]>
Commit | Line | Data |
---|---|---|
4177a14d TL |
1 | package PVE::CLI::pve6to7; |
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; | |
10 | use PVE::API2::Certificates; | |
11 | ||
31e4ad5d | 12 | use PVE::AccessControl; |
4177a14d TL |
13 | use PVE::Ceph::Tools; |
14 | use PVE::Cluster; | |
15 | use PVE::Corosync; | |
16 | use PVE::INotify; | |
17 | use PVE::JSONSchema; | |
1be86a36 | 18 | use PVE::NodeConfig; |
4177a14d TL |
19 | use PVE::RPCEnvironment; |
20 | use PVE::Storage; | |
31e4ad5d | 21 | use PVE::Tools qw(run_command split_list); |
1be86a36 | 22 | use PVE::QemuConfig; |
4177a14d | 23 | use PVE::QemuServer; |
44090816 | 24 | use PVE::VZDump::Common; |
4177a14d TL |
25 | |
26 | use Term::ANSIColor; | |
27 | ||
28 | use PVE::CLIHandler; | |
29 | ||
30 | use base qw(PVE::CLIHandler); | |
31 | ||
32 | my $nodename = PVE::INotify::nodename(); | |
33 | ||
34 | sub setup_environment { | |
35 | PVE::RPCEnvironment->setup_default_cli_env(); | |
36 | } | |
37 | ||
38 | my $min_pve_major = 6; | |
39 | my $min_pve_minor = 4; | |
40 | my $min_pve_pkgrel = 1; | |
41 | ||
42 | my $counters = { | |
43 | pass => 0, | |
44 | skip => 0, | |
45 | warn => 0, | |
46 | fail => 0, | |
47 | }; | |
48 | ||
49 | my $log_line = sub { | |
50 | my ($level, $line) = @_; | |
51 | ||
52 | $counters->{$level}++ if defined($level) && defined($counters->{$level}); | |
53 | ||
54 | print uc($level), ': ' if defined($level); | |
55 | print "$line\n"; | |
56 | }; | |
57 | ||
58 | sub log_pass { | |
59 | print color('green'); | |
60 | $log_line->('pass', @_); | |
61 | print color('reset'); | |
62 | } | |
63 | ||
64 | sub log_info { | |
65 | $log_line->('info', @_); | |
66 | } | |
67 | sub log_skip { | |
68 | $log_line->('skip', @_); | |
69 | } | |
70 | sub log_warn { | |
71 | print color('yellow'); | |
72 | $log_line->('warn', @_); | |
73 | print color('reset'); | |
74 | } | |
75 | sub log_fail { | |
76 | print color('red'); | |
77 | $log_line->('fail', @_); | |
78 | print color('reset'); | |
79 | } | |
80 | ||
81 | my $print_header_first = 1; | |
82 | sub print_header { | |
83 | my ($h) = @_; | |
84 | print "\n" if !$print_header_first; | |
85 | print "= $h =\n\n"; | |
86 | $print_header_first = 0; | |
87 | } | |
88 | ||
89 | my $get_systemd_unit_state = sub { | |
90 | my ($unit) = @_; | |
91 | ||
92 | my $state; | |
93 | my $filter_output = sub { | |
94 | $state = shift; | |
95 | chomp $state; | |
96 | }; | |
97 | eval { | |
98 | run_command(['systemctl', 'is-enabled', "$unit"], outfunc => $filter_output, noerr => 1); | |
99 | return if !defined($state); | |
100 | run_command(['systemctl', 'is-active', "$unit"], outfunc => $filter_output, noerr => 1); | |
101 | }; | |
102 | ||
103 | return $state // 'unknown'; | |
104 | }; | |
105 | my $log_systemd_unit_state = sub { | |
106 | my ($unit, $no_fail_on_inactive) = @_; | |
107 | ||
108 | my $log_method = \&log_warn; | |
109 | ||
110 | my $state = $get_systemd_unit_state->($unit); | |
111 | if ($state eq 'active') { | |
112 | $log_method = \&log_pass; | |
113 | } elsif ($state eq 'inactive') { | |
114 | $log_method = $no_fail_on_inactive ? \&log_warn : \&log_fail; | |
115 | } elsif ($state eq 'failed') { | |
116 | $log_method = \&log_fail; | |
117 | } | |
118 | ||
119 | $log_method->("systemd unit '$unit' is in state '$state'"); | |
120 | }; | |
121 | ||
122 | my $versions; | |
123 | my $get_pkg = sub { | |
124 | my ($pkg) = @_; | |
125 | ||
126 | $versions = eval { PVE::API2::APT->versions({ node => $nodename }) } if !defined($versions); | |
127 | ||
128 | if (!defined($versions)) { | |
129 | my $msg = "unable to retrieve package version information"; | |
130 | $msg .= "- $@" if $@; | |
131 | log_fail("$msg"); | |
132 | return undef; | |
133 | } | |
134 | ||
135 | my $pkgs = [ grep { $_->{Package} eq $pkg } @$versions ]; | |
136 | if (!defined $pkgs || $pkgs == 0) { | |
137 | log_fail("unable to determine installed $pkg version."); | |
138 | return undef; | |
139 | } else { | |
140 | return $pkgs->[0]; | |
141 | } | |
142 | }; | |
143 | ||
144 | sub check_pve_packages { | |
145 | print_header("CHECKING VERSION INFORMATION FOR PVE PACKAGES"); | |
146 | ||
147 | print "Checking for package updates..\n"; | |
148 | my $updates = eval { PVE::API2::APT->list_updates({ node => $nodename }); }; | |
149 | if (!defined($updates)) { | |
150 | log_warn("$@") if $@; | |
151 | log_fail("unable to retrieve list of package updates!"); | |
152 | } elsif (@$updates > 0) { | |
153 | my $pkgs = join(', ', map { $_->{Package} } @$updates); | |
154 | log_warn("updates for the following packages are available:\n $pkgs"); | |
155 | } else { | |
156 | log_pass("all packages uptodate"); | |
157 | } | |
158 | ||
159 | print "\nChecking proxmox-ve package version..\n"; | |
160 | if (defined(my $proxmox_ve = $get_pkg->('proxmox-ve'))) { | |
161 | my $min_pve_ver = "$min_pve_major.$min_pve_minor-$min_pve_pkgrel"; | |
162 | ||
163 | my ($maj, $min, $pkgrel) = $proxmox_ve->{OldVersion} =~ m/^(\d+)\.(\d+)-(\d+)/; | |
164 | ||
165 | my $upgraded = 0; | |
166 | ||
167 | if ($maj > $min_pve_major) { | |
168 | log_pass("already upgraded to Proxmox VE " . ($min_pve_major + 1)); | |
169 | $upgraded = 1; | |
170 | } elsif ($maj >= $min_pve_major && $min >= $min_pve_minor && $pkgrel >= $min_pve_pkgrel) { | |
171 | log_pass("proxmox-ve package has version >= $min_pve_ver"); | |
172 | } else { | |
173 | log_fail("proxmox-ve package is too old, please upgrade to >= $min_pve_ver!"); | |
174 | } | |
175 | ||
176 | my ($krunning, $kinstalled) = (qr/5\.11/, 'pve-kernel-5.11'); | |
177 | if (!$upgraded) { | |
178 | ($krunning, $kinstalled) = (qr/5\.(?:4|11)/, 'pve-kernel-4.15'); | |
179 | } | |
180 | ||
181 | print "\nChecking running kernel version..\n"; | |
182 | my $kernel_ver = $proxmox_ve->{RunningKernel}; | |
183 | if (!defined($kernel_ver)) { | |
184 | log_fail("unable to determine running kernel version."); | |
185 | } elsif ($kernel_ver =~ /^$krunning/) { | |
186 | log_pass("expected running kernel '$kernel_ver'."); | |
187 | } elsif ($get_pkg->($kinstalled)) { | |
188 | log_warn("expected kernel '$kinstalled' intalled but not yet rebooted!"); | |
189 | } else { | |
190 | log_warn("unexpected running and installed kernel '$kernel_ver'."); | |
191 | } | |
192 | } else { | |
193 | log_fail("proxmox-ve package not found!"); | |
194 | } | |
195 | } | |
196 | ||
197 | ||
198 | sub check_storage_health { | |
199 | print_header("CHECKING CONFIGURED STORAGES"); | |
200 | my $cfg = PVE::Storage::config(); | |
201 | ||
202 | my $ctime = time(); | |
203 | ||
204 | my $info = PVE::Storage::storage_info($cfg); | |
205 | ||
206 | foreach my $storeid (keys %$info) { | |
207 | my $d = $info->{$storeid}; | |
208 | if ($d->{enabled}) { | |
209 | if ($d->{type} eq 'sheepdog') { | |
210 | log_fail("storage '$storeid' of type 'sheepdog' is enabled - experimental sheepdog support dropped in PVE 6") | |
211 | } elsif ($d->{active}) { | |
212 | log_pass("storage '$storeid' enabled and active."); | |
213 | } else { | |
214 | log_warn("storage '$storeid' enabled but not active!"); | |
215 | } | |
216 | } else { | |
217 | log_skip("storage '$storeid' disabled."); | |
218 | } | |
219 | } | |
220 | } | |
221 | ||
222 | sub check_cluster_corosync { | |
223 | print_header("CHECKING CLUSTER HEALTH/SETTINGS"); | |
224 | ||
225 | if (!PVE::Corosync::check_conf_exists(1)) { | |
226 | log_skip("standalone node."); | |
227 | return; | |
228 | } | |
229 | ||
230 | $log_systemd_unit_state->('pve-cluster.service'); | |
231 | $log_systemd_unit_state->('corosync.service'); | |
232 | ||
233 | if (PVE::Cluster::check_cfs_quorum(1)) { | |
234 | log_pass("Cluster Filesystem is quorate."); | |
235 | } else { | |
236 | log_fail("Cluster Filesystem readonly, lost quorum?!"); | |
237 | } | |
238 | ||
239 | my $conf = PVE::Cluster::cfs_read_file('corosync.conf'); | |
240 | my $conf_nodelist = PVE::Corosync::nodelist($conf); | |
241 | my $node_votes = 0; | |
242 | ||
243 | print "\nAnalzying quorum settings and state..\n"; | |
244 | if (!defined($conf_nodelist)) { | |
245 | log_fail("unable to retrieve nodelist from corosync.conf"); | |
246 | } else { | |
247 | if (grep { $conf_nodelist->{$_}->{quorum_votes} != 1 } keys %$conf_nodelist) { | |
248 | log_warn("non-default quorum_votes distribution detected!"); | |
249 | } | |
250 | map { $node_votes += $conf_nodelist->{$_}->{quorum_votes} // 0 } keys %$conf_nodelist; | |
251 | } | |
252 | ||
253 | my ($expected_votes, $total_votes); | |
254 | my $filter_output = sub { | |
255 | my $line = shift; | |
256 | ($expected_votes) = $line =~ /^Expected votes:\s*(\d+)\s*$/ | |
257 | if !defined($expected_votes); | |
258 | ($total_votes) = $line =~ /^Total votes:\s*(\d+)\s*$/ | |
259 | if !defined($total_votes); | |
260 | }; | |
261 | eval { | |
262 | run_command(['corosync-quorumtool', '-s'], outfunc => $filter_output, noerr => 1); | |
263 | }; | |
264 | ||
265 | if (!defined($expected_votes)) { | |
266 | log_fail("unable to get expected number of votes, setting to 0."); | |
267 | $expected_votes = 0; | |
268 | } | |
269 | if (!defined($total_votes)) { | |
270 | log_fail("unable to get expected number of votes, setting to 0."); | |
271 | $total_votes = 0; | |
272 | } | |
273 | ||
274 | my $cfs_nodelist = PVE::Cluster::get_clinfo()->{nodelist}; | |
275 | my $offline_nodes = grep { $cfs_nodelist->{$_}->{online} != 1 } keys %$cfs_nodelist; | |
276 | if ($offline_nodes > 0) { | |
277 | log_fail("$offline_nodes nodes are offline!"); | |
278 | } | |
279 | ||
280 | my $qdevice_votes = 0; | |
281 | if (my $qdevice_setup = $conf->{main}->{quorum}->{device}) { | |
282 | $qdevice_votes = $qdevice_setup->{votes} // 1; | |
283 | } | |
284 | ||
285 | log_info("configured votes - nodes: $node_votes"); | |
286 | log_info("configured votes - qdevice: $qdevice_votes"); | |
287 | log_info("current expected votes: $expected_votes"); | |
288 | log_info("current total votes: $total_votes"); | |
289 | ||
290 | log_warn("expected votes set to non-standard value '$expected_votes'.") | |
291 | if $expected_votes != $node_votes + $qdevice_votes; | |
292 | log_warn("total votes < expected votes: $total_votes/$expected_votes!") | |
293 | if $total_votes < $expected_votes; | |
294 | ||
295 | my $conf_nodelist_count = scalar(keys %$conf_nodelist); | |
296 | my $cfs_nodelist_count = scalar(keys %$cfs_nodelist); | |
297 | log_warn("cluster consists of less than three quorum-providing nodes!") | |
298 | if $conf_nodelist_count < 3 && $conf_nodelist_count + $qdevice_votes < 3; | |
299 | ||
300 | log_fail("corosync.conf ($conf_nodelist_count) and pmxcfs ($cfs_nodelist_count) don't agree about size of nodelist.") | |
301 | if $conf_nodelist_count != $cfs_nodelist_count; | |
302 | ||
303 | print "\nChecking nodelist entries..\n"; | |
304 | for my $cs_node (sort keys %$conf_nodelist) { | |
305 | my $entry = $conf_nodelist->{$cs_node}; | |
306 | log_fail("$cs_node: no name entry in corosync.conf.") | |
307 | if !defined($entry->{name}); | |
308 | log_fail("$cs_node: no nodeid configured in corosync.conf.") | |
309 | if !defined($entry->{nodeid}); | |
310 | my $gotLinks = 0; | |
311 | for my $link (0..7) { | |
312 | $gotLinks++ if defined($entry->{"ring${link}_addr"}); | |
313 | } | |
314 | log_fail("$cs_node: no ringX_addr (0 <= X <= 7) link defined in corosync.conf.") if $gotLinks <= 0; | |
315 | ||
316 | my $verify_ring_ip = sub { | |
317 | my $key = shift; | |
318 | if (defined(my $ring = $entry->{$key})) { | |
319 | my ($resolved_ip, undef) = PVE::Corosync::resolve_hostname_like_corosync($ring, $conf); | |
320 | if (defined($resolved_ip)) { | |
321 | if ($resolved_ip ne $ring) { | |
322 | log_warn("$cs_node: $key '$ring' resolves to '$resolved_ip'.\n Consider replacing it with the currently resolved IP address."); | |
323 | } else { | |
324 | log_pass("$cs_node: $key is configured to use IP address '$ring'"); | |
325 | } | |
326 | } else { | |
327 | 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!"); | |
328 | } | |
329 | } | |
330 | }; | |
331 | for my $link (0..7) { | |
332 | $verify_ring_ip->("ring${link}_addr"); | |
333 | } | |
334 | } | |
335 | ||
336 | print "\nChecking totem settings..\n"; | |
337 | my $totem = $conf->{main}->{totem}; | |
338 | my $transport = $totem->{transport}; | |
339 | if (defined($transport)) { | |
340 | if ($transport ne 'knet') { | |
341 | log_fail("Corosync transport explicitly set to '$transport' instead of implicit default!"); | |
342 | } else { | |
343 | log_pass("Corosync transport set to '$transport'."); | |
344 | } | |
345 | } else { | |
346 | log_pass("Corosync transport set to implicit default."); | |
347 | } | |
348 | ||
349 | # TODO: are those values still up-to-date? | |
350 | if ((!defined($totem->{secauth}) || $totem->{secauth} ne 'on') && (!defined($totem->{crypto_cipher}) || $totem->{crypto_cipher} eq 'none')) { | |
351 | log_fail("Corosync authentication/encryption is not explicitly enabled (secauth / crypto_cipher / crypto_hash)!"); | |
352 | } else { | |
353 | if (defined($totem->{crypto_cipher}) && $totem->{crypto_cipher} eq '3des') { | |
354 | log_fail("Corosync encryption cipher set to '3des', no longer supported in Corosync 3.x!"); # FIXME: can be removed? | |
355 | } else { | |
356 | log_pass("Corosync encryption and authentication enabled."); | |
357 | } | |
358 | } | |
359 | ||
360 | print "\n"; | |
361 | log_info("run 'pvecm status' to get detailed cluster status.."); | |
362 | ||
363 | print_header("CHECKING INSTALLED COROSYNC VERSION"); | |
364 | if (defined(my $corosync = $get_pkg->('corosync'))) { | |
365 | if ($corosync->{OldVersion} =~ m/^2\./) { | |
366 | log_fail("corosync 2.x installed, cluster-wide upgrade to 3.x needed!"); | |
367 | } elsif ($corosync->{OldVersion} =~ m/^3\./) { | |
368 | log_pass("corosync 3.x installed."); | |
369 | } else { | |
370 | log_fail("unexpected corosync version installed: $corosync->{OldVersion}!"); | |
371 | } | |
372 | } | |
373 | } | |
374 | ||
375 | sub check_ceph { | |
376 | print_header("CHECKING HYPER-CONVERGED CEPH STATUS"); | |
377 | ||
378 | if (PVE::Ceph::Tools::check_ceph_inited(1)) { | |
379 | log_info("hyper-converged ceph setup detected!"); | |
380 | } else { | |
381 | log_skip("no hyper-converged ceph setup detected!"); | |
382 | return; | |
383 | } | |
384 | ||
385 | log_info("getting Ceph status/health information.."); | |
386 | my $ceph_status = eval { PVE::API2::Ceph->status({ node => $nodename }); }; | |
387 | my $osd_flags = eval { PVE::API2::Ceph->get_flags({ node => $nodename }); }; | |
388 | my $noout_wanted = 1; | |
389 | my $noout = $osd_flags && $osd_flags =~ m/noout/; | |
390 | ||
391 | if (!$ceph_status || !$ceph_status->{health}) { | |
392 | log_fail("unable to determine Ceph status!"); | |
393 | } else { | |
394 | my $ceph_health = $ceph_status->{health}->{status}; | |
395 | if (!$ceph_health) { | |
396 | log_fail("unable to determine Ceph health!"); | |
397 | } elsif ($ceph_health eq 'HEALTH_OK') { | |
398 | log_pass("Ceph health reported as 'HEALTH_OK'."); | |
399 | } elsif ($ceph_health eq 'HEALTH_WARN' && $noout && (keys %{$ceph_status->{health}->{checks}} == 1)) { | |
400 | log_pass("Ceph health reported as 'HEALTH_WARN' with a single failing check and 'noout' flag set."); | |
401 | } else { | |
402 | log_warn("Ceph health reported as '$ceph_health'.\n Use the PVE ". | |
403 | "dashboard or 'ceph -s' to determine the specific issues and try to resolve them."); | |
404 | } | |
405 | } | |
406 | ||
407 | log_info("getting Ceph OSD flags.."); | |
408 | eval { | |
409 | if (!$osd_flags) { | |
410 | log_fail("unable to get Ceph OSD flags!"); | |
411 | } else { | |
412 | if ($osd_flags =~ m/recovery_deletes/ && $osd_flags =~ m/purged_snapdirs/) { | |
413 | log_pass("all PGs have been scrubbed at least once while running Ceph Luminous."); # FIXME: remove? | |
414 | } else { | |
415 | log_fail("missing 'recovery_deletes' and/or 'purged_snapdirs' flag, scrub of all PGs required before upgrading to Nautilus!"); | |
416 | } | |
417 | } | |
418 | }; | |
419 | ||
420 | # TODO: check OSD min-required version, if to low it breaks stuff! | |
421 | ||
422 | log_info("getting Ceph daemon versions.."); | |
423 | my $ceph_versions = eval { PVE::Ceph::Tools::get_cluster_versions(undef, 1); }; | |
424 | if (!$ceph_versions) { | |
425 | log_fail("unable to determine Ceph daemon versions!"); | |
426 | } else { | |
427 | my $services = [ | |
428 | { 'key' => 'mon', 'name' => 'monitor' }, | |
429 | { 'key' => 'mgr', 'name' => 'manager' }, | |
430 | { 'key' => 'mds', 'name' => 'MDS' }, | |
431 | { 'key' => 'osd', 'name' => 'OSD' }, | |
432 | ]; | |
433 | ||
434 | foreach my $service (@$services) { | |
435 | my $name = $service->{name}; | |
436 | if (my $service_versions = $ceph_versions->{$service->{key}}) { | |
437 | if (keys %$service_versions == 0) { | |
438 | log_skip("no running instances detected for daemon type $name."); | |
439 | } elsif (keys %$service_versions == 1) { | |
440 | log_pass("single running version detected for daemon type $name."); | |
441 | } else { | |
442 | log_warn("multiple running versions detected for daemon type $name!"); | |
443 | } | |
444 | } else { | |
445 | log_skip("unable to determine versions of running Ceph $name instances."); | |
446 | } | |
447 | } | |
448 | ||
449 | my $overall_versions = $ceph_versions->{overall}; | |
450 | if (!$overall_versions) { | |
451 | log_warn("unable to determine overall Ceph daemon versions!"); | |
452 | } elsif (keys %$overall_versions == 1) { | |
453 | log_pass("single running overall version detected for all Ceph daemon types."); | |
454 | if ((keys %$overall_versions)[0] =~ /^ceph version 15\./) { | |
455 | $noout_wanted = 0; | |
456 | } | |
457 | } else { | |
458 | log_warn("overall version mismatch detected, check 'ceph versions' output for details!"); | |
459 | } | |
460 | } | |
461 | ||
462 | if ($noout) { | |
463 | if ($noout_wanted) { | |
464 | log_pass("'noout' flag set to prevent rebalancing during cluster-wide upgrades."); | |
465 | } else { | |
466 | log_warn("'noout' flag set, Ceph cluster upgrade seems finished."); | |
467 | } | |
468 | } elsif ($noout_wanted) { | |
469 | log_warn("'noout' flag not set - recommended to prevent rebalancing during upgrades."); | |
470 | } | |
471 | ||
472 | log_info("checking Ceph config.."); | |
473 | my $conf = PVE::Cluster::cfs_read_file('ceph.conf'); | |
474 | if (%$conf) { | |
475 | my $global = $conf->{global}; | |
476 | ||
477 | my $global_monhost = $global->{mon_host} // $global->{"mon host"} // $global->{"mon-host"}; | |
478 | if (!defined($global_monhost)) { | |
479 | 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."); | |
480 | } else { | |
481 | log_pass("Found 'mon_host' entry."); | |
482 | } | |
483 | ||
484 | my $ipv6 = $global->{ms_bind_ipv6} // $global->{"ms bind ipv6"} // $global->{"ms-bind-ipv6"}; | |
485 | if ($ipv6) { | |
486 | my $ipv4 = $global->{ms_bind_ipv4} // $global->{"ms bind ipv4"} // $global->{"ms-bind-ipv4"}; | |
487 | if ($ipv6 eq 'true' && (!defined($ipv4) || $ipv4 ne 'false')) { | |
488 | log_warn("'ms_bind_ipv6' is enabled but 'ms_bind_ipv4' is not disabled.\n Make sure to disable 'ms_bind_ipv4' for ipv6 only clusters, or add an ipv4 network to public/cluster network."); | |
489 | } else { | |
490 | log_pass("'ms_bind_ipv6' is enabled and 'ms_bind_ipv4' disabled"); | |
491 | } | |
492 | } else { | |
493 | log_pass("'ms_bind_ipv6' not enabled"); | |
494 | } | |
495 | ||
496 | if (defined($global->{keyring})) { | |
497 | log_warn("[global] config section contains 'keyring' option, which will prevent services from starting with Nautilus.\n Move 'keyring' option to [client] section instead."); | |
498 | } else { | |
499 | log_pass("no 'keyring' option in [global] section found."); | |
500 | } | |
501 | ||
502 | } else { | |
503 | log_warn("Empty ceph config found"); | |
504 | } | |
505 | ||
506 | my $local_ceph_ver = PVE::Ceph::Tools::get_local_version(1); | |
507 | if (defined($local_ceph_ver)) { | |
508 | if ($local_ceph_ver == 14) { | |
509 | my $ceph_volume_osds = PVE::Ceph::Tools::ceph_volume_list(); | |
510 | my $scanned_osds = PVE::Tools::dir_glob_regex('/etc/ceph/osd', '^.*\.json$'); | |
511 | if (-e '/var/lib/ceph/osd/' && !defined($scanned_osds) && !(keys %$ceph_volume_osds)) { | |
512 | log_warn("local Ceph version is Nautilus, local OSDs detected, but no conversion from ceph-disk to ceph-volume done (yet)."); | |
513 | } | |
514 | } | |
515 | } else { | |
516 | log_fail("unable to determine local Ceph version."); | |
517 | } | |
518 | } | |
519 | ||
44090816 FE |
520 | sub check_backup_retention_settings { |
521 | log_info("Checking backup retention settings.."); | |
522 | ||
523 | my $pass = 1; | |
524 | ||
525 | my $node_has_retention; | |
526 | ||
527 | my $maxfiles_msg = "parameter 'maxfiles' is deprecated with PVE 7.x and will be removed in a " . | |
528 | "future version, use 'prune-backups' instead."; | |
529 | ||
530 | eval { | |
531 | my $confdesc = PVE::VZDump::Common::get_confdesc(); | |
532 | ||
533 | my $fn = "/etc/vzdump.conf"; | |
534 | my $raw = PVE::Tools::file_get_contents($fn); | |
535 | ||
536 | my $conf_schema = { type => 'object', properties => $confdesc, }; | |
537 | my $param = PVE::JSONSchema::parse_config($conf_schema, $fn, $raw); | |
538 | ||
539 | if (defined($param->{maxfiles})) { | |
540 | $pass = 0; | |
541 | log_warn("$fn - $maxfiles_msg"); | |
542 | } | |
543 | ||
544 | $node_has_retention = defined($param->{maxfiles}) || defined($param->{'prune-backups'}); | |
545 | }; | |
546 | if (my $err = $@) { | |
547 | $pass = 0; | |
548 | log_warn("unable to parse node's VZDump configuration - $err"); | |
549 | } | |
550 | ||
551 | my $storage_cfg = PVE::Storage::config(); | |
552 | ||
553 | for my $storeid (keys $storage_cfg->{ids}->%*) { | |
554 | my $scfg = $storage_cfg->{ids}->{$storeid}; | |
555 | ||
556 | if (defined($scfg->{maxfiles})) { | |
557 | $pass = 0; | |
558 | log_warn("storage '$storeid' - $maxfiles_msg"); | |
559 | } | |
560 | ||
561 | next if !$scfg->{content}->{backup}; | |
562 | next if defined($scfg->{maxfiles}) || defined($scfg->{'prune-backups'}); | |
563 | next if $node_has_retention; | |
564 | ||
565 | log_info("storage '$storeid' - no backup retention settings defined - by default, PVE " . | |
566 | "7.x will no longer keep only the last backup, but all backups"); | |
567 | } | |
568 | ||
35933450 FG |
569 | eval { |
570 | my $vzdump_cron = PVE::Cluster::cfs_read_file('vzdump.cron'); | |
44090816 | 571 | |
35933450 FG |
572 | # only warn once, there might be many jobs... |
573 | if (scalar(grep { defined($_->{maxfiles}) } $vzdump_cron->{jobs}->@*)) { | |
574 | $pass = 0; | |
575 | log_warn("/etc/pve/vzdump.cron - $maxfiles_msg"); | |
576 | } | |
577 | }; | |
578 | if (my $err = $@) { | |
44090816 | 579 | $pass = 0; |
35933450 | 580 | log_warn("unable to parse node's VZDump configuration - $err"); |
44090816 FE |
581 | } |
582 | ||
583 | log_pass("no problems found.") if $pass; | |
584 | } | |
585 | ||
0390b62e FE |
586 | sub check_cifs_credential_location { |
587 | log_info("checking CIFS credential location.."); | |
588 | ||
589 | my $regex = qr/^(.*)\.cred$/; | |
590 | ||
591 | my $found; | |
592 | ||
593 | PVE::Tools::dir_glob_foreach('/etc/pve/priv/', $regex, sub { | |
594 | my ($filename) = @_; | |
595 | ||
596 | my ($basename) = $filename =~ $regex; | |
597 | ||
598 | log_warn("CIFS credentials '/etc/pve/priv/$filename' will be moved to " . | |
599 | "'/etc/pve/priv/storage/$basename.pw' during the update"); | |
600 | ||
601 | $found = 1; | |
602 | }); | |
603 | ||
604 | log_pass("no CIFS credentials at outdated location found.") if !$found; | |
605 | } | |
606 | ||
31e4ad5d LS |
607 | sub check_custom_pool_roles { |
608 | log_info("Checking custom roles for pool permissions.."); | |
609 | ||
d817b524 FG |
610 | my $raw = eval { PVE::Tools::file_get_contents('/etc/pve/user.cfg'); }; |
611 | if ($@) { | |
612 | log_fail("Failed to read '/etc/pve/user.cfg' - $@"); | |
613 | return; | |
614 | } | |
31e4ad5d LS |
615 | |
616 | my $roles = {}; | |
b5f89880 FG |
617 | while ($raw =~ /^\s*(.+?)\s*$/gm) { |
618 | my $line = $1; | |
619 | my @data; | |
620 | ||
621 | foreach my $d (split (/:/, $line)) { | |
622 | $d =~ s/^\s+//; | |
623 | $d =~ s/\s+$//; | |
624 | push @data, $d | |
625 | } | |
626 | ||
627 | my $et = shift @data; | |
628 | next if $et ne 'role'; | |
31e4ad5d | 629 | |
b5f89880 | 630 | my ($role, $privlist) = @data; |
31e4ad5d LS |
631 | if (!PVE::AccessControl::verify_rolename($role, 1)) { |
632 | warn "user config - ignore role '$role' - invalid characters in role name\n"; | |
633 | next; | |
634 | } | |
635 | ||
636 | $roles->{$role} = {} if !$roles->{$role}; | |
637 | foreach my $priv (split_list($privlist)) { | |
638 | $roles->{$role}->{$priv} = 1; | |
639 | } | |
640 | } | |
641 | ||
642 | foreach my $role (sort keys %{$roles}) { | |
643 | if (PVE::AccessControl::role_is_special($role)) { | |
644 | next; | |
645 | } | |
646 | ||
647 | if ($role eq "PVEPoolUser") { | |
648 | # the user created a custom role named PVEPoolUser | |
649 | log_fail("Custom role '$role' has a restricted name - a built-in role 'PVEPoolUser' will be available with the upgrade"); | |
650 | } else { | |
651 | log_pass("Custom role '$role' has no restricted name"); | |
652 | } | |
653 | ||
654 | my $perms = $roles->{$role}; | |
655 | if ($perms->{'Pool.Allocate'} && $perms->{'Pool.Audit'}) { | |
656 | log_pass("Custom role '$role' contains updated pool permissions"); | |
657 | } elsif ($perms->{'Pool.Allocate'}) { | |
658 | log_warn("Custom role '$role' contains permission 'Pool.Allocate' - to ensure same behavior add 'Pool.Audit' to this role"); | |
659 | } else { | |
660 | log_pass("Custom role '$role' contains no permissions that need to be updated"); | |
661 | } | |
662 | } | |
663 | } | |
664 | ||
56142f8c TL |
665 | my sub check_max_length { |
666 | my ($raw, $max_length, $warning) = @_; | |
667 | log_warn($warning) if defined($raw) && length($raw) > $max_length; | |
668 | } | |
669 | ||
1be86a36 LS |
670 | sub check_description_lengths { |
671 | log_info("Checking node and guest description/note legnth.."); | |
672 | ||
56142f8c TL |
673 | my @affected_nodes = grep { |
674 | my $desc = PVE::NodeConfig::load_config($_)->{desc}; | |
675 | defined($desc) && length($desc) > 64 * 1024 | |
676 | } PVE::Cluster::get_nodelist(); | |
1be86a36 | 677 | |
56142f8c TL |
678 | if (scalar(@affected_nodes) > 0) { |
679 | log_warn("Node config description of the following nodes too long for new limit of 64 KiB:\n " | |
680 | . join(', ', @affected_nodes)); | |
681 | } else { | |
682 | log_pass("All node config descriptions fit in the new limit of 64 KiB"); | |
1be86a36 LS |
683 | } |
684 | ||
56142f8c TL |
685 | my $affected_guests = []; |
686 | ||
687 | my $cts = PVE::LXC::config_list(); | |
688 | for my $vmid (sort { $a <=> $b } keys %$cts) { | |
689 | my $desc = PVE::LXC::Config->load_config($vmid)->{description}; | |
690 | push @$affected_guests, "CT $vmid" if defined($desc) && length($desc) > 8 * 1024; | |
691 | } | |
1be86a36 | 692 | my $vms = PVE::QemuServer::config_list(); |
56142f8c TL |
693 | for my $vmid (sort { $a <=> $b } keys %$vms) { |
694 | my $desc = PVE::QemuConfig->load_config($vmid)->{description}; | |
695 | push @$affected_guests, "VM $vmid" if defined($desc) && length($desc) > 8 * 1024; | |
696 | } | |
697 | if (scalar($affected_guests->@*) > 0) { | |
698 | log_warn("Node config description of the following nodes too long for new limit of 64 KiB:\n" | |
699 | ." * " . join("\n * ", $affected_guests->@*)); | |
700 | } else { | |
701 | log_pass("All guest config descriptions fit in the new limit of 8 KiB"); | |
1be86a36 LS |
702 | } |
703 | } | |
704 | ||
09828717 FE |
705 | sub check_storage_content { |
706 | log_info("Scanning for guest images on storages without images/rootdir content type.."); | |
707 | ||
708 | my $found; | |
709 | ||
710 | my $storage_cfg = PVE::Storage::config(); | |
711 | ||
712 | for my $storeid (keys $storage_cfg->{ids}->%*) { | |
713 | my $scfg = $storage_cfg->{ids}->{$storeid}; | |
714 | ||
715 | next if !PVE::Storage::storage_check_enabled($storage_cfg, $storeid, undef, 1); | |
716 | ||
717 | next if $scfg->{content}->{images}; | |
718 | next if $scfg->{content}->{rootdir}; | |
719 | ||
720 | # Skip 'iscsi(direct)' (and foreign plugins with potentially similiar behavior) with 'none', | |
721 | # because that means "use LUNs directly" and vdisk_list() in PVE 6.x still lists those. | |
722 | # It's enough to *not* skip 'dir', because it is the only other storage that supports 'none' | |
723 | # and 'images' or 'rootdir', hence being potentially misconfigured. | |
724 | next if $scfg->{type} ne 'dir' && $scfg->{content}->{none}; | |
725 | ||
726 | my $res = PVE::Storage::vdisk_list($storage_cfg, $storeid); | |
727 | my $disk_list = $res->{$storeid}; | |
728 | ||
729 | my @volumes = map { $_->{volid} } $disk_list->@*; | |
730 | ||
731 | if (scalar(@volumes) > 0) { | |
732 | $found = 1; | |
733 | log_warn("storage '$storeid' - neither content type 'images' nor 'rootdir' " . | |
734 | "configured, but found guest volume(s) " . join(',', @volumes)); | |
735 | } | |
736 | } | |
737 | ||
738 | if ($found) { | |
739 | log_warn("PVE 7.0 enforces stricter content type checks. Guests referencing the above " . | |
740 | "volumes will not work until the storage configuration is fixed."); | |
741 | } else { | |
742 | log_pass("none found"); | |
743 | } | |
744 | } | |
745 | ||
4177a14d TL |
746 | sub check_misc { |
747 | print_header("MISCELLANEOUS CHECKS"); | |
748 | my $ssh_config = eval { PVE::Tools::file_get_contents('/root/.ssh/config') }; | |
749 | if (defined($ssh_config)) { | |
750 | log_fail("Unsupported SSH Cipher configured for root in /root/.ssh/config: $1") | |
751 | if $ssh_config =~ /^Ciphers .*(blowfish|arcfour|3des).*$/m; | |
752 | } else { | |
753 | log_skip("No SSH config file found."); | |
754 | } | |
755 | ||
756 | log_info("Checking common daemon services.."); | |
757 | $log_systemd_unit_state->('pveproxy.service'); | |
758 | $log_systemd_unit_state->('pvedaemon.service'); | |
759 | $log_systemd_unit_state->('pvestatd.service'); | |
760 | ||
761 | my $root_free = PVE::Tools::df('/', 10); | |
21aa5446 TL |
762 | log_warn("Less than 4 GiB free space on root file system.") |
763 | if defined($root_free) && $root_free->{avail} < 4*1024*1024*1024; | |
4177a14d TL |
764 | |
765 | log_info("Checking for running guests.."); | |
766 | my $running_guests = 0; | |
767 | ||
768 | my $vms = eval { PVE::API2::Qemu->vmlist({ node => $nodename }) }; | |
769 | log_warn("Failed to retrieve information about this node's VMs - $@") if $@; | |
770 | $running_guests += grep { $_->{status} eq 'running' } @$vms if defined($vms); | |
771 | ||
772 | my $cts = eval { PVE::API2::LXC->vmlist({ node => $nodename }) }; | |
773 | log_warn("Failed to retrieve information about this node's CTs - $@") if $@; | |
774 | $running_guests += grep { $_->{status} eq 'running' } @$cts if defined($cts); | |
775 | ||
776 | if ($running_guests > 0) { | |
777 | log_warn("$running_guests running guest(s) detected - consider migrating or stopping them.") | |
778 | } else { | |
779 | log_pass("no running guest detected.") | |
780 | } | |
781 | ||
782 | log_info("Checking if the local node's hostname '$nodename' is resolvable.."); | |
783 | my $local_ip = eval { PVE::Network::get_ip_from_hostname($nodename) }; | |
784 | if ($@) { | |
785 | log_warn("Failed to resolve hostname '$nodename' to IP - $@"); | |
786 | } else { | |
787 | log_info("Checking if resolved IP is configured on local node.."); | |
788 | my $cidr = Net::IP::ip_is_ipv6($local_ip) ? "$local_ip/128" : "$local_ip/32"; | |
789 | my $configured_ips = PVE::Network::get_local_ip_from_cidr($cidr); | |
790 | my $ip_count = scalar(@$configured_ips); | |
791 | ||
792 | if ($ip_count <= 0) { | |
793 | log_fail("Resolved node IP '$local_ip' not configured or active for '$nodename'"); | |
794 | } elsif ($ip_count > 1) { | |
795 | log_warn("Resolved node IP '$local_ip' active on multiple ($ip_count) interfaces!"); | |
796 | } else { | |
797 | log_pass("Resolved node IP '$local_ip' configured and active on single interface."); | |
798 | } | |
799 | } | |
800 | ||
801 | log_info("Check node certificate's RSA key size"); | |
802 | my $certs = PVE::API2::Certificates->info({ node => $nodename }); | |
803 | my $certs_check = { | |
804 | 'rsaEncryption' => { | |
805 | minsize => 2048, | |
806 | name => 'RSA', | |
807 | }, | |
808 | 'id-ecPublicKey' => { | |
809 | minsize => 224, | |
810 | name => 'ECC', | |
811 | }, | |
812 | }; | |
813 | ||
814 | my $certs_check_failed = 0; | |
815 | foreach my $cert (@$certs) { | |
816 | my ($type, $size, $fn) = $cert->@{qw(public-key-type public-key-bits filename)}; | |
817 | ||
818 | if (!defined($type) || !defined($size)) { | |
819 | log_warn("'$fn': cannot check certificate, failed to get it's type or size!"); | |
820 | } | |
821 | ||
822 | my $check = $certs_check->{$type}; | |
823 | if (!defined($check)) { | |
824 | log_warn("'$fn': certificate's public key type '$type' unknown, check Debian Busters release notes"); | |
825 | next; | |
826 | } | |
827 | ||
828 | if ($size < $check->{minsize}) { | |
829 | log_fail("'$fn', certificate's $check->{name} public key size is less than 2048 bit"); | |
830 | $certs_check_failed = 1; | |
831 | } else { | |
832 | log_pass("Certificate '$fn' passed Debian Busters security level for TLS connections ($size >= 2048)"); | |
833 | } | |
834 | } | |
44090816 FE |
835 | |
836 | check_backup_retention_settings(); | |
0390b62e | 837 | check_cifs_credential_location(); |
31e4ad5d | 838 | check_custom_pool_roles(); |
1be86a36 | 839 | check_description_lengths(); |
09828717 | 840 | check_storage_content(); |
4177a14d TL |
841 | } |
842 | ||
843 | __PACKAGE__->register_method ({ | |
844 | name => 'checklist', | |
845 | path => 'checklist', | |
846 | method => 'GET', | |
847 | description => 'Check (pre-/post-)upgrade conditions.', | |
848 | parameters => { | |
849 | additionalProperties => 0, | |
850 | properties => { | |
851 | }, | |
852 | }, | |
853 | returns => { type => 'null' }, | |
854 | code => sub { | |
855 | my ($param) = @_; | |
856 | ||
857 | check_pve_packages(); | |
858 | check_cluster_corosync(); | |
859 | check_ceph(); | |
860 | check_storage_health(); | |
861 | check_misc(); | |
862 | ||
863 | print_header("SUMMARY"); | |
864 | ||
865 | my $total = 0; | |
866 | $total += $_ for values %$counters; | |
867 | ||
868 | print "TOTAL: $total\n"; | |
869 | print colored("PASSED: $counters->{pass}\n", 'green'); | |
870 | print "SKIPPED: $counters->{skip}\n"; | |
871 | print colored("WARNINGS: $counters->{warn}\n", 'yellow'); | |
872 | print colored("FAILURES: $counters->{fail}\n", 'red'); | |
873 | ||
874 | if ($counters->{warn} > 0 || $counters->{fail} > 0) { | |
875 | my $color = $counters->{fail} > 0 ? 'red' : 'yellow'; | |
876 | print colored("\nATTENTION: Please check the output for detailed information!\n", $color); | |
877 | print colored("Try to solve the problems one at a time and then run this checklist tool again.\n", $color) if $counters->{fail} > 0; | |
878 | } | |
879 | ||
880 | return undef; | |
881 | }}); | |
882 | ||
883 | our $cmddef = [ __PACKAGE__, 'checklist', [], {}]; | |
884 | ||
885 | # for now drop all unknown params and just check | |
886 | @ARGV = (); | |
887 | ||
888 | 1; |