Line data Source code
1 : /*
2 : * Copyright (C) 2013 Andrea Mazzoleni
3 : *
4 : * This program is free software: you can redistribute it and/or modify
5 : * it under the terms of the GNU General Public License as published by
6 : * the Free Software Foundation, either version 3 of the License, or
7 : * (at your option) any later version.
8 : *
9 : * This program is distributed in the hope that it will be useful,
10 : * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 : * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 : * GNU General Public License for more details.
13 : *
14 : * You should have received a copy of the GNU General Public License
15 : * along with this program. If not, see <http://www.gnu.org/licenses/>.
16 : */
17 :
18 : #include "portable.h"
19 :
20 : #include "support.h"
21 : #include "elem.h"
22 : #include "state.h"
23 : #include "parity.h"
24 : #include "handle.h"
25 : #include "raid/raid.h"
26 :
27 : /****************************************************************************/
28 : /* status */
29 :
30 27 : unsigned day_ago(time_t ref, time_t now)
31 : {
32 : /* in case some dates is in the future */
33 27 : if (now < ref)
34 0 : return 0;
35 :
36 27 : return (now - ref) / (24 * 3600);
37 : }
38 :
39 : #define GRAPH_COLUMN 70
40 : #define GRAPH_ROW 15
41 :
42 : /**
43 : * Bit used to mark unscrubbed time info.
44 : */
45 : #define TIME_NEW 1
46 :
47 15 : int state_status(struct snapraid_state* state)
48 : {
49 : block_off_t blockmax;
50 : block_off_t i;
51 : time_t* timemap;
52 : time_t now;
53 : block_off_t bad;
54 : block_off_t bad_first;
55 : block_off_t bad_last;
56 : block_off_t rehash;
57 : block_off_t count;
58 : unsigned l;
59 : unsigned dayoldest, daymedian, daynewest;
60 : unsigned bar_scrubbed[GRAPH_COLUMN];
61 : unsigned bar_new[GRAPH_COLUMN];
62 : unsigned barpos;
63 : unsigned barmax;
64 : time_t oldest, newest, median;
65 : unsigned x, y;
66 : tommy_node* node_disk;
67 : unsigned file_count;
68 : unsigned file_fragmented;
69 : unsigned extra_fragment;
70 : unsigned file_zerosubsecond;
71 : uint64_t file_size;
72 : uint64_t file_block_count;
73 : uint64_t file_block_free;
74 : block_off_t parity_block_free;
75 : unsigned unsynced_blocks;
76 : unsigned unscrubbed_blocks;
77 : uint64_t all_wasted;
78 : int free_not_zero;
79 :
80 : /* get the present time */
81 15 : now = time(0);
82 :
83 : /* keep track if at least a free info is available */
84 15 : free_not_zero = 0;
85 :
86 15 : blockmax = parity_allocated_size(state);
87 :
88 15 : log_tag("summary:block_size:%u\n", state->block_size);
89 15 : log_tag("summary:parity_block_count:%u\n", blockmax);
90 :
91 : /* get the minimum parity free space */
92 15 : parity_block_free = state->parity[0].free_blocks;
93 105 : for (l = 0; l < state->level; ++l) {
94 90 : log_tag("summary:parity_block_total:%s:%u\n", lev_config_name(l), state->parity[l].total_blocks);
95 90 : log_tag("summary:parity_block_free:%s:%u\n", lev_config_name(l), state->parity[l].free_blocks);
96 90 : if (state->parity[l].free_blocks < parity_block_free)
97 1 : parity_block_free = state->parity[l].free_blocks;
98 90 : if (state->parity[l].free_blocks != 0)
99 79 : free_not_zero = 1;
100 : }
101 15 : log_tag("summary:parity_block_free_min:%u\n", parity_block_free);
102 :
103 15 : printf("SnapRAID status report:\n");
104 15 : printf("\n");
105 15 : printf(" Files Fragmented Excess Wasted Used Free Use Name\n");
106 15 : printf(" Files Fragments GB GB GB\n");
107 :
108 : /* count fragments */
109 15 : file_count = 0;
110 15 : file_size = 0;
111 15 : file_block_count = 0;
112 15 : file_block_free = 0;
113 15 : file_fragmented = 0;
114 15 : extra_fragment = 0;
115 15 : file_zerosubsecond = 0;
116 15 : all_wasted = 0;
117 105 : for (node_disk = state->disklist; node_disk != 0; node_disk = node_disk->next) {
118 90 : struct snapraid_disk* disk = node_disk->data;
119 : tommy_node* node;
120 : block_off_t j;
121 90 : unsigned disk_file_count = 0;
122 90 : unsigned disk_file_fragmented = 0;
123 90 : unsigned disk_extra_fragment = 0;
124 90 : unsigned disk_file_zerosubsecond = 0;
125 90 : block_off_t disk_block_count = 0;
126 90 : uint64_t disk_file_size = 0;
127 90 : block_off_t disk_block_latest_used = 0;
128 : block_off_t disk_block_max_by_space;
129 : block_off_t disk_block_max_by_parity;
130 : block_off_t disk_block_max;
131 : int64_t wasted;
132 :
133 : /* for each file in the disk */
134 90 : node = disk->filelist;
135 95624 : while (node) {
136 : struct snapraid_file* file;
137 :
138 95534 : file = node->data;
139 95534 : node = node->next; /* next node */
140 :
141 95534 : if (file->mtime_nsec == STAT_NSEC_INVALID
142 95534 : || file->mtime_nsec == 0
143 : ) {
144 0 : ++file_zerosubsecond;
145 0 : ++disk_file_zerosubsecond;
146 0 : if (disk_file_zerosubsecond < 50)
147 0 : log_tag("zerosubsecond:%s:%s: \n", disk->name, file->sub);
148 0 : if (disk_file_zerosubsecond == 50)
149 0 : log_tag("zerosubsecond:%s:%s: (more follow)\n", disk->name, file->sub);
150 : }
151 :
152 : /* check fragmentation */
153 95534 : if (file->blockmax != 0) {
154 : block_off_t prev_pos;
155 : block_off_t last_pos;
156 : int fragmented;
157 :
158 95396 : fragmented = 0;
159 95396 : prev_pos = fs_file2par_get(disk, file, 0);
160 235735 : for (j = 1; j < file->blockmax; ++j) {
161 140339 : block_off_t parity_pos = fs_file2par_get(disk, file, j);
162 140339 : if (prev_pos + 1 != parity_pos) {
163 698 : fragmented = 1;
164 698 : ++extra_fragment;
165 698 : ++disk_extra_fragment;
166 : }
167 140339 : prev_pos = parity_pos;
168 : }
169 :
170 : /* keep track of latest block used */
171 95396 : last_pos = fs_file2par_get(disk, file, file->blockmax - 1);
172 95396 : if (last_pos > disk_block_latest_used) {
173 92537 : disk_block_latest_used = last_pos;
174 : }
175 :
176 95396 : if (fragmented) {
177 582 : ++file_fragmented;
178 582 : ++disk_file_fragmented;
179 : }
180 :
181 95396 : disk_block_count += file->blockmax;
182 : }
183 :
184 : /* count files */
185 95534 : ++file_count;
186 95534 : ++disk_file_count;
187 95534 : file_size += file->size;
188 95534 : file_block_count += file->blockmax;
189 95534 : disk_file_size += file->size;
190 : }
191 :
192 90 : if (disk->free_blocks != 0)
193 53 : free_not_zero = 1;
194 :
195 : /* get the free block info */
196 90 : disk_block_max_by_space = disk_block_count + disk->free_blocks;
197 90 : disk_block_max_by_parity = blockmax + parity_block_free;
198 :
199 : /* the maximum usable space in a disk is limited by the smallest */
200 : /* of the disk size and the parity size */
201 : /* the wasted space is the space that we have to leave */
202 : /* free on the data disk, when the parity is filled up */
203 90 : if (disk_block_max_by_space < disk_block_max_by_parity) {
204 78 : disk_block_max = disk_block_max_by_space;
205 : } else {
206 12 : disk_block_max = disk_block_max_by_parity;
207 : }
208 :
209 : /* wasted space is the difference of the two maximum size */
210 : /* if negative, it's extra space available in parity */
211 90 : wasted = (int64_t)disk_block_max_by_space - (int64_t)disk_block_max_by_parity;
212 90 : wasted *= state->block_size;
213 :
214 90 : if (wasted > 0)
215 6 : all_wasted += wasted;
216 90 : file_block_free += disk_block_max - disk_block_count;
217 :
218 90 : printf("%8u", disk_file_count);
219 90 : printf("%8u", disk_file_fragmented);
220 90 : printf("%8u", disk_extra_fragment);
221 90 : if (wasted < -100LL * GIGA) {
222 78 : printf(" -");
223 : } else {
224 12 : printf("%8.1f", (double)wasted / GIGA);
225 : }
226 90 : printf("%8" PRIu64, disk_file_size / GIGA);
227 :
228 90 : if (disk_block_max == 0 && disk_block_count == 0) {
229 : /* if the disk is empty and we don't have the free space info */
230 37 : printf(" -");
231 37 : printf(" - ");
232 : } else {
233 53 : printf("%8" PRIu64, (disk_block_max - disk_block_count) * (uint64_t)state->block_size / GIGA);
234 53 : printf(" %3u%%", muldiv(disk_block_count, 100, disk_block_max));
235 : }
236 90 : printf(" %s\n", disk->name);
237 :
238 90 : log_tag("summary:disk_file_count:%s:%u\n", disk->name, disk_file_count);
239 90 : log_tag("summary:disk_block_count:%s:%u\n", disk->name, disk_block_count);
240 90 : log_tag("summary:disk_fragmented_file_count:%s:%u\n", disk->name, disk_file_fragmented);
241 90 : log_tag("summary:disk_excess_fragment_count:%s:%u\n", disk->name, disk_extra_fragment);
242 90 : log_tag("summary:disk_zerosubsecond_file_count:%s:%u\n", disk->name, disk_file_zerosubsecond);
243 90 : log_tag("summary:disk_file_size:%s:%" PRIu64 "\n", disk->name, disk_file_size);
244 90 : log_tag("summary:disk_block_allocated:%s:%u\n", disk->name, disk_block_latest_used + 1);
245 90 : log_tag("summary:disk_block_total:%s:%u\n", disk->name, disk->total_blocks);
246 90 : log_tag("summary:disk_block_free:%s:%u\n", disk->name, disk->free_blocks);
247 90 : log_tag("summary:disk_block_max_by_space:%s:%u\n", disk->name, disk_block_max_by_space);
248 90 : log_tag("summary:disk_block_max_by_parity:%s:%u\n", disk->name, disk_block_max_by_parity);
249 90 : log_tag("summary:disk_block_max:%s:%u\n", disk->name, disk_block_max);
250 90 : log_tag("summary:disk_space_wasted:%s:%" PRId64 "\n", disk->name, wasted);
251 : }
252 :
253 : /* totals */
254 15 : printf(" --------------------------------------------------------------------------\n");
255 15 : printf("%8u", file_count);
256 15 : printf("%8u", file_fragmented);
257 15 : printf("%8u", extra_fragment);
258 15 : printf("%8.1f", (double)all_wasted / GIGA);
259 15 : printf("%8" PRIu64, file_size / GIGA);
260 15 : printf("%8" PRIu64, file_block_free * state->block_size / GIGA);
261 15 : printf(" %3u%%", muldiv(file_block_count, 100, file_block_count + file_block_free));
262 15 : printf("\n");
263 :
264 : /* warn about invalid data free info */
265 15 : if (!free_not_zero)
266 1 : printf("\nWARNING! Free space info will be valid after the first sync.\n");
267 :
268 15 : log_tag("summary:file_count:%u\n", file_count);
269 15 : log_tag("summary:file_block_count:%" PRIu64 "\n", file_block_count);
270 15 : log_tag("summary:fragmented_file_count:%u\n", file_fragmented);
271 15 : log_tag("summary:excess_fragment_count:%u\n", extra_fragment);
272 15 : log_tag("summary:zerosubsecond_file_count:%u\n", file_zerosubsecond);
273 15 : log_tag("summary:file_size:%" PRIu64 "\n", file_size);
274 15 : log_tag("summary:parity_size:%" PRIu64 "\n", blockmax * (uint64_t)state->block_size);
275 15 : log_tag("summary:parity_size_max:%" PRIu64 "\n", (blockmax + parity_block_free) * (uint64_t)state->block_size);
276 15 : log_tag("summary:hash:%s\n", hash_config_name(state->hash));
277 15 : log_tag("summary:prev_hash:%s\n", hash_config_name(state->prevhash));
278 15 : log_tag("summary:best_hash:%s\n", hash_config_name(state->besthash));
279 15 : log_flush();
280 :
281 : /* copy the info a temp vector, and count bad/rehash/unsynced blocks */
282 15 : timemap = malloc_nofail(blockmax * sizeof(time_t));
283 15 : bad = 0;
284 15 : bad_first = 0;
285 15 : bad_last = 0;
286 15 : count = 0;
287 15 : rehash = 0;
288 15 : unsynced_blocks = 0;
289 15 : unscrubbed_blocks = 0;
290 15 : log_tag("block_count:%u\n", blockmax);
291 39823 : for (i = 0; i < blockmax; ++i) {
292 : int one_invalid;
293 : int one_valid;
294 :
295 39808 : snapraid_info info = info_get(&state->infoarr, i);
296 :
297 : /* for each disk */
298 39808 : one_invalid = 0;
299 39808 : one_valid = 0;
300 278656 : for (node_disk = state->disklist; node_disk != 0; node_disk = node_disk->next) {
301 238848 : struct snapraid_disk* disk = node_disk->data;
302 238848 : struct snapraid_block* block = fs_par2block_find(disk, i);
303 :
304 238848 : if (block_has_file(block))
305 235735 : one_valid = 1;
306 238848 : if (block_has_invalid_parity(block))
307 56 : one_invalid = 1;
308 : }
309 :
310 : /* if both valid and invalid, we need to update */
311 39808 : if (one_invalid && one_valid) {
312 40 : ++unsynced_blocks;
313 : }
314 :
315 : /* skip unused blocks */
316 39808 : if (info != 0) {
317 : time_t scrub_time;
318 :
319 39808 : if (info_get_bad(info)) {
320 1772 : if (bad == 0)
321 2 : bad_first = i;
322 1772 : bad_last = i;
323 1772 : ++bad;
324 : }
325 :
326 39808 : if (info_get_rehash(info))
327 4607 : ++rehash;
328 :
329 39808 : scrub_time = info_get_time(info);
330 :
331 39808 : if (info_get_justsynced(info)) {
332 21356 : ++unscrubbed_blocks;
333 :
334 : /* mark the time as not scrubbed */
335 21356 : scrub_time |= TIME_NEW;
336 : }
337 :
338 39808 : timemap[count++] = scrub_time;
339 : }
340 :
341 39808 : if (state->opt.gui) {
342 0 : if (info != 0)
343 0 : log_tag("block:%u:%" PRIu64 ":%s:%s:%s:%s\n", i, (uint64_t)info_get_time(info), one_valid ? "used" : "", one_invalid ? "unsynced" : "", info_get_bad(info) ? "bad" : "", info_get_rehash(info) ? "rehash" : "");
344 : else
345 0 : log_tag("block_noinfo:%u:%s:%s\n", i, one_valid ? "used" : "", one_invalid ? "unsynced" : "");
346 : }
347 : }
348 :
349 15 : log_tag("summary:has_unsynced:%u\n", unsynced_blocks);
350 15 : log_tag("summary:has_unscrubbed:%u\n", unscrubbed_blocks);
351 15 : log_tag("summary:has_rehash:%u\n", rehash);
352 15 : log_tag("summary:has_bad:%u:%u:%u\n", bad, bad_first, bad_last);
353 15 : log_flush();
354 :
355 15 : if (!count) {
356 6 : log_fatal("The array is empty.\n");
357 6 : free(timemap);
358 6 : return 0;
359 : }
360 :
361 : /* sort the info to get the time info */
362 9 : qsort(timemap, count, sizeof(time_t), time_compare);
363 :
364 : /* output the info map */
365 9 : i = 0;
366 9 : log_tag("info_count:%u\n", count);
367 26 : while (i < count) {
368 17 : unsigned j = i + 1;
369 39808 : while (j < count && timemap[i] == timemap[j])
370 39791 : ++j;
371 17 : if ((timemap[i] & TIME_NEW) == 0) {
372 7 : log_tag("info_time:%" PRIu64 ":%u:scrubbed\n", (uint64_t)timemap[i], j - i);
373 : } else {
374 10 : log_tag("info_time:%" PRIu64 ":%u:new\n", (uint64_t)(timemap[i] & ~TIME_NEW), j - i);
375 : }
376 17 : i = j;
377 : }
378 :
379 9 : oldest = timemap[0];
380 9 : median = timemap[count / 2];
381 9 : newest = timemap[count - 1];
382 9 : dayoldest = day_ago(oldest, now);
383 9 : daymedian = day_ago(median, now);
384 9 : daynewest = day_ago(newest, now);
385 :
386 : /* compute graph limits */
387 9 : barpos = 0;
388 9 : barmax = 0;
389 639 : for (i = 0; i < GRAPH_COLUMN; ++i) {
390 : time_t limit;
391 : unsigned step_scrubbed, step_new;
392 :
393 630 : limit = oldest + (newest - oldest) * (i + 1) / GRAPH_COLUMN;
394 :
395 630 : step_scrubbed = 0;
396 630 : step_new = 0;
397 40438 : while (barpos < count && timemap[barpos] <= limit) {
398 39808 : if ((timemap[barpos] & TIME_NEW) != 0)
399 21356 : ++step_new;
400 : else
401 18452 : ++step_scrubbed;
402 39808 : ++barpos;
403 : }
404 :
405 630 : if (step_new + step_scrubbed > barmax)
406 12 : barmax = step_new + step_scrubbed;
407 :
408 630 : bar_scrubbed[i] = step_scrubbed;
409 630 : bar_new[i] = step_new;
410 : }
411 :
412 9 : printf("\n\n");
413 :
414 : /* print the graph */
415 144 : for (y = 0; y < GRAPH_ROW; ++y) {
416 135 : if (y == 0)
417 9 : printf("%3u%%|", barmax * 100 / count);
418 126 : else if (y == GRAPH_ROW - 1)
419 9 : printf(" 0%%|");
420 117 : else if (y == GRAPH_ROW / 2)
421 9 : printf("%3u%%|", barmax * 50 / count);
422 : else
423 108 : printf(" |");
424 9585 : for (x = 0; x < GRAPH_COLUMN; ++x) {
425 9450 : unsigned pivot_upper = barmax * (GRAPH_ROW - y) / GRAPH_ROW;
426 9450 : unsigned pivot_lower = barmax * (GRAPH_ROW - 1 - y) / GRAPH_ROW;
427 9450 : unsigned both = bar_scrubbed[x] + bar_new[x];
428 9450 : unsigned scrubbed = bar_scrubbed[x];
429 :
430 9450 : if (both > pivot_upper) {
431 156 : if (scrubbed > pivot_lower)
432 61 : printf("*");
433 : else
434 95 : printf("o");
435 9294 : } else if (both > pivot_lower) {
436 16 : if (scrubbed == both)
437 6 : printf("*");
438 : else
439 10 : printf("o");
440 : } else {
441 9278 : if (y == GRAPH_ROW - 1)
442 614 : printf("_");
443 : else
444 8664 : printf(" ");
445 : }
446 : }
447 135 : printf("\n");
448 : }
449 9 : printf(" %3u days ago of the last scrub/sync %3u\n", dayoldest, daynewest);
450 :
451 9 : printf("\n");
452 :
453 9 : printf("The oldest block was scrubbed %u days ago, the median %u, the newest %u.\n", dayoldest, daymedian, daynewest);
454 :
455 9 : printf("\n");
456 :
457 9 : if (newest > now) {
458 0 : printf("WARNING! You have scrub dates in the future! The next sync/scrub will truncate them!\n");
459 : }
460 :
461 9 : if (unsynced_blocks) {
462 1 : printf("WARNING! The array is NOT fully synced.\n");
463 1 : printf("You have a sync in progress at %u%%.\n", muldiv(blockmax - unsynced_blocks, 100, blockmax));
464 : } else {
465 8 : printf("No sync is in progress.\n");
466 : }
467 :
468 9 : if (unscrubbed_blocks) {
469 7 : printf("%u%% of the array is not scrubbed.\n", muldiv_upper(unscrubbed_blocks, 100, blockmax));
470 : } else {
471 2 : printf("The full array was scrubbed at least one time.\n");
472 : }
473 :
474 9 : if (file_zerosubsecond) {
475 0 : printf("You have %u files with a zero sub-second timestamp.\n", file_zerosubsecond);
476 0 : printf("Run 'snapraid touch' to set their sub-second timestamps to a non-zero value.\n");
477 : } else {
478 9 : printf("No file has a zero sub-second timestamp.\n");
479 : }
480 :
481 9 : if (rehash) {
482 1 : printf("You have a rehash in progress at %u%%.\n", muldiv(count - rehash, 100, count));
483 : } else {
484 8 : if (state->besthash != state->hash) {
485 7 : printf("No rehash is in progress, but for optimal performance one is recommended.\n");
486 : } else {
487 1 : printf("No rehash is in progress or needed.\n");
488 : }
489 : }
490 :
491 9 : if (bad) {
492 2 : printf("DANGER! In the array there are %u errors!\n\n", bad);
493 :
494 2 : if (bad_last - bad_first + 1 == bad) {
495 0 : printf("They are from block %u to %u.\n", bad_first, bad_last);
496 : } else {
497 : block_off_t bad_range;
498 : block_off_t bad_count;
499 : block_off_t range_start;
500 : block_off_t range_count;
501 :
502 2 : printf("They are from block %u to %u, specifically at blocks:", bad_first, bad_last);
503 :
504 : /* print some of the errors */
505 2 : bad_range = 0;
506 2 : bad_count = 0;
507 2 : range_start = 0;
508 2 : range_count = 0;
509 885 : for (i = 0; i <= blockmax; ++i) { /* one extra iteration to print the final range */
510 885 : snapraid_info info = 0;
511 885 : int is_bad = 0;
512 :
513 885 : if (i < blockmax) {
514 885 : info = info_get(&state->infoarr, i);
515 885 : if (info != 0) /* unused blocks are never bad */
516 885 : is_bad = info_get_bad(info);
517 : }
518 :
519 885 : if (is_bad) {
520 : /* create or extend the range */
521 258 : if (!range_count)
522 202 : range_start = i;
523 258 : ++range_count;
524 : } else {
525 : /* break the range */
526 627 : if (range_count) {
527 202 : if (range_count == 1) {
528 157 : printf(" %u", range_start);
529 : } else {
530 45 : printf(" %u-%u", range_start, range_start + range_count - 1);
531 : }
532 202 : bad_count += range_count;
533 202 : ++bad_range;
534 202 : range_count = 0;
535 : }
536 : }
537 :
538 885 : if (bad_range > 100) {
539 2 : printf(" and %u more...", bad - bad_count);
540 2 : break;
541 : }
542 : }
543 :
544 2 : printf("\n");
545 : }
546 :
547 2 : printf("\n");
548 :
549 2 : printf("To fix them use the command 'snapraid -e fix'.\n");
550 2 : printf("The errors will disappear from the 'status' at the next 'scrub' command.\n");
551 : } else {
552 7 : printf("No error detected.\n");
553 : }
554 :
555 : /* free the temp vector */
556 9 : free(timemap);
557 :
558 9 : return 0;
559 : }
560 :
|
