | |
45 | fail => 0, | |
46 | }; | |
47 | ||
48 | my $log_line = sub { | |
49 | my ($level, $line) = @_; | |
50 | ||
51 | $counters->{$level}++ if defined($level) && defined($counters->{$level}); | |
52 | ||
53 | print uc($level), ': ' if defined($level); | |
54 | print "$line\n"; | |
55 | }; | |
56 | ||
57 | sub log_pass { | |
23bba5a4 | 58 | print color('green'); |
c62eb5b2 | 59 | $log_line->('pass', @_); |
23bba5a4 | 60 | print color('reset'); |
c62eb5b2 FG |
61 | } |
62 | ||
63 | sub log_info { | |
64 | $log_line->('info', @_); | |
65 | } | |
66 | sub log_skip { | |
67 | $log_line->('skip', @_); | |
68 | } | |
69 | sub log_warn { | |
23bba5a4 | 70 | print color('yellow'); |
c62eb5b2 | 71 | $log_line->('warn', @_); |
23bba5a4 | 72 | print color('reset'); |
c62eb5b2 FG |
73 | } |
74 | sub log_fail { | |
23bba5a4 | 75 | print color('red'); |
c62eb5b2 | 76 | $log_line->('fail', @_); |
23bba5a4 | 77 | print color('reset'); |
c62eb5b2 FG |
78 | } |
79 | ||
f3d40afa TL |
80 | my $print_header_first = 1; |
81 | sub print_header { | |
82 | my ($h) = @_; | |
83 | print "\n" if !$print_header_first; | |
84 | print "= $h =\n\n"; | |
85 | $print_header_first = 0; | |
86 | } | |
87 | ||
57d96361 TL |
88 | my $get_systemd_unit_state = sub { |
89 | my ($unit) = @_; | |
90 | ||
91 | my $state; | |
92 | my $filter_output = sub { | |
93 | $state = shift; | |
94 | chomp $state; | |
95 | }; | |
96 | eval { | |
97 | run_command(['systemctl', 'is-enabled', "$unit"], outfunc => $filter_output, noerr => 1); | |
98 | return if !defined($state); | |
99 | run_command(['systemctl', 'is-active', "$unit"], outfunc => $filter_output, noerr => 1); | |
100 | }; | |
101 | ||
102 | return $state // 'unknown'; | |
103 | }; | |
104 | my $log_systemd_unit_state = sub { | |
105 | my ($unit, $no_fail_on_inactive) = @_; | |
106 | ||
107 | my $log_method = \&log_warn; | |
108 | ||
109 | my $state = $get_systemd_unit_state->($unit); | |
110 | if ($state eq 'active') { | |
111 | $log_method = \&log_pass; | |
112 | } elsif ($state eq 'inactive') { | |
113 | $log_method = $no_fail_on_inactive ? \&log_warn : \&log_fail; | |
114 | } elsif ($state eq 'failed') { | |
115 | $log_method = \&log_fail; | |
116 | } | |
117 | ||
118 | $log_method->("systemd unit '$unit' is in state '$state'"); | |
119 | }; | |
120 | ||
663ae4d8 | 121 | my $versions; |
c62eb5b2 FG |
122 | my $get_pkg = sub { |
123 | my ($pkg) = @_; | |
124 | ||
663ae4d8 | 125 | $versions = eval { PVE::API2::APT->versions({ node => $nodename }) } if !defined($versions); |
c62eb5b2 FG |
126 | |
127 | if (!defined($versions)) { | |
128 | my $msg = "unable to retrieve package version information"; | |
129 | $msg .= "- $@" if $@; | |
130 | log_fail("$msg"); | |
131 | return undef; | |
132 | } | |
133 | ||
134 | my $pkgs = [ grep { $_->{Package} eq $pkg } @$versions ]; | |
135 | if (!defined $pkgs || $pkgs == 0) { | |
136 | log_fail("unable to determine installed $pkg version."); | |
137 | return undef; | |
138 | } else { | |
139 | return $pkgs->[0]; | |
140 | } | |
141 | }; | |
142 | ||
797ba640 TL |
143 | my $apt_cache; |
144 | my $get_apt_package_state = sub { | |
145 | my ($pkg) = @_; | |
146 | ||
147 | if (!$apt_cache) { | |
148 | if (!($apt_cache = AptPkg::Cache->new())) { | |
149 | log_fail("unable to initialize AptPkg::Cache\n"); # should not happen at all | |
150 | return undef; | |
151 | } | |
152 | } | |
153 | my $p = $apt_cache->{'linux-image-amd64'}; | |
154 | if ($p) { | |
155 | return $p->{SelectedState}; | |
156 | } | |
157 | return undef; | |
158 | }; | |
159 | ||
160 | my $is_pkg_installed = sub { | |
161 | my ($pkg) = @_; | |
162 | ||
163 | my $state = $get_apt_package_state->($pkg); | |
164 | return undef if !defined($state); # better die? | |
165 | return lc($state) eq 'install'; | |
166 | }; | |
167 | ||
e9ad9ab4 FG |
168 | # taken from pve-cluster 6.0-4 |
169 | my $resolve_hostname_like_corosync = sub { | |
170 | my ($hostname, $corosync_conf) = @_; | |
171 | ||
172 | my $corosync_strategy = $corosync_conf->{main}->{totem}->{ip_version}; | |
173 | $corosync_strategy = lc ($corosync_strategy // "ipv6-4"); | |
174 | ||
175 | my $match_ip_and_version = sub { | |
176 | my ($addr) = @_; | |
177 | ||
178 | return undef if !defined($addr); | |
179 | ||
180 | if ($addr =~ m/^$IPV4RE$/) { | |
181 | return ($addr, 4); | |
182 | } elsif ($addr =~ m/^$IPV6RE$/) { | |
183 | return ($addr, 6); | |
184 | } | |
185 | ||
186 | return undef; | |
187 | }; | |
188 | ||
189 | my ($resolved_ip, $ip_version) = $match_ip_and_version->($hostname); | |
190 | ||
191 | return ($resolved_ip, $ip_version) if defined($resolved_ip); | |
192 | ||
193 | my $resolved_ip4; | |
194 | my $resolved_ip6; | |
195 | ||
196 | my @resolved_raw; | |
197 | eval { @resolved_raw = PVE::Tools::getaddrinfo_all($hostname); }; | |
198 | ||
199 | return undef if ($@ || !@resolved_raw); | |
200 | ||
201 | foreach my $socket_info (@resolved_raw) { | |
202 | next if !$socket_info->{addr}; | |
203 | ||
204 | my ($family, undef, $host) = PVE::Tools::unpack_sockaddr_in46($socket_info->{addr}); | |
205 | ||
206 | if ($family == AF_INET && !defined($resolved_ip4)) { | |
207 | $resolved_ip4 = inet_ntop(AF_INET, $host); | |
208 | } elsif ($family == AF_INET6 && !defined($resolved_ip6)) { | |
209 | $resolved_ip6 = inet_ntop(AF_INET6, $host); | |
210 | } | |
211 | ||
212 | last if defined($resolved_ip4) && defined($resolved_ip6); | |
213 | } | |
214 | ||
215 | # corosync_strategy specifies the order in which IP addresses are resolved | |
216 | # by corosync. We need to match that order, to ensure we create firewall | |
217 | # rules for the correct address family. | |
218 | if ($corosync_strategy eq "ipv4") { | |
219 | $resolved_ip = $resolved_ip4; | |
220 | } elsif ($corosync_strategy eq "ipv6") { | |
221 | $resolved_ip = $resolved_ip6; | |
222 | } elsif ($corosync_strategy eq "ipv6-4") { | |
223 | $resolved_ip = $resolved_ip6 // $resolved_ip4; | |
224 | } elsif ($corosync_strategy eq "ipv4-6") { | |
225 | $resolved_ip = $resolved_ip4 // $resolved_ip6; | |
226 | } | |
227 | ||
228 | return $match_ip_and_version->($resolved_ip); | |
229 | }; | |
230 | ||
c62eb5b2 | 231 | sub check_pve_packages { |
f3d40afa | 232 | print_header("CHECKING VERSION INFORMATION FOR PVE PACKAGES"); |
c62eb5b2 | 233 | |
f3d40afa | 234 | print "Checking for package updates..\n"; |
c62eb5b2 FG |
235 | my $updates = eval { PVE::API2::APT->list_updates({ node => $nodename }); }; |
236 | if (!defined($updates)) { | |
237 | log_warn("$@") if $@; | |
238 | log_fail("unable to retrieve list of package updates!"); | |
239 | } elsif (@$updates > 0) { | |
240 | my $pkgs = join(', ', map { $_->{Package} } @$updates); | |
3f5db7fa | 241 | log_warn("updates for the following packages are available:\n $pkgs"); |
c62eb5b2 FG |
242 | } else { |
243 | log_pass("all packages uptodate"); | |
244 | } | |
245 | ||
246 | print "\nChecking proxmox-ve package version..\n"; | |
247 | if (defined(my $proxmox_ve = $get_pkg->('proxmox-ve'))) { | |
6afec5da | 248 | my $min_pve_ver = "$min_pve_major.$min_pve_minor-$min_pve_pkgrel"; |
c62eb5b2 | 249 | |
6afec5da TL |
250 | my ($maj, $min, $pkgrel) = $proxmox_ve->{OldVersion} =~ m/^(\d+)\.(\d+)-(\d+)/; |
251 | ||
86a3955d TL |
252 | my $upgraded = 0; |
253 | ||
6afec5da TL |
254 | if ($maj > $min_pve_major) { |
255 | log_pass("already upgraded to Proxmox VE " . ($min_pve_major + 1)); | |
86a3955d | 256 | $upgraded = 1; |
6afec5da | 257 | } elsif ($maj >= $min_pve_major && $min >= $min_pve_minor && $pkgrel >= $min_pve_pkgrel) { |
c62eb5b2 FG |
258 | log_pass("proxmox-ve package has version >= $min_pve_ver"); |
259 | } else { | |
260 | log_fail("proxmox-ve package is too old, please upgrade to >= $min_pve_ver!"); | |
261 | } | |
86a3955d TL |
262 | |
263 | my ($krunning, $kinstalled) = (qr/5\./, 'pve-kernel-5.0'); | |
264 | if (!$upgraded) { | |
265 | ($krunning, $kinstalled) = (qr/4\.15/, 'pve-kernel-4.15'); | |
266 | } | |
267 | ||
7f4b3ea5 FG |
268 | print "\nChecking running kernel version..\n"; |
269 | my $kernel_ver = $proxmox_ve->{RunningKernel}; | |
270 | if (!defined($kernel_ver)) { | |
271 | log_fail("unable to determine running kernel version."); | |
86a3955d | 272 | } elsif ($kernel_ver =~ /^$krunning/) { |
7f4b3ea5 | 273 | log_pass("expected running kernel '$kernel_ver'."); |
86a3955d TL |
274 | } elsif ($get_pkg->($kinstalled)) { |
275 | log_warn("expected kernel '$kinstalled' intalled but not yet rebooted!"); | |
7f4b3ea5 | 276 | } else { |
86a3955d | 277 | log_warn("unexpected running and installed kernel '$kernel_ver'."); |
7f4b3ea5 | 278 | } |
aa517822 DC |
279 | |
280 | } | |
797ba640 TL |
281 | print "\nChecking for installed stock Debian Kernel..\n"; |
282 | if ($is_pkg_installed->('linux-image-amd64')) { | |
283 | log_fail("Stock Debian kernel package installed. Please remove package 'linux-image-amd64'."); | |
aa517822 | 284 | } else { |
797ba640 | 285 | log_pass("Stock Debian kernel package not installed."); |
c62eb5b2 FG |
286 | } |
287 | } | |
288 | ||
eb7fda99 DC |
289 | sub get_vms_with_vmx { |
290 | my $res = { | |
291 | cpu => [], | |
292 | flag => [], | |
293 | }; | |
294 | my $vmlist = PVE::QemuServer::vzlist(); | |
450957c0 | 295 | |
eb7fda99 DC |
296 | foreach my $vmid ( sort { $a <=> $b } keys %$vmlist ) { |
297 | my $pid = $vmlist->{$vmid}->{pid}; | |
298 | next if !$pid; # skip not running vms | |
450957c0 | 299 | |
eb7fda99 DC |
300 | my $cmdline = eval { PVE::Tools::file_get_contents("/proc/$pid/cmdline") }; |
301 | if ($cmdline) { | |
302 | my @args = split(/\0/, $cmdline); | |
303 | for (my $i = 0; $i < scalar(@args); $i++) { | |
304 | next if !$args[$i] || $args[$i] !~ m/^-?-cpu$/; | |
450957c0 | 305 | |
eb7fda99 DC |
306 | my $cpuarg = $args[$i+1]; |
307 | if ($cpuarg =~ m/^(host|max)/) { | |
308 | push @{$res->{cpu}}, $vmid; | |
309 | } elsif ($cpuarg =~ m/\+(vmx|svm)/) { | |
310 | push @{$res->{flag}}, $vmid; | |
311 | } | |
312 | } | |
313 | } | |
314 | } | |
450957c0 TL |
315 | |
316 | $res = undef if (scalar(@{$res->{cpu}}) + scalar(@{$res->{flag}})) <= 0; | |
317 | ||
318 | return $res; | |
eb7fda99 DC |
319 | } |
320 | ||
7bf30805 | 321 | sub check_kvm_nested { |
450957c0 TL |
322 | log_info("Checking KVM nesting support, which breaks live migration for VMs using it.."); |
323 | ||
7bf30805 DC |
324 | my $module_sysdir = "/sys/module"; |
325 | if (-e "$module_sysdir/kvm_amd") { | |
326 | $module_sysdir .= "/kvm_amd/parameters"; | |
327 | } elsif (-e "$module_sysdir/kvm_intel") { | |
328 | $module_sysdir .= "/kvm_intel/parameters"; | |
329 | } else { | |
330 | log_skip("no kvm module found"); | |
331 | return; | |
332 | } | |
333 | ||
334 | if (-f "$module_sysdir/nested") { | |
335 | my $val = eval { PVE::Tools::file_read_firstline("$module_sysdir/nested") }; | |
336 | if ($val && $val =~ m/Y|1/) { | |
450957c0 TL |
337 | my $list = get_vms_with_vmx(); |
338 | if (!defined($list)) { | |
339 | log_pass("KVM nested parameter set, but currently no VM with a 'vmx' or 'svm' flag is running."); | |
eb7fda99 | 340 | } else { |
450957c0 | 341 | my $warnmsg = "KVM nested enabled. It will not be possible to live migrate the following running VMs to PVE 6:\n"; |
eb7fda99 | 342 | if (@{$list->{cpu}}) { |
450957c0 | 343 | $warnmsg .= " VMID(s) with cputype 'host' or 'max': " . join(',', @{$list->{cpu}}) . "\n"; |
eb7fda99 | 344 | } |
eb7fda99 | 345 | if (@{$list->{flag}}) { |
450957c0 | 346 | $warnmsg .= " VMID(s) with enforced cpu flag 'vmx' or 'svm': " . join(',', @{$list->{flag}}) . "\n"; |
eb7fda99 | 347 | } |
eb7fda99 DC |
348 | log_warn($warnmsg); |
349 | } | |
7bf30805 DC |
350 | } else { |
351 | log_pass("KVM nested parameter not set.") | |
352 | } | |
353 | } else { | |
354 | log_skip("KVM nested parameter not found."); | |
355 | } | |
356 | } | |
357 | ||
a0841973 DC |
358 | sub check_vms_with_uefi { |
359 | log_info("Checking VMs with OVMF enabled, which may need manual intervention..."); | |
360 | ||
361 | my $vmlist = PVE::QemuServer::vzlist(); | |
362 | ||
363 | my $vms = []; | |
364 | ||
365 | foreach my $vmid ( sort { $a <=> $b } keys %$vmlist ) { | |
366 | my $conf = PVE::QemuConfig->load_config($vmid); | |
367 | if ($conf->{bios} && $conf->{bios} eq 'ovmf' && $conf->{efidisk0}) { | |
368 | my $disk = PVE::QemuServer::parse_drive('efidisk0', $conf->{efidisk0}); | |
369 | if (!defined($disk->{size}) || $disk->{size} > 128*1024) { | |
370 | # all efidisks bigger than the default 128k and those | |
371 | # without size in the config | |
372 | push @$vms, $vmid; | |
373 | } elsif ($disk->{file} !~ /\.(raw|qcow2|vmdk)$/) { | |
374 | # all efidisks not on file storage | |
375 | push @$vms, $vmid; | |
376 | } | |
377 | } | |
378 | } | |
379 | ||
380 | if (scalar(@$vms) > 0) { | |
381 | my $warnmsg = "VMs with OVMF configured and potentially broken EFI disks: \n"; | |
382 | $warnmsg .= " " . join(',', @$vms); | |
383 | $warnmsg .= "\nThere may be manual intervention required. See Known upgrade issues for details\n"; | |
384 | log_warn($warnmsg); | |
385 | } else { | |
386 | log_pass("No VMs with OVMF and potentially broken EFI disk found."); | |
387 | } | |
388 | } | |
389 | ||
c62eb5b2 | 390 | sub check_storage_health { |
f3d40afa | 391 | print_header("CHECKING CONFIGURED STORAGES"); |
c62eb5b2 FG |
392 | my $cfg = PVE::Storage::config(); |
393 | ||
394 | my $ctime = time(); | |
395 | ||
396 | my $info = PVE::Storage::storage_info($cfg); | |
397 | ||
398 | foreach my $storeid (keys %$info) { | |
399 | my $d = $info->{$storeid}; | |
400 | if ($d->{enabled}) { | |
cd772160 | 401 | if ($d->{type} eq 'sheepdog') { |
92c64229 | 402 | log_fail("storage '$storeid' of type 'sheepdog' is enabled - experimental sheepdog support dropped in PVE 6") |
cd772160 | 403 | } elsif ($d->{active}) { |
c62eb5b2 FG |
404 | log_pass("storage '$storeid' enabled and active."); |
405 | } else { | |
406 | log_warn("storage '$storeid' enabled but not active!"); | |
407 | } | |
408 | } else { | |
409 | log_skip("storage '$storeid' disabled."); | |
410 | } | |
411 | } | |
412 | } | |
413 | ||
414 | sub check_cluster_corosync { | |
f3d40afa | 415 | print_header("CHECKING CLUSTER HEALTH/SETTINGS"); |
c62eb5b2 FG |
416 | |
417 | if (!PVE::Corosync::check_conf_exists(1)) { | |
418 | log_skip("standalone node."); | |
419 | return; | |
420 | } | |
421 | ||
57d96361 TL |
422 | $log_systemd_unit_state->('pve-cluster.service'); |
423 | $log_systemd_unit_state->('corosync.service'); | |
424 | ||
c62eb5b2 | 425 | if (PVE::Cluster::check_cfs_quorum(1)) { |
084625c2 | 426 | log_pass("Cluster Filesystem is quorate."); |
c62eb5b2 | 427 | } else { |
084625c2 | 428 | log_fail("Cluster Filesystem readonly, lost quorum?!"); |
c62eb5b2 FG |
429 | } |
430 | ||
431 | my $conf = PVE::Cluster::cfs_read_file('corosync.conf'); | |
432 | my $conf_nodelist = PVE::Corosync::nodelist($conf); | |
fbcc7737 | 433 | my $node_votes = 0; |
c62eb5b2 | 434 | |
77f8b841 | 435 | print "\nAnalzying quorum settings and state..\n"; |
c62eb5b2 FG |
436 | if (!defined($conf_nodelist)) { |
437 | log_fail("unable to retrieve nodelist from corosync.conf"); | |
fbcc7737 FG |
438 | } else { |
439 | if (grep { $conf_nodelist->{$_}->{quorum_votes} != 1 } keys %$conf_nodelist) { | |
440 | log_warn("non-default quorum_votes distribution detected!"); | |
441 | } | |
442 | map { $node_votes += $conf_nodelist->{$_}->{quorum_votes} // 0 } keys %$conf_nodelist; | |
443 | } | |
444 | ||
445 | my ($expected_votes, $total_votes); | |
446 | my $filter_output = sub { | |
447 | my $line = shift; | |
448 | ($expected_votes) = $line =~ /^Expected votes:\s*(\d+)\s*$/ | |
449 | if !defined($expected_votes); | |
450 | ($total_votes) = $line =~ /^Total votes:\s*(\d+)\s*$/ | |
451 | if !defined($total_votes); | |
452 | }; | |
453 | eval { | |
b60916cc | 454 | run_command(['corosync-quorumtool', '-s'], outfunc => $filter_output, noerr => 1); |
fbcc7737 FG |
455 | }; |
456 | ||
457 | if (!defined($expected_votes)) { | |
458 | log_fail("unable to get expected number of votes, setting to 0."); | |
459 | $expected_votes = 0; | |
460 | } | |
461 | if (!defined($total_votes)) { | |
462 | log_fail("unable to get expected number of votes, setting to 0."); | |
a6fc492a | 463 | $total_votes = 0; |
c62eb5b2 FG |
464 | } |
465 | ||
466 | my $cfs_nodelist = PVE::Cluster::get_clinfo()->{nodelist}; | |
467 | my $offline_nodes = grep { $cfs_nodelist->{$_}->{online} != 1 } keys %$cfs_nodelist; | |
468 | if ($offline_nodes > 0) { | |
469 | log_fail("$offline_nodes nodes are offline!"); | |
470 | } | |
471 | ||
fbcc7737 FG |
472 | my $qdevice_votes = 0; |
473 | if (my $qdevice_setup = $conf->{main}->{quorum}->{device}) { | |
474 | $qdevice_votes = $qdevice_setup->{votes} // 1; | |
475 | } | |
476 | ||
477 | log_info("configured votes - nodes: $node_votes"); | |
478 | log_info("configured votes - qdevice: $qdevice_votes"); | |
479 | log_info("current expected votes: $expected_votes"); | |
480 | log_info("current total votes: $total_votes"); | |
481 | ||
482 | log_warn("expected votes set to non-standard value '$expected_votes'.") | |
483 | if $expected_votes != $node_votes + $qdevice_votes; | |
484 | log_warn("total votes < expected votes: $total_votes/$expected_votes!") | |
485 | if $total_votes < $expected_votes; | |
486 | ||
c62eb5b2 FG |
487 | my $conf_nodelist_count = scalar(keys %$conf_nodelist); |
488 | my $cfs_nodelist_count = scalar(keys %$cfs_nodelist); | |
489 | log_warn("cluster consists of less than three nodes!") | |
fbcc7737 | 490 | if $conf_nodelist_count < 3 && $conf_nodelist_count + $qdevice_votes < 3; |
c62eb5b2 FG |
491 | |
492 | log_fail("corosync.conf ($conf_nodelist_count) and pmxcfs ($cfs_nodelist_count) don't agree about size of nodelist.") | |
493 | if $conf_nodelist_count != $cfs_nodelist_count; | |
494 | ||
77f8b841 | 495 | print "\nChecking nodelist entries..\n"; |
c62eb5b2 FG |
496 | foreach my $cs_node (keys %$conf_nodelist) { |
497 | my $entry = $conf_nodelist->{$cs_node}; | |
b3a6a2ae | 498 | log_fail("$cs_node: no name entry in corosync.conf.") |
c62eb5b2 | 499 | if !defined($entry->{name}); |
b3a6a2ae | 500 | log_fail("$cs_node: no nodeid configured in corosync.conf.") |
c62eb5b2 | 501 | if !defined($entry->{nodeid}); |
a6524bbe FG |
502 | log_fail("$cs_node: neither ring0_addr nor ring1_addr defined in corosync.conf.") |
503 | if !defined($entry->{ring0_addr}) && !defined($entry->{ring1_addr}); | |
c62eb5b2 FG |
504 | |
505 | my $verify_ring_ip = sub { | |
506 | my $key = shift; | |
507 | my $ring = $entry->{$key}; | |
b78378f3 FG |
508 | if (defined($ring)) { |
509 | my ($resolved_ip, undef) = $resolve_hostname_like_corosync->($ring, $conf); | |
510 | if (defined($resolved_ip)) { | |
511 | if ($resolved_ip ne $ring) { | |
b3a6a2ae | 512 | log_warn("$cs_node: $key '$ring' resolves to '$resolved_ip'.\n Consider replacing it with the currently resolved IP address."); |
b78378f3 | 513 | } else { |
b3a6a2ae | 514 | log_pass("$cs_node: $key is configured to use IP address '$ring'"); |
b78378f3 FG |
515 | } |
516 | } else { | |
b3a6a2ae | 517 | log_fail("$cs_node: unable to resolve $key '$ring' to an IP address according to Corosync's resolve strategy - cluster will potentially fail with Corosync 3.x/kronosnet!"); |
b78378f3 | 518 | } |
c62eb5b2 FG |
519 | } |
520 | }; | |
521 | $verify_ring_ip->('ring0_addr'); | |
522 | $verify_ring_ip->('ring1_addr'); | |
523 | } | |
524 | ||
77f8b841 | 525 | print "\nChecking totem settings..\n"; |
c62eb5b2 | 526 | my $totem = $conf->{main}->{totem}; |
c62eb5b2 FG |
527 | my $transport = $totem->{transport}; |
528 | if (defined($transport)) { | |
e0505600 FG |
529 | if ($transport ne 'knet') { |
530 | log_fail("Corosync transport explicitly set to '$transport' instead of implicit default!"); | |
531 | } else { | |
532 | log_pass("Corosync transport set to '$transport'."); | |
533 | } | |
534 | } else { | |
535 | log_pass("Corosync transport set to implicit default."); | |
c62eb5b2 FG |
536 | } |
537 | ||
538 | if ((!defined($totem->{secauth}) || $totem->{secauth} ne 'on') && (!defined($totem->{crypto_cipher}) || $totem->{crypto_cipher} eq 'none')) { | |
539 | log_fail("Corosync authentication/encryption is not explicitly enabled (secauth / crypto_cipher / crypto_hash)!"); | |
e0505600 FG |
540 | } else { |
541 | if (defined($totem->{crypto_cipher}) && $totem->{crypto_cipher} eq '3des') { | |
542 | log_fail("Corosync encryption cipher set to '3des', no longer supported in Corosync 3.x!"); | |
543 | } else { | |
544 | log_pass("Corosync encryption and authentication enabled."); | |
545 | } | |
c62eb5b2 FG |
546 | } |
547 | ||
02e79acc | 548 | print "\n"; |
0e25ab80 | 549 | log_info("run 'pvecm status' to get detailed cluster status.."); |
c62eb5b2 | 550 | |
f3d40afa | 551 | print_header("CHECKING INSTALLED COROSYNC VERSION"); |
c62eb5b2 FG |
552 | if (defined(my $corosync = $get_pkg->('corosync'))) { |
553 | if ($corosync->{OldVersion} =~ m/^2\./) { | |
554 | log_fail("corosync 2.x installed, cluster-wide upgrade to 3.x needed!"); | |
555 | } elsif ($corosync->{OldVersion} =~ m/^3\./) { | |
556 | log_pass("corosync 3.x installed."); | |
557 | } else { | |
558 | log_fail("unexpected corosync version installed: $corosync->{OldVersion}!"); | |
559 | } | |
560 | } | |
561 | } | |
562 | ||
563 | sub check_ceph { | |
f3d40afa | 564 | print_header("CHECKING HYPER-CONVERGED CEPH STATUS"); |
c62eb5b2 FG |
565 | |
566 | if (PVE::Ceph::Tools::check_ceph_inited(1)) { | |
567 | log_info("hyper-converged ceph setup detected!"); | |
568 | } else { | |
569 | log_skip("no hyper-converged ceph setup detected!"); | |
570 | return; | |
571 | } | |
572 | ||
573 | log_info("getting Ceph status/health information.."); | |
574 | my $ceph_status = eval { PVE::API2::Ceph->status({ node => $nodename }); }; | |
575 | my $osd_flags = eval { PVE::API2::Ceph->get_flags({ node => $nodename }); }; | |
a75241fb FG |
576 | my $noout_wanted = 1; |
577 | my $noout = $osd_flags =~ m/noout/ if $osd_flags; | |
c62eb5b2 FG |
578 | |
579 | if (!$ceph_status || !$ceph_status->{health}) { | |
580 | log_fail("unable to determine Ceph status!"); | |
581 | } else { | |
582 | my $ceph_health = $ceph_status->{health}->{status}; | |
583 | if (!$ceph_health) { | |
584 | log_fail("unable to determine Ceph health!"); | |
585 | } elsif ($ceph_health eq 'HEALTH_OK') { | |
586 | log_pass("Ceph health reported as 'HEALTH_OK'."); | |
587 | } elsif ($ceph_health eq 'HEALTH_WARN' && $noout && (keys %{$ceph_status->{health}->{checks}} == 1)) { | |
588 | log_pass("Ceph health reported as 'HEALTH_WARN' with a single failing check and 'noout' flag set."); | |
589 | } else { | |
dc7eea01 TL |
590 | log_warn("Ceph health reported as '$ceph_health'.\n Use the PVE ". |
591 | "dashboard or 'ceph -s' to determine the specific issues and try to resolve them."); | |
c62eb5b2 FG |
592 | } |
593 | } | |
594 | ||
595 | log_info("getting Ceph OSD flags.."); | |
596 | eval { | |
597 | if (!$osd_flags) { | |
598 | log_fail("unable to get Ceph OSD flags!"); | |
599 | } else { | |
600 | if ($osd_flags =~ m/recovery_deletes/ && $osd_flags =~ m/purged_snapdirs/) { | |
601 | log_pass("all PGs have been scrubbed at least once while running Ceph Luminous."); | |
602 | } else { | |
603 | log_fail("missing 'recovery_deletes' and/or 'purged_snapdirs' flag, scrub of all PGs required before upgrading to Nautilus!"); | |
604 | } | |
c62eb5b2 FG |
605 | } |
606 | }; | |
607 | ||
608 | log_info("getting Ceph daemon versions.."); | |
609 | my $ceph_versions = eval { PVE::Ceph::Tools::get_cluster_versions(undef, 1); }; | |
610 | if (!$ceph_versions) { | |
611 | log_fail("unable to determine Ceph daemon versions!"); | |
612 | } else { | |
613 | my $services = [ | |
614 | { 'key' => 'mon', 'name' => 'monitor' }, | |
615 | { 'key' => 'mgr', 'name' => 'manager' }, | |
616 | { 'key' => 'mds', 'name' => 'MDS' }, | |
617 | { 'key' => 'osd', 'name' => 'OSD' }, | |
618 | ]; | |
619 | ||
620 | foreach my $service (@$services) { | |
621 | my $name = $service->{name}; | |
622 | if (my $service_versions = $ceph_versions->{$service->{key}}) { | |
623 | if (keys %$service_versions == 0) { | |
624 | log_skip("no running instances detected for daemon type $name."); | |
625 | } elsif (keys %$service_versions == 1) { | |
626 | log_pass("single running version detected for daemon type $name."); | |
627 | } else { | |
628 | log_warn("multiple running versions detected for daemon type $name!"); | |
629 | } | |
630 | } else { | |
631 | log_skip("unable to determine versions of running Ceph $name instances."); | |
632 | } | |
633 | } | |
634 | ||
635 | my $overall_versions = $ceph_versions->{overall}; | |
636 | if (!$overall_versions) { | |
637 | log_warn("unable to determine overall Ceph daemon versions!"); | |
638 | } elsif (keys %$overall_versions == 1) { | |
639 | log_pass("single running overall version detected for all Ceph daemon types."); | |
a75241fb FG |
640 | if ((keys %$overall_versions)[0] =~ /^ceph version 14\./) { |
641 | $noout_wanted = 0; | |
642 | } | |
c62eb5b2 FG |
643 | } else { |
644 | log_warn("overall version mismatch detected, check 'ceph versions' output for details!"); | |
645 | } | |
646 | } | |
5d2ae292 | 647 | |
a75241fb FG |
648 | if ($noout) { |
649 | if ($noout_wanted) { | |
7a6994bc | 650 | log_pass("'noout' flag set to prevent rebalancing during cluster-wide upgrades."); |
a75241fb | 651 | } else { |
7a6994bc | 652 | log_warn("'noout' flag set, Ceph cluster upgrade seems finished."); |
a75241fb FG |
653 | } |
654 | } elsif ($noout_wanted) { | |
7a6994bc | 655 | log_warn("'noout' flag not set - recommended to prevent rebalancing during upgrades."); |
a75241fb FG |
656 | } |
657 | ||
42bbe3ed DC |
658 | log_info("checking Ceph config.."); |
659 | my $conf = PVE::Cluster::cfs_read_file('ceph.conf'); | |
90ec2ddb | 660 | if (%$conf) { |
42bbe3ed | 661 | my $global = $conf->{global}; |
8ade3b35 TL |
662 | |
663 | my $global_monhost = $global->{mon_host} // $global->{"mon host"} // $global->{"mon-host"}; | |
664 | if (!defined($global_monhost)) { | |
665 | log_warn("No 'mon_host' entry found in ceph config.\n It's recommended to add mon_host with all monitor addresses (without ports) to the global section."); | |
42bbe3ed | 666 | } else { |
7a6994bc | 667 | log_pass("Found 'mon_host' entry."); |
42bbe3ed | 668 | } |
c0d772c4 | 669 | |
6cc9ab61 TL |
670 | my $ipv6 = $global->{ms_bind_ipv6} // $global->{"ms bind ipv6"} // $global->{"ms-bind-ipv6"}; |
671 | if ($ipv6) { | |
672 | my $ipv4 = $global->{ms_bind_ipv4} // $global->{"ms bind ipv4"} // $global->{"ms-bind-ipv4"}; | |
c0d772c4 | 673 | if ($ipv6 eq 'true' && (!defined($ipv4) || $ipv4 ne 'false')) { |
7a6994bc | 674 | log_warn("'ms_bind_ipv6' is enabled but 'ms_bind_ipv4' is not disabled.\n Make sure to disable 'ms_bind_ipv4' for ipv6 only clusters, or add an ipv4 network to public/cluster network."); |
c0d772c4 | 675 | } else { |
7a6994bc | 676 | log_pass("'ms_bind_ipv6' is enabled and 'ms_bind_ipv4' disabled"); |
c0d772c4 DC |
677 | } |
678 | } else { | |
7a6994bc | 679 | log_pass("'ms_bind_ipv6' not enabled"); |
c0d772c4 | 680 | } |
a500f71b FG |
681 | |
682 | if (defined($global->{keyring})) { | |
683 | log_warn("[global] config section contains 'keyring' option, which will prevent services from starting with Nautilus.\n Move 'keyring' option to [client] section instead."); | |
684 | } else { | |
685 | log_pass("no 'keyring' option in [global] section found."); | |
686 | } | |
687 | ||
42bbe3ed | 688 | } else { |
90ec2ddb | 689 | log_warn("Empty ceph config found"); |
42bbe3ed DC |
690 | } |
691 | ||
5d2ae292 FG |
692 | my $local_ceph_ver = PVE::Ceph::Tools::get_local_version(1); |
693 | if (defined($local_ceph_ver)) { | |
694 | if ($local_ceph_ver == 14) { | |
695 | my $scanned_osds = PVE::Tools::dir_glob_regex('/etc/ceph/osd', '^.*\.json$'); | |
696 | if (-e '/var/lib/ceph/osd/' && !defined($scanned_osds)) { | |
697 | log_warn("local Ceph version is Nautilus, local OSDs detected, but no conversion from ceph-disk to ceph-volume done (yet)."); | |
698 | } | |
699 | } | |
700 | } else { | |
701 | log_fail("unable to determine local Ceph version."); | |
702 | } | |
c62eb5b2 FG |
703 | } |
704 | ||
705 | sub check_misc { | |
f3d40afa | 706 | print_header("MISCELLANEOUS CHECKS"); |
c62eb5b2 | 707 | my $ssh_config = eval { PVE::Tools::file_get_contents('/root/.ssh/config') }; |
5d100641 FG |
708 | if (defined($ssh_config)) { |
709 | log_fail("Unsupported SSH Cipher configured for root in /root/.ssh/config: $1") | |
710 | if $ssh_config =~ /^Ciphers .*(blowfish|arcfour|3des).*$/m; | |
711 | } else { | |
712 | log_skip("No SSH config file found."); | |
713 | } | |
c62eb5b2 | 714 | |
e5196e84 TL |
715 | log_info("Checking common daemon services.."); |
716 | $log_systemd_unit_state->('pveproxy.service'); | |
717 | $log_systemd_unit_state->('pvedaemon.service'); | |
718 | $log_systemd_unit_state->('pvestatd.service'); | |
719 | ||
c62eb5b2 FG |
720 | my $root_free = PVE::Tools::df('/', 10); |
721 | log_warn("Less than 2G free space on root file system.") | |
722 | if defined($root_free) && $root_free->{avail} < 2*1024*1024*1024; | |
723 | ||
22ad5ba6 | 724 | log_info("Checking for running guests.."); |
c62eb5b2 | 725 | my $running_guests = 0; |
22ad5ba6 | 726 | |
c62eb5b2 FG |
727 | my $vms = eval { PVE::API2::Qemu->vmlist({ node => $nodename }) }; |
728 | log_warn("Failed to retrieve information about this node's VMs - $@") if $@; | |
22ad5ba6 TL |
729 | $running_guests += grep { $_->{status} eq 'running' } @$vms if defined($vms); |
730 | ||
c62eb5b2 FG |
731 | my $cts = eval { PVE::API2::LXC->vmlist({ node => $nodename }) }; |
732 | log_warn("Failed to retrieve information about this node's CTs - $@") if $@; | |
22ad5ba6 TL |
733 | $running_guests += grep { $_->{status} eq 'running' } @$cts if defined($cts); |
734 | ||
735 | if ($running_guests > 0) { | |
736 | log_warn("$running_guests running guest(s) detected - consider migrating or stopping them.") | |
737 | } else { | |
738 | log_pass("no running guest detected.") | |
739 | } | |
9c2f3b73 | 740 | |
33866e3c | 741 | log_info("Checking if the local node's hostname '$nodename' is resolvable.."); |
f94e33c0 | 742 | my $local_ip = eval { PVE::Network::get_ip_from_hostname($nodename) }; |
9c2f3b73 | 743 | if ($@) { |
f94e33c0 | 744 | log_warn("Failed to resolve hostname '$nodename' to IP - $@"); |
9c2f3b73 | 745 | } else { |
329cf22a | 746 | log_info("Checking if resolved IP is configured on local node.."); |
9c2f3b73 ML |
747 | my $cidr = Net::IP::ip_is_ipv6($local_ip) ? "$local_ip/128" : "$local_ip/32"; |
748 | my $configured_ips = PVE::Network::get_local_ip_from_cidr($cidr); | |
749 | my $ip_count = scalar(@$configured_ips); | |
ebd4de69 TL |
750 | |
751 | if ($ip_count <= 0) { | |
f94e33c0 | 752 | log_fail("Resolved node IP '$local_ip' not configured or active for '$nodename'"); |
ebd4de69 TL |
753 | } elsif ($ip_count > 1) { |
754 | log_warn("Resolved node IP '$local_ip' active on multiple ($ip_count) interfaces!"); | |
22ad5ba6 | 755 | } else { |
329cf22a | 756 | log_pass("Resolved node IP '$local_ip' configured and active on single interface."); |
ebd4de69 | 757 | } |
9c2f3b73 | 758 | } |
b0a8f08b | 759 | |
12511b12 | 760 | log_info("Check node certificate's RSA key size"); |
70effe27 | 761 | my $certs = PVE::API2::Certificates->info({ node => $nodename }); |
12511b12 TL |
762 | my $certs_check = { |
763 | 'rsaEncryption' => { | |
764 | minsize => 2048, | |
765 | name => 'RSA', | |
766 | }, | |
767 | 'id-ecPublicKey' => { | |
768 | minsize => 224, | |
769 | name => 'ECC', | |
770 | }, | |
771 | }; | |
772 | ||
773 | my $certs_check_failed = 0; | |
774 | foreach my $cert (@$certs) { | |
775 | my ($type, $size, $fn) = $cert->@{qw(public-key-type public-key-bits filename)}; | |
776 | ||
777 | if (!defined($type) || !defined($size)) { | |
778 | log_warn("'$fn': cannot check certificate, failed to get it's type or size!"); | |
779 | } | |
780 | ||
781 | my $check = $certs_check->{$type}; | |
782 | if (!defined($check)) { | |
783 | log_warn("'$fn': certificate's public key type '$type' unknown, check Debian Busters release notes"); | |
784 | next; | |
785 | } | |
786 | ||
787 | if ($size < $check->{minsize}) { | |
788 | log_fail("'$fn', certificate's $check->{name} public key size is less than 2048 bit"); | |
789 | $certs_check_failed = 1; | |
790 | } else { | |
791 | log_pass("Certificate '$fn' passed Debian Busters security level for TLS connections ($size >= 2048)"); | |
70effe27 AA |
792 | } |
793 | } | |
70effe27 | 794 | |
b0a8f08b | 795 | check_kvm_nested(); |
a0841973 DC |
796 | |
797 | check_vms_with_uefi(); | |
c62eb5b2 FG |
798 | } |
799 | ||
800 | __PACKAGE__->register_method ({ | |
801 | name => 'checklist', | |
802 | path => 'checklist', | |
803 | method => 'GET', | |
804 | description => 'Check (pre-/post-)upgrade conditions.', | |
805 | parameters => { | |
806 | additionalProperties => 0, | |
807 | properties => { | |
808 | }, | |
809 | }, | |
810 | returns => { type => 'null' }, | |
811 | code => sub { | |
812 | my ($param) = @_; | |
813 | ||
814 | check_pve_packages(); | |
815 | check_cluster_corosync(); | |
816 | check_ceph(); | |
817 | check_storage_health(); | |
818 | check_misc(); | |
819 | ||
f3d40afa | 820 | print_header("SUMMARY"); |
f814c077 TL |
821 | |
822 | my $total = 0; | |
823 | $total += $_ for values %$counters; | |
824 | ||
825 | print "TOTAL: $total\n"; | |
52d469af TL |
826 | print colored("PASSED: $counters->{pass}\n", 'green'); |
827 | print "SKIPPED: $counters->{skip}\n"; | |
23bba5a4 FG |
828 | print colored("WARNINGS: $counters->{warn}\n", 'yellow'); |
829 | print colored("FAILURES: $counters->{fail}\n", 'red'); | |
c62eb5b2 | 830 | |
e35e351f TL |
831 | if ($counters->{warn} > 0 || $counters->{fail} > 0) { |
832 | my $color = $counters->{fail} > 0 ? 'red' : 'yellow'; | |
833 | print colored("\nATTENTION: Please check the output for detailed information!\n", $color); | |
64ef9363 | 834 | print colored("Try to solve the problems one at a time and then run this checklist tool again.\n", $color) if $counters->{fail} > 0; |
e35e351f | 835 | } |
c62eb5b2 FG |
836 | |
837 | return undef; | |
838 | }}); | |
839 | ||
89abf04d TL |
840 | our $cmddef = [ __PACKAGE__, 'checklist', [], {}]; |
841 | ||
842 | # for now drop all unknown params and just check | |
843 | @ARGV = (); | |
c62eb5b2 FG |
844 | |
845 | 1; |