]>
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 | ||
5b3bad37 | 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; | |
18 | use PVE::RPCEnvironment; | |
19 | use PVE::Storage; | |
5b3bad37 | 20 | use PVE::Tools qw(run_command split_list); |
4177a14d | 21 | use PVE::QemuServer; |
e522a2db | 22 | use PVE::VZDump::Common; |
4177a14d TL |
23 | |
24 | use Term::ANSIColor; | |
25 | ||
26 | use PVE::CLIHandler; | |
27 | ||
28 | use base qw(PVE::CLIHandler); | |
29 | ||
30 | my $nodename = PVE::INotify::nodename(); | |
31 | ||
32 | sub setup_environment { | |
33 | PVE::RPCEnvironment->setup_default_cli_env(); | |
34 | } | |
35 | ||
36 | my $min_pve_major = 6; | |
37 | my $min_pve_minor = 4; | |
38 | my $min_pve_pkgrel = 1; | |
39 | ||
40 | my $counters = { | |
41 | pass => 0, | |
42 | skip => 0, | |
43 | warn => 0, | |
44 | fail => 0, | |
45 | }; | |
46 | ||
47 | my $log_line = sub { | |
48 | my ($level, $line) = @_; | |
49 | ||
50 | $counters->{$level}++ if defined($level) && defined($counters->{$level}); | |
51 | ||
52 | print uc($level), ': ' if defined($level); | |
53 | print "$line\n"; | |
54 | }; | |
55 | ||
56 | sub log_pass { | |
57 | print color('green'); | |
58 | $log_line->('pass', @_); | |
59 | print color('reset'); | |
60 | } | |
61 | ||
62 | sub log_info { | |
63 | $log_line->('info', @_); | |
64 | } | |
65 | sub log_skip { | |
66 | $log_line->('skip', @_); | |
67 | } | |
68 | sub log_warn { | |
69 | print color('yellow'); | |
70 | $log_line->('warn', @_); | |
71 | print color('reset'); | |
72 | } | |
73 | sub log_fail { | |
74 | print color('red'); | |
75 | $log_line->('fail', @_); | |
76 | print color('reset'); | |
77 | } | |
78 | ||
79 | my $print_header_first = 1; | |
80 | sub print_header { | |
81 | my ($h) = @_; | |
82 | print "\n" if !$print_header_first; | |
83 | print "= $h =\n\n"; | |
84 | $print_header_first = 0; | |
85 | } | |
86 | ||
87 | my $get_systemd_unit_state = sub { | |
88 | my ($unit) = @_; | |
89 | ||
90 | my $state; | |
91 | my $filter_output = sub { | |
92 | $state = shift; | |
93 | chomp $state; | |
94 | }; | |
95 | eval { | |
96 | run_command(['systemctl', 'is-enabled', "$unit"], outfunc => $filter_output, noerr => 1); | |
97 | return if !defined($state); | |
98 | run_command(['systemctl', 'is-active', "$unit"], outfunc => $filter_output, noerr => 1); | |
99 | }; | |
100 | ||
101 | return $state // 'unknown'; | |
102 | }; | |
103 | my $log_systemd_unit_state = sub { | |
104 | my ($unit, $no_fail_on_inactive) = @_; | |
105 | ||
106 | my $log_method = \&log_warn; | |
107 | ||
108 | my $state = $get_systemd_unit_state->($unit); | |
109 | if ($state eq 'active') { | |
110 | $log_method = \&log_pass; | |
111 | } elsif ($state eq 'inactive') { | |
112 | $log_method = $no_fail_on_inactive ? \&log_warn : \&log_fail; | |
113 | } elsif ($state eq 'failed') { | |
114 | $log_method = \&log_fail; | |
115 | } | |
116 | ||
117 | $log_method->("systemd unit '$unit' is in state '$state'"); | |
118 | }; | |
119 | ||
120 | my $versions; | |
121 | my $get_pkg = sub { | |
122 | my ($pkg) = @_; | |
123 | ||
124 | $versions = eval { PVE::API2::APT->versions({ node => $nodename }) } if !defined($versions); | |
125 | ||
126 | if (!defined($versions)) { | |
127 | my $msg = "unable to retrieve package version information"; | |
128 | $msg .= "- $@" if $@; | |
129 | log_fail("$msg"); | |
130 | return undef; | |
131 | } | |
132 | ||
133 | my $pkgs = [ grep { $_->{Package} eq $pkg } @$versions ]; | |
134 | if (!defined $pkgs || $pkgs == 0) { | |
135 | log_fail("unable to determine installed $pkg version."); | |
136 | return undef; | |
137 | } else { | |
138 | return $pkgs->[0]; | |
139 | } | |
140 | }; | |
141 | ||
142 | sub check_pve_packages { | |
143 | print_header("CHECKING VERSION INFORMATION FOR PVE PACKAGES"); | |
144 | ||
145 | print "Checking for package updates..\n"; | |
146 | my $updates = eval { PVE::API2::APT->list_updates({ node => $nodename }); }; | |
147 | if (!defined($updates)) { | |
148 | log_warn("$@") if $@; | |
149 | log_fail("unable to retrieve list of package updates!"); | |
150 | } elsif (@$updates > 0) { | |
151 | my $pkgs = join(', ', map { $_->{Package} } @$updates); | |
152 | log_warn("updates for the following packages are available:\n $pkgs"); | |
153 | } else { | |
154 | log_pass("all packages uptodate"); | |
155 | } | |
156 | ||
157 | print "\nChecking proxmox-ve package version..\n"; | |
158 | if (defined(my $proxmox_ve = $get_pkg->('proxmox-ve'))) { | |
159 | my $min_pve_ver = "$min_pve_major.$min_pve_minor-$min_pve_pkgrel"; | |
160 | ||
161 | my ($maj, $min, $pkgrel) = $proxmox_ve->{OldVersion} =~ m/^(\d+)\.(\d+)-(\d+)/; | |
162 | ||
163 | my $upgraded = 0; | |
164 | ||
165 | if ($maj > $min_pve_major) { | |
166 | log_pass("already upgraded to Proxmox VE " . ($min_pve_major + 1)); | |
167 | $upgraded = 1; | |
168 | } elsif ($maj >= $min_pve_major && $min >= $min_pve_minor && $pkgrel >= $min_pve_pkgrel) { | |
169 | log_pass("proxmox-ve package has version >= $min_pve_ver"); | |
170 | } else { | |
171 | log_fail("proxmox-ve package is too old, please upgrade to >= $min_pve_ver!"); | |
172 | } | |
173 | ||
174 | my ($krunning, $kinstalled) = (qr/5\.11/, 'pve-kernel-5.11'); | |
175 | if (!$upgraded) { | |
176 | ($krunning, $kinstalled) = (qr/5\.(?:4|11)/, 'pve-kernel-4.15'); | |
177 | } | |
178 | ||
179 | print "\nChecking running kernel version..\n"; | |
180 | my $kernel_ver = $proxmox_ve->{RunningKernel}; | |
181 | if (!defined($kernel_ver)) { | |
182 | log_fail("unable to determine running kernel version."); | |
183 | } elsif ($kernel_ver =~ /^$krunning/) { | |
184 | log_pass("expected running kernel '$kernel_ver'."); | |
185 | } elsif ($get_pkg->($kinstalled)) { | |
186 | log_warn("expected kernel '$kinstalled' intalled but not yet rebooted!"); | |
187 | } else { | |
188 | log_warn("unexpected running and installed kernel '$kernel_ver'."); | |
189 | } | |
190 | } else { | |
191 | log_fail("proxmox-ve package not found!"); | |
192 | } | |
193 | } | |
194 | ||
195 | ||
196 | sub check_storage_health { | |
197 | print_header("CHECKING CONFIGURED STORAGES"); | |
198 | my $cfg = PVE::Storage::config(); | |
199 | ||
200 | my $ctime = time(); | |
201 | ||
202 | my $info = PVE::Storage::storage_info($cfg); | |
203 | ||
204 | foreach my $storeid (keys %$info) { | |
205 | my $d = $info->{$storeid}; | |
206 | if ($d->{enabled}) { | |
207 | if ($d->{type} eq 'sheepdog') { | |
208 | log_fail("storage '$storeid' of type 'sheepdog' is enabled - experimental sheepdog support dropped in PVE 6") | |
209 | } elsif ($d->{active}) { | |
210 | log_pass("storage '$storeid' enabled and active."); | |
211 | } else { | |
212 | log_warn("storage '$storeid' enabled but not active!"); | |
213 | } | |
214 | } else { | |
215 | log_skip("storage '$storeid' disabled."); | |
216 | } | |
217 | } | |
218 | } | |
219 | ||
220 | sub check_cluster_corosync { | |
221 | print_header("CHECKING CLUSTER HEALTH/SETTINGS"); | |
222 | ||
223 | if (!PVE::Corosync::check_conf_exists(1)) { | |
224 | log_skip("standalone node."); | |
225 | return; | |
226 | } | |
227 | ||
228 | $log_systemd_unit_state->('pve-cluster.service'); | |
229 | $log_systemd_unit_state->('corosync.service'); | |
230 | ||
231 | if (PVE::Cluster::check_cfs_quorum(1)) { | |
232 | log_pass("Cluster Filesystem is quorate."); | |
233 | } else { | |
234 | log_fail("Cluster Filesystem readonly, lost quorum?!"); | |
235 | } | |
236 | ||
237 | my $conf = PVE::Cluster::cfs_read_file('corosync.conf'); | |
238 | my $conf_nodelist = PVE::Corosync::nodelist($conf); | |
239 | my $node_votes = 0; | |
240 | ||
241 | print "\nAnalzying quorum settings and state..\n"; | |
242 | if (!defined($conf_nodelist)) { | |
243 | log_fail("unable to retrieve nodelist from corosync.conf"); | |
244 | } else { | |
245 | if (grep { $conf_nodelist->{$_}->{quorum_votes} != 1 } keys %$conf_nodelist) { | |
246 | log_warn("non-default quorum_votes distribution detected!"); | |
247 | } | |
248 | map { $node_votes += $conf_nodelist->{$_}->{quorum_votes} // 0 } keys %$conf_nodelist; | |
249 | } | |
250 | ||
251 | my ($expected_votes, $total_votes); | |
252 | my $filter_output = sub { | |
253 | my $line = shift; | |
254 | ($expected_votes) = $line =~ /^Expected votes:\s*(\d+)\s*$/ | |
255 | if !defined($expected_votes); | |
256 | ($total_votes) = $line =~ /^Total votes:\s*(\d+)\s*$/ | |
257 | if !defined($total_votes); | |
258 | }; | |
259 | eval { | |
260 | run_command(['corosync-quorumtool', '-s'], outfunc => $filter_output, noerr => 1); | |
261 | }; | |
262 | ||
263 | if (!defined($expected_votes)) { | |
264 | log_fail("unable to get expected number of votes, setting to 0."); | |
265 | $expected_votes = 0; | |
266 | } | |
267 | if (!defined($total_votes)) { | |
268 | log_fail("unable to get expected number of votes, setting to 0."); | |
269 | $total_votes = 0; | |
270 | } | |
271 | ||
272 | my $cfs_nodelist = PVE::Cluster::get_clinfo()->{nodelist}; | |
273 | my $offline_nodes = grep { $cfs_nodelist->{$_}->{online} != 1 } keys %$cfs_nodelist; | |
274 | if ($offline_nodes > 0) { | |
275 | log_fail("$offline_nodes nodes are offline!"); | |
276 | } | |
277 | ||
278 | my $qdevice_votes = 0; | |
279 | if (my $qdevice_setup = $conf->{main}->{quorum}->{device}) { | |
280 | $qdevice_votes = $qdevice_setup->{votes} // 1; | |
281 | } | |
282 | ||
283 | log_info("configured votes - nodes: $node_votes"); | |
284 | log_info("configured votes - qdevice: $qdevice_votes"); | |
285 | log_info("current expected votes: $expected_votes"); | |
286 | log_info("current total votes: $total_votes"); | |
287 | ||
288 | log_warn("expected votes set to non-standard value '$expected_votes'.") | |
289 | if $expected_votes != $node_votes + $qdevice_votes; | |
290 | log_warn("total votes < expected votes: $total_votes/$expected_votes!") | |
291 | if $total_votes < $expected_votes; | |
292 | ||
293 | my $conf_nodelist_count = scalar(keys %$conf_nodelist); | |
294 | my $cfs_nodelist_count = scalar(keys %$cfs_nodelist); | |
295 | log_warn("cluster consists of less than three quorum-providing nodes!") | |
296 | if $conf_nodelist_count < 3 && $conf_nodelist_count + $qdevice_votes < 3; | |
297 | ||
298 | log_fail("corosync.conf ($conf_nodelist_count) and pmxcfs ($cfs_nodelist_count) don't agree about size of nodelist.") | |
299 | if $conf_nodelist_count != $cfs_nodelist_count; | |
300 | ||
301 | print "\nChecking nodelist entries..\n"; | |
302 | for my $cs_node (sort keys %$conf_nodelist) { | |
303 | my $entry = $conf_nodelist->{$cs_node}; | |
304 | log_fail("$cs_node: no name entry in corosync.conf.") | |
305 | if !defined($entry->{name}); | |
306 | log_fail("$cs_node: no nodeid configured in corosync.conf.") | |
307 | if !defined($entry->{nodeid}); | |
308 | my $gotLinks = 0; | |
309 | for my $link (0..7) { | |
310 | $gotLinks++ if defined($entry->{"ring${link}_addr"}); | |
311 | } | |
312 | log_fail("$cs_node: no ringX_addr (0 <= X <= 7) link defined in corosync.conf.") if $gotLinks <= 0; | |
313 | ||
314 | my $verify_ring_ip = sub { | |
315 | my $key = shift; | |
316 | if (defined(my $ring = $entry->{$key})) { | |
317 | my ($resolved_ip, undef) = PVE::Corosync::resolve_hostname_like_corosync($ring, $conf); | |
318 | if (defined($resolved_ip)) { | |
319 | if ($resolved_ip ne $ring) { | |
320 | log_warn("$cs_node: $key '$ring' resolves to '$resolved_ip'.\n Consider replacing it with the currently resolved IP address."); | |
321 | } else { | |
322 | log_pass("$cs_node: $key is configured to use IP address '$ring'"); | |
323 | } | |
324 | } else { | |
325 | 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!"); | |
326 | } | |
327 | } | |
328 | }; | |
329 | for my $link (0..7) { | |
330 | $verify_ring_ip->("ring${link}_addr"); | |
331 | } | |
332 | } | |
333 | ||
334 | print "\nChecking totem settings..\n"; | |
335 | my $totem = $conf->{main}->{totem}; | |
336 | my $transport = $totem->{transport}; | |
337 | if (defined($transport)) { | |
338 | if ($transport ne 'knet') { | |
339 | log_fail("Corosync transport explicitly set to '$transport' instead of implicit default!"); | |
340 | } else { | |
341 | log_pass("Corosync transport set to '$transport'."); | |
342 | } | |
343 | } else { | |
344 | log_pass("Corosync transport set to implicit default."); | |
345 | } | |
346 | ||
347 | # TODO: are those values still up-to-date? | |
348 | if ((!defined($totem->{secauth}) || $totem->{secauth} ne 'on') && (!defined($totem->{crypto_cipher}) || $totem->{crypto_cipher} eq 'none')) { | |
349 | log_fail("Corosync authentication/encryption is not explicitly enabled (secauth / crypto_cipher / crypto_hash)!"); | |
350 | } else { | |
351 | if (defined($totem->{crypto_cipher}) && $totem->{crypto_cipher} eq '3des') { | |
352 | log_fail("Corosync encryption cipher set to '3des', no longer supported in Corosync 3.x!"); # FIXME: can be removed? | |
353 | } else { | |
354 | log_pass("Corosync encryption and authentication enabled."); | |
355 | } | |
356 | } | |
357 | ||
358 | print "\n"; | |
359 | log_info("run 'pvecm status' to get detailed cluster status.."); | |
360 | ||
361 | print_header("CHECKING INSTALLED COROSYNC VERSION"); | |
362 | if (defined(my $corosync = $get_pkg->('corosync'))) { | |
363 | if ($corosync->{OldVersion} =~ m/^2\./) { | |
364 | log_fail("corosync 2.x installed, cluster-wide upgrade to 3.x needed!"); | |
365 | } elsif ($corosync->{OldVersion} =~ m/^3\./) { | |
366 | log_pass("corosync 3.x installed."); | |
367 | } else { | |
368 | log_fail("unexpected corosync version installed: $corosync->{OldVersion}!"); | |
369 | } | |
370 | } | |
371 | } | |
372 | ||
373 | sub check_ceph { | |
374 | print_header("CHECKING HYPER-CONVERGED CEPH STATUS"); | |
375 | ||
376 | if (PVE::Ceph::Tools::check_ceph_inited(1)) { | |
377 | log_info("hyper-converged ceph setup detected!"); | |
378 | } else { | |
379 | log_skip("no hyper-converged ceph setup detected!"); | |
380 | return; | |
381 | } | |
382 | ||
383 | log_info("getting Ceph status/health information.."); | |
384 | my $ceph_status = eval { PVE::API2::Ceph->status({ node => $nodename }); }; | |
385 | my $osd_flags = eval { PVE::API2::Ceph->get_flags({ node => $nodename }); }; | |
386 | my $noout_wanted = 1; | |
387 | my $noout = $osd_flags && $osd_flags =~ m/noout/; | |
388 | ||
389 | if (!$ceph_status || !$ceph_status->{health}) { | |
390 | log_fail("unable to determine Ceph status!"); | |
391 | } else { | |
392 | my $ceph_health = $ceph_status->{health}->{status}; | |
393 | if (!$ceph_health) { | |
394 | log_fail("unable to determine Ceph health!"); | |
395 | } elsif ($ceph_health eq 'HEALTH_OK') { | |
396 | log_pass("Ceph health reported as 'HEALTH_OK'."); | |
397 | } elsif ($ceph_health eq 'HEALTH_WARN' && $noout && (keys %{$ceph_status->{health}->{checks}} == 1)) { | |
398 | log_pass("Ceph health reported as 'HEALTH_WARN' with a single failing check and 'noout' flag set."); | |
399 | } else { | |
400 | log_warn("Ceph health reported as '$ceph_health'.\n Use the PVE ". | |
401 | "dashboard or 'ceph -s' to determine the specific issues and try to resolve them."); | |
402 | } | |
403 | } | |
404 | ||
405 | log_info("getting Ceph OSD flags.."); | |
406 | eval { | |
407 | if (!$osd_flags) { | |
408 | log_fail("unable to get Ceph OSD flags!"); | |
409 | } else { | |
410 | if ($osd_flags =~ m/recovery_deletes/ && $osd_flags =~ m/purged_snapdirs/) { | |
411 | log_pass("all PGs have been scrubbed at least once while running Ceph Luminous."); # FIXME: remove? | |
412 | } else { | |
413 | log_fail("missing 'recovery_deletes' and/or 'purged_snapdirs' flag, scrub of all PGs required before upgrading to Nautilus!"); | |
414 | } | |
415 | } | |
416 | }; | |
417 | ||
418 | # TODO: check OSD min-required version, if to low it breaks stuff! | |
419 | ||
420 | log_info("getting Ceph daemon versions.."); | |
421 | my $ceph_versions = eval { PVE::Ceph::Tools::get_cluster_versions(undef, 1); }; | |
422 | if (!$ceph_versions) { | |
423 | log_fail("unable to determine Ceph daemon versions!"); | |
424 | } else { | |
425 | my $services = [ | |
426 | { 'key' => 'mon', 'name' => 'monitor' }, | |
427 | { 'key' => 'mgr', 'name' => 'manager' }, | |
428 | { 'key' => 'mds', 'name' => 'MDS' }, | |
429 | { 'key' => 'osd', 'name' => 'OSD' }, | |
430 | ]; | |
431 | ||
432 | foreach my $service (@$services) { | |
433 | my $name = $service->{name}; | |
434 | if (my $service_versions = $ceph_versions->{$service->{key}}) { | |
435 | if (keys %$service_versions == 0) { | |
436 | log_skip("no running instances detected for daemon type $name."); | |
437 | } elsif (keys %$service_versions == 1) { | |
438 | log_pass("single running version detected for daemon type $name."); | |
439 | } else { | |
440 | log_warn("multiple running versions detected for daemon type $name!"); | |
441 | } | |
442 | } else { | |
443 | log_skip("unable to determine versions of running Ceph $name instances."); | |
444 | } | |
445 | } | |
446 | ||
447 | my $overall_versions = $ceph_versions->{overall}; | |
448 | if (!$overall_versions) { | |
449 | log_warn("unable to determine overall Ceph daemon versions!"); | |
450 | } elsif (keys %$overall_versions == 1) { | |
451 | log_pass("single running overall version detected for all Ceph daemon types."); | |
452 | if ((keys %$overall_versions)[0] =~ /^ceph version 15\./) { | |
453 | $noout_wanted = 0; | |
454 | } | |
455 | } else { | |
456 | log_warn("overall version mismatch detected, check 'ceph versions' output for details!"); | |
457 | } | |
458 | } | |
459 | ||
460 | if ($noout) { | |
461 | if ($noout_wanted) { | |
462 | log_pass("'noout' flag set to prevent rebalancing during cluster-wide upgrades."); | |
463 | } else { | |
464 | log_warn("'noout' flag set, Ceph cluster upgrade seems finished."); | |
465 | } | |
466 | } elsif ($noout_wanted) { | |
467 | log_warn("'noout' flag not set - recommended to prevent rebalancing during upgrades."); | |
468 | } | |
469 | ||
470 | log_info("checking Ceph config.."); | |
471 | my $conf = PVE::Cluster::cfs_read_file('ceph.conf'); | |
472 | if (%$conf) { | |
473 | my $global = $conf->{global}; | |
474 | ||
475 | my $global_monhost = $global->{mon_host} // $global->{"mon host"} // $global->{"mon-host"}; | |
476 | if (!defined($global_monhost)) { | |
477 | 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."); | |
478 | } else { | |
479 | log_pass("Found 'mon_host' entry."); | |
480 | } | |
481 | ||
482 | my $ipv6 = $global->{ms_bind_ipv6} // $global->{"ms bind ipv6"} // $global->{"ms-bind-ipv6"}; | |
483 | if ($ipv6) { | |
484 | my $ipv4 = $global->{ms_bind_ipv4} // $global->{"ms bind ipv4"} // $global->{"ms-bind-ipv4"}; | |
485 | if ($ipv6 eq 'true' && (!defined($ipv4) || $ipv4 ne 'false')) { | |
486 | log_warn("'ms_bind_ipv6' is enabled but 'ms_bind_ipv4' is not disabled.\n Make sure to disable 'ms_bind_ipv4' for ipv6 only clusters, or add an ipv4 network to public/cluster network."); | |
487 | } else { | |
488 | log_pass("'ms_bind_ipv6' is enabled and 'ms_bind_ipv4' disabled"); | |
489 | } | |
490 | } else { | |
491 | log_pass("'ms_bind_ipv6' not enabled"); | |
492 | } | |
493 | ||
494 | if (defined($global->{keyring})) { | |
495 | log_warn("[global] config section contains 'keyring' option, which will prevent services from starting with Nautilus.\n Move 'keyring' option to [client] section instead."); | |
496 | } else { | |
497 | log_pass("no 'keyring' option in [global] section found."); | |
498 | } | |
499 | ||
500 | } else { | |
501 | log_warn("Empty ceph config found"); | |
502 | } | |
503 | ||
504 | my $local_ceph_ver = PVE::Ceph::Tools::get_local_version(1); | |
505 | if (defined($local_ceph_ver)) { | |
506 | if ($local_ceph_ver == 14) { | |
507 | my $ceph_volume_osds = PVE::Ceph::Tools::ceph_volume_list(); | |
508 | my $scanned_osds = PVE::Tools::dir_glob_regex('/etc/ceph/osd', '^.*\.json$'); | |
509 | if (-e '/var/lib/ceph/osd/' && !defined($scanned_osds) && !(keys %$ceph_volume_osds)) { | |
510 | log_warn("local Ceph version is Nautilus, local OSDs detected, but no conversion from ceph-disk to ceph-volume done (yet)."); | |
511 | } | |
512 | } | |
513 | } else { | |
514 | log_fail("unable to determine local Ceph version."); | |
515 | } | |
516 | } | |
517 | ||
e522a2db FE |
518 | sub check_backup_retention_settings { |
519 | log_info("Checking backup retention settings.."); | |
520 | ||
521 | my $pass = 1; | |
522 | ||
523 | my $node_has_retention; | |
524 | ||
525 | my $maxfiles_msg = "parameter 'maxfiles' is deprecated with PVE 7.x and will be removed in a " . | |
526 | "future version, use 'prune-backups' instead."; | |
527 | ||
528 | eval { | |
529 | my $confdesc = PVE::VZDump::Common::get_confdesc(); | |
530 | ||
531 | my $fn = "/etc/vzdump.conf"; | |
532 | my $raw = PVE::Tools::file_get_contents($fn); | |
533 | ||
534 | my $conf_schema = { type => 'object', properties => $confdesc, }; | |
535 | my $param = PVE::JSONSchema::parse_config($conf_schema, $fn, $raw); | |
536 | ||
537 | if (defined($param->{maxfiles})) { | |
538 | $pass = 0; | |
539 | log_warn("$fn - $maxfiles_msg"); | |
540 | } | |
541 | ||
542 | $node_has_retention = defined($param->{maxfiles}) || defined($param->{'prune-backups'}); | |
543 | }; | |
544 | if (my $err = $@) { | |
545 | $pass = 0; | |
546 | log_warn("unable to parse node's VZDump configuration - $err"); | |
547 | } | |
548 | ||
549 | my $storage_cfg = PVE::Storage::config(); | |
550 | ||
551 | for my $storeid (keys $storage_cfg->{ids}->%*) { | |
552 | my $scfg = $storage_cfg->{ids}->{$storeid}; | |
553 | ||
554 | if (defined($scfg->{maxfiles})) { | |
555 | $pass = 0; | |
556 | log_warn("storage '$storeid' - $maxfiles_msg"); | |
557 | } | |
558 | ||
559 | next if !$scfg->{content}->{backup}; | |
560 | next if defined($scfg->{maxfiles}) || defined($scfg->{'prune-backups'}); | |
561 | next if $node_has_retention; | |
562 | ||
563 | log_info("storage '$storeid' - no backup retention settings defined - by default, PVE " . | |
564 | "7.x will no longer keep only the last backup, but all backups"); | |
565 | } | |
566 | ||
437ade60 FG |
567 | eval { |
568 | my $vzdump_cron = PVE::Cluster::cfs_read_file('vzdump.cron'); | |
e522a2db | 569 | |
437ade60 FG |
570 | # only warn once, there might be many jobs... |
571 | if (scalar(grep { defined($_->{maxfiles}) } $vzdump_cron->{jobs}->@*)) { | |
572 | $pass = 0; | |
573 | log_warn("/etc/pve/vzdump.cron - $maxfiles_msg"); | |
574 | } | |
575 | }; | |
576 | if (my $err = $@) { | |
e522a2db | 577 | $pass = 0; |
437ade60 | 578 | log_warn("unable to parse node's VZDump configuration - $err"); |
e522a2db FE |
579 | } |
580 | ||
581 | log_pass("no problems found.") if $pass; | |
582 | } | |
583 | ||
5df8b555 FE |
584 | sub check_cifs_credential_location { |
585 | log_info("checking CIFS credential location.."); | |
586 | ||
587 | my $regex = qr/^(.*)\.cred$/; | |
588 | ||
589 | my $found; | |
590 | ||
591 | PVE::Tools::dir_glob_foreach('/etc/pve/priv/', $regex, sub { | |
592 | my ($filename) = @_; | |
593 | ||
594 | my ($basename) = $filename =~ $regex; | |
595 | ||
596 | log_warn("CIFS credentials '/etc/pve/priv/$filename' will be moved to " . | |
597 | "'/etc/pve/priv/storage/$basename.pw' during the update"); | |
598 | ||
599 | $found = 1; | |
600 | }); | |
601 | ||
602 | log_pass("no CIFS credentials at outdated location found.") if !$found; | |
603 | } | |
604 | ||
5b3bad37 LS |
605 | sub check_custom_pool_roles { |
606 | log_info("Checking custom roles for pool permissions.."); | |
607 | ||
055b6967 FG |
608 | my $raw = eval { PVE::Tools::file_get_contents('/etc/pve/user.cfg'); }; |
609 | if ($@) { | |
610 | log_fail("Failed to read '/etc/pve/user.cfg' - $@"); | |
611 | return; | |
612 | } | |
5b3bad37 LS |
613 | |
614 | my $roles = {}; | |
65b46cad FG |
615 | while ($raw =~ /^\s*(.+?)\s*$/gm) { |
616 | my $line = $1; | |
617 | my @data; | |
618 | ||
619 | foreach my $d (split (/:/, $line)) { | |
620 | $d =~ s/^\s+//; | |
621 | $d =~ s/\s+$//; | |
622 | push @data, $d | |
623 | } | |
624 | ||
625 | my $et = shift @data; | |
626 | next if $et ne 'role'; | |
5b3bad37 | 627 | |
65b46cad | 628 | my ($role, $privlist) = @data; |
5b3bad37 LS |
629 | if (!PVE::AccessControl::verify_rolename($role, 1)) { |
630 | warn "user config - ignore role '$role' - invalid characters in role name\n"; | |
631 | next; | |
632 | } | |
633 | ||
634 | $roles->{$role} = {} if !$roles->{$role}; | |
635 | foreach my $priv (split_list($privlist)) { | |
636 | $roles->{$role}->{$priv} = 1; | |
637 | } | |
638 | } | |
639 | ||
640 | foreach my $role (sort keys %{$roles}) { | |
641 | if (PVE::AccessControl::role_is_special($role)) { | |
642 | next; | |
643 | } | |
644 | ||
645 | if ($role eq "PVEPoolUser") { | |
646 | # the user created a custom role named PVEPoolUser | |
647 | log_fail("Custom role '$role' has a restricted name - a built-in role 'PVEPoolUser' will be available with the upgrade"); | |
648 | } else { | |
649 | log_pass("Custom role '$role' has no restricted name"); | |
650 | } | |
651 | ||
652 | my $perms = $roles->{$role}; | |
653 | if ($perms->{'Pool.Allocate'} && $perms->{'Pool.Audit'}) { | |
654 | log_pass("Custom role '$role' contains updated pool permissions"); | |
655 | } elsif ($perms->{'Pool.Allocate'}) { | |
656 | log_warn("Custom role '$role' contains permission 'Pool.Allocate' - to ensure same behavior add 'Pool.Audit' to this role"); | |
657 | } else { | |
658 | log_pass("Custom role '$role' contains no permissions that need to be updated"); | |
659 | } | |
660 | } | |
661 | } | |
662 | ||
4177a14d TL |
663 | sub check_misc { |
664 | print_header("MISCELLANEOUS CHECKS"); | |
665 | my $ssh_config = eval { PVE::Tools::file_get_contents('/root/.ssh/config') }; | |
666 | if (defined($ssh_config)) { | |
667 | log_fail("Unsupported SSH Cipher configured for root in /root/.ssh/config: $1") | |
668 | if $ssh_config =~ /^Ciphers .*(blowfish|arcfour|3des).*$/m; | |
669 | } else { | |
670 | log_skip("No SSH config file found."); | |
671 | } | |
672 | ||
673 | log_info("Checking common daemon services.."); | |
674 | $log_systemd_unit_state->('pveproxy.service'); | |
675 | $log_systemd_unit_state->('pvedaemon.service'); | |
676 | $log_systemd_unit_state->('pvestatd.service'); | |
677 | ||
678 | my $root_free = PVE::Tools::df('/', 10); | |
679 | log_warn("Less than 2G free space on root file system.") | |
680 | if defined($root_free) && $root_free->{avail} < 2*1024*1024*1024; | |
681 | ||
682 | log_info("Checking for running guests.."); | |
683 | my $running_guests = 0; | |
684 | ||
685 | my $vms = eval { PVE::API2::Qemu->vmlist({ node => $nodename }) }; | |
686 | log_warn("Failed to retrieve information about this node's VMs - $@") if $@; | |
687 | $running_guests += grep { $_->{status} eq 'running' } @$vms if defined($vms); | |
688 | ||
689 | my $cts = eval { PVE::API2::LXC->vmlist({ node => $nodename }) }; | |
690 | log_warn("Failed to retrieve information about this node's CTs - $@") if $@; | |
691 | $running_guests += grep { $_->{status} eq 'running' } @$cts if defined($cts); | |
692 | ||
693 | if ($running_guests > 0) { | |
694 | log_warn("$running_guests running guest(s) detected - consider migrating or stopping them.") | |
695 | } else { | |
696 | log_pass("no running guest detected.") | |
697 | } | |
698 | ||
699 | log_info("Checking if the local node's hostname '$nodename' is resolvable.."); | |
700 | my $local_ip = eval { PVE::Network::get_ip_from_hostname($nodename) }; | |
701 | if ($@) { | |
702 | log_warn("Failed to resolve hostname '$nodename' to IP - $@"); | |
703 | } else { | |
704 | log_info("Checking if resolved IP is configured on local node.."); | |
705 | my $cidr = Net::IP::ip_is_ipv6($local_ip) ? "$local_ip/128" : "$local_ip/32"; | |
706 | my $configured_ips = PVE::Network::get_local_ip_from_cidr($cidr); | |
707 | my $ip_count = scalar(@$configured_ips); | |
708 | ||
709 | if ($ip_count <= 0) { | |
710 | log_fail("Resolved node IP '$local_ip' not configured or active for '$nodename'"); | |
711 | } elsif ($ip_count > 1) { | |
712 | log_warn("Resolved node IP '$local_ip' active on multiple ($ip_count) interfaces!"); | |
713 | } else { | |
714 | log_pass("Resolved node IP '$local_ip' configured and active on single interface."); | |
715 | } | |
716 | } | |
717 | ||
718 | log_info("Check node certificate's RSA key size"); | |
719 | my $certs = PVE::API2::Certificates->info({ node => $nodename }); | |
720 | my $certs_check = { | |
721 | 'rsaEncryption' => { | |
722 | minsize => 2048, | |
723 | name => 'RSA', | |
724 | }, | |
725 | 'id-ecPublicKey' => { | |
726 | minsize => 224, | |
727 | name => 'ECC', | |
728 | }, | |
729 | }; | |
730 | ||
731 | my $certs_check_failed = 0; | |
732 | foreach my $cert (@$certs) { | |
733 | my ($type, $size, $fn) = $cert->@{qw(public-key-type public-key-bits filename)}; | |
734 | ||
735 | if (!defined($type) || !defined($size)) { | |
736 | log_warn("'$fn': cannot check certificate, failed to get it's type or size!"); | |
737 | } | |
738 | ||
739 | my $check = $certs_check->{$type}; | |
740 | if (!defined($check)) { | |
741 | log_warn("'$fn': certificate's public key type '$type' unknown, check Debian Busters release notes"); | |
742 | next; | |
743 | } | |
744 | ||
745 | if ($size < $check->{minsize}) { | |
746 | log_fail("'$fn', certificate's $check->{name} public key size is less than 2048 bit"); | |
747 | $certs_check_failed = 1; | |
748 | } else { | |
749 | log_pass("Certificate '$fn' passed Debian Busters security level for TLS connections ($size >= 2048)"); | |
750 | } | |
751 | } | |
e522a2db FE |
752 | |
753 | check_backup_retention_settings(); | |
5df8b555 | 754 | check_cifs_credential_location(); |
5b3bad37 | 755 | check_custom_pool_roles(); |
4177a14d TL |
756 | } |
757 | ||
758 | __PACKAGE__->register_method ({ | |
759 | name => 'checklist', | |
760 | path => 'checklist', | |
761 | method => 'GET', | |
762 | description => 'Check (pre-/post-)upgrade conditions.', | |
763 | parameters => { | |
764 | additionalProperties => 0, | |
765 | properties => { | |
766 | }, | |
767 | }, | |
768 | returns => { type => 'null' }, | |
769 | code => sub { | |
770 | my ($param) = @_; | |
771 | ||
772 | check_pve_packages(); | |
773 | check_cluster_corosync(); | |
774 | check_ceph(); | |
775 | check_storage_health(); | |
776 | check_misc(); | |
777 | ||
778 | print_header("SUMMARY"); | |
779 | ||
780 | my $total = 0; | |
781 | $total += $_ for values %$counters; | |
782 | ||
783 | print "TOTAL: $total\n"; | |
784 | print colored("PASSED: $counters->{pass}\n", 'green'); | |
785 | print "SKIPPED: $counters->{skip}\n"; | |
786 | print colored("WARNINGS: $counters->{warn}\n", 'yellow'); | |
787 | print colored("FAILURES: $counters->{fail}\n", 'red'); | |
788 | ||
789 | if ($counters->{warn} > 0 || $counters->{fail} > 0) { | |
790 | my $color = $counters->{fail} > 0 ? 'red' : 'yellow'; | |
791 | print colored("\nATTENTION: Please check the output for detailed information!\n", $color); | |
792 | print colored("Try to solve the problems one at a time and then run this checklist tool again.\n", $color) if $counters->{fail} > 0; | |
793 | } | |
794 | ||
795 | return undef; | |
796 | }}); | |
797 | ||
798 | our $cmddef = [ __PACKAGE__, 'checklist', [], {}]; | |
799 | ||
800 | # for now drop all unknown params and just check | |
801 | @ARGV = (); | |
802 | ||
803 | 1; |