1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117
|
/*
* Copyright (C) 2011 Andrea Mazzoleni
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "portable.h"
#include "support.h"
#include "util.h"
#include "elem.h"
#include "import.h"
#include "search.h"
#include "state.h"
#include "parity.h"
#include "handle.h"
#include "raid/raid.h"
#include "raid/combo.h"
/****************************************************************************/
/* check */
/**
* A block that failed the hash check, or that was deleted.
*/
struct failed_struct {
/**
* If we know for sure that the block is garbage or missing
* and it needs to be recovered and rewritten to the disk.
*/
int is_bad;
/**
* If that we have recovered may be not updated data,
* an old version, or just garbage.
*
* Essentially, it means that we are not sure what we have recovered
* is really correct. It's just our best guess.
*
* These "recovered" block are also written to the disk if the block is marked as ::is_bad.
* But these files are marked also as FILE_IS_DAMAGED, and then renamed to .unrecoverable.
*
* Note that this could happen only for CHG blocks.
*/
int is_outofdate;
unsigned index; /**< Index of the failed block. */
struct snapraid_block* block; /**< The failed block */
struct snapraid_disk* disk; /**< The failed disk. */
struct snapraid_file* file; /**< The failed file. 0 for DELETED block. */
block_off_t file_pos; /**< Offset inside the file */
struct snapraid_handle* handle; /**< The handle containing the failed block, or 0 for a DELETED block */
};
/**
* Check if a block hash matches the specified buffer.
* Return ==0 if equal
*/
static int blockcmp(struct snapraid_state* state, int rehash, struct snapraid_block* block, unsigned pos_size, unsigned char* buffer, unsigned char* buffer_zero)
{
unsigned char hash[HASH_MAX];
/* now compute the hash of the valid part */
if (rehash) {
memhash(state->prevhash, state->prevhashseed, hash, buffer, pos_size);
} else {
memhash(state->hash, state->hashseed, hash, buffer, pos_size);
}
/* compare the hash */
if (memcmp(hash, block->hash, BLOCK_HASH_SIZE) != 0) {
return -1;
}
/* compare to the end of the block */
if (pos_size < state->block_size) {
if (memcmp(buffer + pos_size, buffer_zero + pos_size, state->block_size - pos_size) != 0) {
return -1;
}
}
return 0;
}
/**
* Check if the hash of all the failed block we are expecting to recover are now matching.
*/
static int is_hash_matching(struct snapraid_state* state, int rehash, unsigned diskmax, struct failed_struct* failed, unsigned* failed_map, unsigned failed_count, void** buffer, void* buffer_zero)
{
unsigned j;
int hash_checked;
hash_checked = 0; /* keep track if we check at least one block */
/* check if the recovered blocks are OK */
for (j = 0; j < failed_count; ++j) {
/* if we are expected to recover this block */
if (!failed[failed_map[j]].is_outofdate
/* if the block has a hash to check */
&& block_has_updated_hash(failed[failed_map[j]].block)
) {
/* if a hash doesn't match, fail the check */
unsigned pos_size = file_block_size(failed[failed_map[j]].file, failed[failed_map[j]].file_pos, state->block_size);
if (blockcmp(state, rehash, failed[failed_map[j]].block, pos_size, buffer[failed[failed_map[j]].index], buffer_zero) != 0) {
log_tag("hash_error: Hash mismatch on entry %u\n", failed_map[j]);
return 0;
}
hash_checked = 1;
}
}
/* if nothing checked, we reject it */
/* note that we are excluding this case at upper level */
/* but checking again doesn't hurt */
if (!hash_checked) {
/* LCOV_EXCL_START */
return 0;
/* LCOV_EXCL_STOP */
}
/* if we checked something, and no block failed the check */
/* recompute all the redundancy information */
raid_gen(diskmax, state->level, state->block_size, buffer);
return 1;
}
/**
* Check if specified parity is now matching with a recomputed one.
*/
static int is_parity_matching(struct snapraid_state* state, unsigned diskmax, unsigned i, void** buffer, void** buffer_recov)
{
/* recompute parity, note that we don't need parity over i */
raid_gen(diskmax, i + 1, state->block_size, buffer);
/* if the recovered parity block matches */
if (memcmp(buffer[diskmax + i], buffer_recov[i], state->block_size) == 0) {
/* recompute all the redundancy information */
raid_gen(diskmax, state->level, state->block_size, buffer);
return 1;
}
return 0;
}
/**
* Repair errors.
* Return <0 if failure for missing strategy, >0 if data is wrong and we cannot rebuild correctly, 0 on success.
* If success, the parity are computed in the buffer variable.
*/
static int repair_step(struct snapraid_state* state, int rehash, unsigned pos, unsigned diskmax, struct failed_struct* failed, unsigned* failed_map, unsigned failed_count, void** buffer, void** buffer_recov, void* buffer_zero)
{
unsigned i, n;
int error;
int has_hash;
int id[LEV_MAX];
int ip[LEV_MAX];
/* no fix required, already checked at higher level, but just to be sure */
if (failed_count == 0) {
/* LCOV_EXCL_START */
/* recompute only the parity */
raid_gen(diskmax, state->level, state->block_size, buffer);
return 0;
/* LCOV_EXCL_STOP */
}
n = state->level;
error = 0;
/* setup vector of failed disk indexes */
for (i = 0; i < failed_count; ++i)
id[i] = failed[failed_map[i]].index;
/* check if there is at least a failed block that can be checked for correctness using the hash */
/* if there isn't, we have to sacrifice a parity block to check that the result is correct */
has_hash = 0;
for (i = 0; i < failed_count; ++i) {
/* if we are expected to recover this block */
if (!failed[failed_map[i]].is_outofdate
/* if the block has a hash to check */
&& block_has_updated_hash(failed[failed_map[i]].block)
)
has_hash = 1;
}
/* if we don't have a hash, but we have an extra parity */
/* (strictly-less failures than number of parities) */
if (!has_hash && failed_count < n) {
/* number of parity to use, one more to check the recovering */
unsigned r = failed_count + 1;
/* all combinations (r of n) parities */
combination_first(r, n, ip);
do {
/* if a parity is missing, do nothing */
for (i = 0; i < r; ++i) {
if (buffer_recov[ip[i]] == 0)
break;
}
if (i != r)
continue;
/* copy the parities to use, one less because the last is used for checking */
for (i = 0; i < r - 1; ++i)
memcpy(buffer[diskmax + ip[i]], buffer_recov[ip[i]], state->block_size);
/* recover using one less parity, the ip[r-1] one */
raid_data(r - 1, id, ip, diskmax, state->block_size, buffer);
/* use the remaining ip[r-1] parity to check the result */
if (is_parity_matching(state, diskmax, ip[r - 1], buffer, buffer_recov))
return 0;
/* log */
log_tag("parity_error:%u:", pos);
for (i = 0; i < r; ++i) {
if (i != 0)
log_tag("/");
log_tag("%s", lev_config_name(ip[i]));
}
log_tag(":parity: Parity mismatch\n");
++error;
} while (combination_next(r, n, ip));
}
/* if we have a hash, and enough parities */
/* (less-or-equal failures than number of parities) */
if (has_hash && failed_count <= n) {
/* number of parities to use equal at the number of failures */
unsigned r = failed_count;
/* all combinations (r of n) parities */
combination_first(r, n, ip);
do {
/* if a parity is missing, do nothing */
for (i = 0; i < r; ++i) {
if (buffer_recov[ip[i]] == 0)
break;
}
if (i != r)
continue;
/* copy the parities to use */
for (i = 0; i < r; ++i)
memcpy(buffer[diskmax + ip[i]], buffer_recov[ip[i]], state->block_size);
/* recover */
raid_data(r, id, ip, diskmax, state->block_size, buffer);
/* use the hash to check the result */
if (is_hash_matching(state, rehash, diskmax, failed, failed_map, failed_count, buffer, buffer_zero))
return 0;
/* log */
log_tag("parity_error:%u:", pos);
for (i = 0; i < r; ++i) {
if (i != 0)
log_tag("/");
log_tag("%s", lev_config_name(ip[i]));
}
log_tag(":hash: Hash mismatch\n");
++error;
} while (combination_next(r, n, ip));
}
/* return the number of failed attempts, or -1 if no strategy */
if (error)
return error;
log_tag("strategy_error:%u: No strategy to recover from %u failures with %u parity %s hash\n",
pos, failed_count, n, has_hash ? "with" : "without");
return -1;
}
static int repair(struct snapraid_state* state, int rehash, unsigned pos, unsigned diskmax, struct failed_struct* failed, unsigned* failed_map, unsigned failed_count, void** buffer, void** buffer_recov, void* buffer_zero)
{
int ret;
int error;
unsigned j;
int n;
int something_to_recover;
int something_unsynced;
char esc_buffer[ESC_MAX];
error = 0;
/* if nothing failed, just recompute the parity */
if (failed_count == 0) {
raid_gen(diskmax, state->level, state->block_size, buffer);
return 0;
}
/* logs the status */
for (j = 0; j < failed_count; ++j) {
const char* desc;
const char* hash;
const char* data;
struct snapraid_block* block = failed[j].block;
unsigned block_state = block_state_get(block);
switch (block_state) {
case BLOCK_STATE_DELETED : desc = "delete"; break;
case BLOCK_STATE_CHG : desc = "change"; break;
case BLOCK_STATE_REP : desc = "replace"; break;
case BLOCK_STATE_BLK : desc = "block"; break;
/* LCOV_EXCL_START */
default : desc = "unknown"; break;
/* LCOV_EXCL_STOP */
}
if (hash_is_invalid(block->hash)) {
hash = "lost";
} else if (hash_is_zero(block->hash)) {
hash = "zero";
} else {
hash = "known";
}
if (failed[j].is_bad)
data = "bad";
else
data = "good";
if (failed[j].file) {
struct snapraid_disk* disk = failed[j].disk;
struct snapraid_file* file = failed[j].file;
block_off_t file_pos = failed[j].file_pos;
log_tag("entry:%u:%s:%s:%s:%s:%s:%u:\n", j, desc, hash, data, disk->name, esc_tag(file->sub, esc_buffer), file_pos);
} else {
log_tag("entry:%u:%s:%s:%s:\n", j, desc, hash, data);
}
}
/* Here we have to try two different strategies to recover, because in case the 'sync' */
/* process is aborted, we don't know if the parity data is really updated just like after 'sync', */
/* or if it still represents the state before the 'sync'. */
/* Note that if the 'sync' ends normally, we don't have any DELETED, REP and CHG blocks */
/* and the two strategies are identical */
/* As first, we assume that the parity IS updated for the current state */
/* and that we are going to recover the state after the last 'sync'. */
/* In this case, parity contains info from BLK, REP and CHG blocks, */
/* but not for DELETED. */
/* We need to put in the recovering process only the bad blocks, because all the */
/* others already contains the correct data read from disk, and the parity is correctly computed for them. */
/* We are interested to recover BLK, REP and CHG blocks if they are marked as bad, */
/* but we are not interested in DELETED ones. */
n = 0;
something_to_recover = 0; /* keep track if there is at least one block to fix */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad) {
unsigned block_state = block_state_get(failed[j].block);
assert(block_state != BLOCK_STATE_DELETED); /* we cannot have bad DELETED blocks */
/* if we have the hash for it */
if ((block_state == BLOCK_STATE_BLK || block_state == BLOCK_STATE_REP)
/* try to fetch the block using the known hash */
&& (state_import_fetch(state, rehash, failed[j].block, buffer[failed[j].index]) == 0
|| state_search_fetch(state, rehash, failed[j].file, failed[j].file_pos, failed[j].block, buffer[failed[j].index]) == 0)
) {
/* we already have corrected it! */
log_tag("hash_import: Fixed entry %u\n", j);
} else {
/* otherwise try to recover it */
failed_map[n] = j;
++n;
/* we have something to try to recover */
something_to_recover = 1;
}
}
}
/* if nothing to fix */
if (!something_to_recover) {
log_tag("recover_sync:%u:%u: Skipped for already recovered\n", pos, n);
/* recompute only the parity */
raid_gen(diskmax, state->level, state->block_size, buffer);
return 0;
}
ret = repair_step(state, rehash, pos, diskmax, failed, failed_map, n, buffer, buffer_recov, buffer_zero);
if (ret == 0) {
/* reprocess the CHG blocks, for which we don't have a hash to check */
/* if they were BAD we have to use some heuristics to ensure that we have recovered */
/* the state after the sync. If unsure, we assume the worst case */
for (j = 0; j < failed_count; ++j) {
/* we take care only of BAD blocks we have to write back */
if (failed[j].is_bad) {
unsigned block_state = block_state_get(failed[j].block);
/* BLK and REP blocks are always OK, because at this point */
/* we have already checked their hash */
if (block_state != BLOCK_STATE_CHG) {
assert(block_state == BLOCK_STATE_BLK || block_state == BLOCK_STATE_REP);
continue;
}
/* for CHG blocks we have to 'guess' if they are correct or not */
/* if the hash is invalid we cannot check the result */
/* this could happen if we have lost this information */
/* after an aborted sync */
if (hash_is_invalid(failed[j].block->hash)) {
/* it may contain garbage */
failed[j].is_outofdate = 1;
log_tag("hash_unknown: Unknown hash on entry %u\n", j);
} else if (hash_is_zero(failed[j].block->hash)) {
/* if the block is not filled with 0, we are sure to have */
/* restored it to the state after the 'sync' */
/* instead, if the block is filled with 0, it could be either that the */
/* block after the sync is really filled by 0, or that */
/* we restored the block before the 'sync'. */
if (memcmp(buffer[failed[j].index], buffer_zero, state->block_size) == 0) {
/* it may contain garbage */
failed[j].is_outofdate = 1;
log_tag("hash_unknown: Maybe old zero on entry %u\n", j);
}
} else {
/* if the hash is different than the previous one, we are sure to have */
/* restored it to the state after the 'sync' */
/* instead, if the hash matches, it could be either that the */
/* block after the sync has this hash, or that */
/* we restored the block before the 'sync'. */
unsigned pos_size = file_block_size(failed[j].file, failed[j].file_pos, state->block_size);
if (blockcmp(state, rehash, failed[j].block, pos_size, buffer[failed[j].index], buffer_zero) == 0) {
/* it may contain garbage */
failed[j].is_outofdate = 1;
log_tag("hash_unknown: Maybe old data on entry %u\n", j);
}
}
}
}
return 0;
}
if (ret > 0)
error += ret;
if (ret < 0)
log_tag("recover_sync:%u:%u: Failed with no attempts\n", pos, n);
else
log_tag("recover_sync:%u:%u: Failed with %d attempts\n", pos, n, ret);
/* Now assume that the parity IS NOT updated at the current state, */
/* but still represent the state before the last 'sync' process. */
/* In this case, parity contains info from BLK, REP (old version), CHG (old version) and DELETED blocks, */
/* but not for REP (new version) and CHG (new version). */
/* We are interested to recover BLK ones marked as bad, */
/* but we are not interested to recover CHG (new version) and REP (new version) blocks, */
/* even if marked as bad, because we don't have parity for them and it's just impossible, */
/* and we are not interested to recover DELETED ones. */
n = 0;
something_to_recover = 0; /* keep track if there is at least one block to fix */
something_unsynced = 0; /* keep track if we have some unsynced info to process */
for (j = 0; j < failed_count; ++j) {
unsigned block_state = block_state_get(failed[j].block);
if (block_state == BLOCK_STATE_DELETED
|| block_state == BLOCK_STATE_CHG
|| block_state == BLOCK_STATE_REP
) {
/* If the block is CHG, REP or DELETED, we don't have the original content of block, */
/* and we must try to recover it. */
/* This apply to CHG and REP blocks even if they are not marked bad, */
/* because the parity is computed with old content, and not with the new one. */
/* Note that this recovering is done just to make possible to recover any other BLK one, */
/* we are not really interested in DELETED, CHG (old version) and REP (old version). */
something_unsynced = 1;
if (block_state == BLOCK_STATE_CHG
&& hash_is_zero(failed[j].block->hash)
) {
/* If the block was a ZERO block, restore it to the original 0 as before the 'sync' */
/* We do this to just allow recovering of other BLK ones */
memset(buffer[failed[j].index], 0, state->block_size);
/* note that from now the buffer is definitively lost */
/* we can do this only because it's the last retry of recovering */
/* try to fetch the old block using the old hash for CHG and DELETED blocks */
} else if ((block_state == BLOCK_STATE_CHG || block_state == BLOCK_STATE_DELETED)
&& hash_is_unique(failed[j].block->hash)
&& state_import_fetch(state, rehash, failed[j].block, buffer[failed[j].index]) == 0) {
/* note that from now the buffer is definitively lost */
/* we can do this only because it's the last retry of recovering */
} else {
/* otherwise try to recover it */
failed_map[n] = j;
++n;
/* note that we don't set something_to_recover, because we are */
/* not really interested to recover *only* old blocks. */
}
/* avoid to use the hash of this block to verify the recovering */
/* this applies to REP blocks because we are going to recover the old state */
/* and the REP hash represent the new one */
/* it also applies to CHG and DELETE blocks because we want to have */
/* a successful recovering only if a BLK one is matching */
failed[j].is_outofdate = 1;
} else if (failed[j].is_bad) {
/* If the block is bad we don't know its content, and we try to recover it */
/* At this point, we can have only BLK ones */
assert(block_state == BLOCK_STATE_BLK);
/* we have something we are interested to recover */
something_to_recover = 1;
/* we try to recover it */
failed_map[n] = j;
++n;
}
}
/* if nothing to fix, we just don't try */
/* if nothing unsynced we also don't retry, because it's the same try as before */
if (something_to_recover && something_unsynced) {
ret = repair_step(state, rehash, pos, diskmax, failed, failed_map, n, buffer, buffer_recov, buffer_zero);
if (ret == 0) {
/* reprocess the REP and CHG blocks, for which we have recovered and old state */
/* that we don't want to save into disk */
/* we have already marked them, but we redo it for logging */
for (j = 0; j < failed_count; ++j) {
/* we take care only of BAD blocks we have to write back */
if (failed[j].is_bad) {
unsigned block_state = block_state_get(failed[j].block);
if (block_state == BLOCK_STATE_CHG
|| block_state == BLOCK_STATE_REP
) {
/* mark that we have restored an old state */
/* and we don't want to write it to the disk */
failed[j].is_outofdate = 1;
log_tag("hash_unknown: Surely old data on entry %u\n", j);
}
}
}
return 0;
}
if (ret > 0)
error += ret;
if (ret < 0)
log_tag("recover_unsync:%u:%u: Failed with no attempts\n", pos, n);
else
log_tag("recover_unsync:%u:%u: Failed with %d attempts\n", pos, n, ret);
} else {
log_tag("recover_unsync:%u:%u: Skipped for%s%s\n", pos, n,
!something_to_recover ? " nothing to recover" : "",
!something_unsynced ? " nothing unsynced" : ""
);
}
/* return the number of failed attempts, or -1 if no strategy */
if (error)
return error;
else
return -1;
}
/**
* Post process all the files at the specified block index ::i.
* For each file, if we are at the last block, closes it,
* adjust the timestamp, and print the result.
*
* This works only if the whole file is processed, including its last block.
* This doesn't always happen, like with an explicit end block.
*
* In such case, the check/fix command won't report any information of the
* files partially checked.
*/
static int file_post(struct snapraid_state* state, int fix, unsigned i, struct snapraid_handle* handle, unsigned diskmax)
{
unsigned j;
int ret;
char esc_buffer[ESC_MAX];
char esc_buffer_alt[ESC_MAX];
/* for all the files print the final status, and does the final time fix */
/* we also ensure to close files after processing the last block */
for (j = 0; j < diskmax; ++j) {
struct snapraid_block* block;
struct snapraid_disk* disk;
struct snapraid_file* collide_file;
struct snapraid_file* file;
block_off_t file_pos;
uint64_t inode;
disk = handle[j].disk;
if (!disk) {
/* if no disk, nothing to do */
continue;
}
block = fs_par2block_find(disk, i);
if (!block_has_file(block)) {
/* if no file, nothing to do */
continue;
}
file = fs_par2file_get(disk, i, &file_pos);
/* if it isn't the last block in the file */
if (!file_block_is_last(file, file_pos)) {
/* nothing to do */
continue;
}
/* if the file is excluded, we have nothing to adjust as the file is never written */
if (file_flag_has(file, FILE_IS_EXCLUDED)
|| (state->opt.syncedonly && file_flag_has(file, FILE_IS_UNSYNCED))) {
/* nothing to do, but close the file */
goto close_and_continue;
}
/* finish the fix process if it's the last block of the files */
if (fix) {
/* mark that we finished with this file */
/* to identify later any NOT finished ones */
file_flag_set(file, FILE_IS_FINISHED);
/* if the file is damaged, meaning that a fix failed */
if (file_flag_has(file, FILE_IS_DAMAGED)) {
/* rename it to .unrecoverable */
char path[PATH_MAX];
char path_to[PATH_MAX];
pathprint(path, sizeof(path), "%s%s", disk->dir, file->sub);
pathprint(path_to, sizeof(path_to), "%s%s.unrecoverable", disk->dir, file->sub);
/* ensure to close the file before renaming */
if (handle[j].file == file) {
ret = handle_close(&handle[j]);
if (ret != 0) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Close error. %s\n", i, disk->name, esc_tag(file->sub, esc_buffer), strerror(errno));
log_fatal("DANGER! Unexpected close error in a data disk.\n");
return -1;
/* LCOV_EXCL_STOP */
}
}
ret = rename(path, path_to);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("Error renaming '%s' to '%s'. %s.\n", path, path_to, strerror(errno));
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
return -1;
/* LCOV_EXCL_STOP */
}
log_tag("status:unrecoverable:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("unrecoverable %s\n", fmt_term(disk, file->sub, esc_buffer));
/* and do not set the time if damaged */
goto close_and_continue;
}
/* if the file is not fixed, meaning that it is untouched */
if (!file_flag_has(file, FILE_IS_FIXED)) {
/* nothing to do, but close the file */
goto close_and_continue;
}
/* if the file is closed or different than the one expected, reopen it */
/* a different open file could happen when filtering for bad blocks */
if (handle[j].file != file) {
/* close a potential different file */
ret = handle_close(&handle[j]);
if (ret != 0) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Close error. %s\n", i, disk->name, esc_tag(handle[j].file->sub, esc_buffer), strerror(errno));
log_fatal("DANGER! Unexpected close error in a data disk.\n");
return -1;
/* LCOV_EXCL_STOP */
}
/* reopen it as readonly, as to set the mtime readonly access it's enough */
/* we know that the file exists because it has the FILE_IS_FIXED tag */
ret = handle_open(&handle[j], file, state->file_mode, log_error, 0);
if (ret != 0) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Open error. %s\n", i, disk->name, esc_tag(file->sub, esc_buffer), strerror(errno));
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
return -1;
/* LCOV_EXCL_STOP */
}
}
log_tag("status:recovered:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("recovered %s\n", fmt_term(disk, file->sub, esc_buffer));
inode = handle[j].st.st_ino;
/* search for the corresponding inode */
collide_file = tommy_hashdyn_search(&disk->inodeset, file_inode_compare_to_arg, &inode, file_inode_hash(inode));
/* if the inode is already in the database and it refers at a different file name, */
/* we can fix the file time ONLY if the time and size allow to differentiate */
/* between the two files */
/* for example, suppose we delete a bunch of files with all the same size and time, */
/* when recreating them the inodes may be reused in a different order, */
/* and at the next sync some files may have matching inode/size/time even if different name */
/* not allowing sync to detect that the file is changed and not renamed */
if (!collide_file /* if not in the database, there is no collision */
|| strcmp(collide_file->sub, file->sub) == 0 /* if the name is the same, it's the right collision */
|| collide_file->size != file->size /* if the size is different, the collision is identified */
|| collide_file->mtime_sec != file->mtime_sec /* if the mtime is different, the collision is identified */
|| collide_file->mtime_nsec != file->mtime_nsec /* same for mtime_nsec */
) {
/* set the original modification time */
ret = handle_utime(&handle[j]);
if (ret == -1) {
/* LCOV_EXCL_START */
/* mark the file as damaged */
file_flag_set(file, FILE_IS_DAMAGED);
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
return -1;
/* LCOV_EXCL_STOP */
}
} else {
log_tag("collision:%s:%s:%s: Not setting modification time to avoid inode collision\n", disk->name, esc_tag(file->sub, esc_buffer), esc_tag(collide_file->sub, esc_buffer_alt));
}
} else {
/* we are not fixing, but only checking */
/* print just the final status */
if (file_flag_has(file, FILE_IS_DAMAGED)) {
if (state->opt.auditonly) {
log_tag("status:damaged:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("damaged %s\n", fmt_term(disk, file->sub, esc_buffer));
} else {
log_tag("status:unrecoverable:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("unrecoverable %s\n", fmt_term(disk, file->sub, esc_buffer));
}
} else if (file_flag_has(file, FILE_IS_FIXED)) {
log_tag("status:recoverable:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("recoverable %s\n", fmt_term(disk, file->sub, esc_buffer));
} else {
/* we don't use msg_verbose() because it also goes into the log */
if (msg_level >= MSG_VERBOSE) {
log_tag("status:correct:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("correct %s\n", fmt_term(disk, file->sub, esc_buffer));
}
}
}
close_and_continue:
/* if the opened file is the correct one, close it */
/* in case of excluded and fragmented files it's possible */
/* that the opened file is not the current one */
if (handle[j].file == file) {
/* ensure to close the file just after finishing with it */
/* to avoid to keep it open without any possible use */
ret = handle_close(&handle[j]);
if (ret != 0) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Close error. %s\n", i, disk->name, esc_tag(file->sub, esc_buffer), strerror(errno));
log_fatal("DANGER! Unexpected close error in a data disk.\n");
return -1;
/* LCOV_EXCL_STOP */
}
}
}
return 0;
}
/**
* Check if we have to process the specified block index ::i.
*/
static int block_is_enabled(struct snapraid_state* state, block_off_t i, struct snapraid_handle* handle, unsigned diskmax)
{
unsigned j;
unsigned l;
/* filter for bad blocks */
if (state->opt.badblockonly) {
snapraid_info info;
/* get block specific info */
info = info_get(&state->infoarr, i);
/*
* Filter specifically only for bad blocks
*/
return info_get_bad(info);
}
/* filter for the parity */
if (state->opt.badfileonly) {
snapraid_info info;
/* get block specific info */
info = info_get(&state->infoarr, i);
/*
* If the block is bad, it has to be processed
*
* This is not necessary in normal cases because if a block is bad,
* it necessary needs to have a file related to it, and files with
* bad blocks are fully included.
*
* But some files may be excluded by additional filter options,
* so it's not always true, and this ensures to always check all
* the bad blocks.
*/
if (info_get_bad(info))
return 1;
} else {
/* if a parity is not excluded, include all blocks, even unused ones */
for (l = 0; l < state->level; ++l) {
if (!state->parity[l].is_excluded_by_filter) {
return 1;
}
}
}
/* filter for the files */
for (j = 0; j < diskmax; ++j) {
struct snapraid_block* block;
/* if no disk, nothing to check */
if (!handle[j].disk)
continue;
block = fs_par2block_find(handle[j].disk, i);
/* try to recover all files, even the ones without hash */
/* because in some cases we can recover also them */
if (block_has_file(block)) {
struct snapraid_file* file = fs_par2file_get(handle[j].disk, i, 0);
if (!file_flag_has(file, FILE_IS_EXCLUDED)) { /* only if the file is not filtered out */
return 1;
}
}
}
return 0;
}
static int state_check_process(struct snapraid_state* state, int fix, struct snapraid_parity_handle** parity, block_off_t blockstart, block_off_t blockmax)
{
struct snapraid_handle* handle;
unsigned diskmax;
block_off_t i;
unsigned j;
void* buffer_alloc;
void** buffer;
unsigned buffermax;
int ret;
data_off_t countsize;
block_off_t countpos;
block_off_t countmax;
unsigned error;
unsigned unrecoverable_error;
unsigned recovered_error;
struct failed_struct* failed;
unsigned* failed_map;
unsigned l;
char esc_buffer[ESC_MAX];
char esc_buffer_alt[ESC_MAX];
bit_vect_t* block_enabled;
struct snapraid_bw bw;
handle = handle_mapping(state, &diskmax);
/* initialize the bandwith context */
bw_init(&bw, state->opt.bwlimit);
/* share the bandwidth context with all handles */
for (j = 0; j < diskmax; ++j)
handle[j].bw = &bw;
for (j = 0; j < state->level; ++j)
if (parity[j])
parity[j]->bw = &bw;
/* we need 1 * data + 2 * parity + 1 * zero */
buffermax = diskmax + 2 * state->level + 1;
buffer = malloc_nofail_vector_align(diskmax, buffermax, state->block_size, &buffer_alloc);
if (!state->opt.skip_self)
mtest_vector(buffermax, state->block_size, buffer);
/* fill up the zero buffer */
memset(buffer[buffermax - 1], 0, state->block_size);
raid_zero(buffer[buffermax - 1]);
failed = malloc_nofail(diskmax * sizeof(struct failed_struct));
failed_map = malloc_nofail(diskmax * sizeof(unsigned));
error = 0;
unrecoverable_error = 0;
recovered_error = 0;
msg_progress("Selecting...\n");
/* first count the number of blocks to process */
countmax = 0;
block_enabled = calloc_nofail(1, bit_vect_size(blockmax)); /* preinitialize to 0 */
for (i = blockstart; i < blockmax; ++i) {
if (!block_is_enabled(state, i, handle, diskmax))
continue;
bit_vect_set(block_enabled, i);
++countmax;
}
if (fix)
msg_progress("Fixing...\n");
else if (!state->opt.auditonly)
msg_progress("Checking...\n");
else
msg_progress("Hashing...\n");
/* check all the blocks in files */
countsize = 0;
countpos = 0;
if (!state_progress_begin(state, blockstart, blockmax, countmax))
goto end;
for (i = blockstart; i < blockmax; ++i) {
unsigned failed_count;
int valid_parity;
int used_parity;
snapraid_info info;
int rehash;
if (!bit_vect_test(block_enabled, i)) {
/* continue with the next block */
continue;
}
/* If we have valid parity, and it makes sense to check its content. */
/* If we already know that the parity is invalid, we just read the file */
/* but we don't report parity errors */
/* Note that with auditonly, we anyway skip the full parity check, */
/* because we also don't read it at all */
valid_parity = 1;
/* If the parity is used by at least one file */
used_parity = 0;
/* keep track of the number of failed blocks */
failed_count = 0;
/* get block specific info */
info = info_get(&state->infoarr, i);
/* if we have to use the old hash */
rehash = info_get_rehash(info);
/* for each disk, process the block */
for (j = 0; j < diskmax; ++j) {
int read_size;
unsigned char hash[HASH_MAX];
struct snapraid_disk* disk;
struct snapraid_block* block;
struct snapraid_file* file;
block_off_t file_pos;
unsigned block_state;
/* if the disk position is not used */
disk = handle[j].disk;
if (!disk) {
/* use an empty block */
memset(buffer[j], 0, state->block_size);
continue;
}
/* if the disk block is not used */
block = fs_par2block_find(disk, i);
if (block == BLOCK_NULL) {
/* use an empty block */
memset(buffer[j], 0, state->block_size);
continue;
}
/* get the state of the block */
block_state = block_state_get(block);
/* if the parity is not valid */
if (block_has_invalid_parity(block)) {
/* mark the parity as invalid, and don't try to check/fix it */
/* because it will be recomputed at the next sync */
valid_parity = 0;
/* follow */
}
/* if the block is DELETED */
if (block_state == BLOCK_STATE_DELETED) {
/* use an empty block */
memset(buffer[j], 0, state->block_size);
/* store it in the failed set, because potentially */
/* the parity may be still computed with the previous content */
failed[failed_count].is_bad = 0; /* note that is_bad==0 <=> file==0 */
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = 0;
failed[failed_count].file_pos = 0;
failed[failed_count].handle = 0;
++failed_count;
continue;
}
/* here we are sure that the parity is used by a file */
used_parity = 1;
/* get the file of this block */
file = fs_par2file_get(disk, i, &file_pos);
/* if we are only hashing, we can skip excluded files and don't even read them */
if (state->opt.auditonly && file_flag_has(file, FILE_IS_EXCLUDED)) {
/* use an empty block */
/* in true, this is unnecessary, because we are not checking any parity */
/* but we keep it for completeness */
memset(buffer[j], 0, state->block_size);
continue;
}
/* if the file is closed or different than the current one */
if (handle[j].file == 0 || handle[j].file != file) {
/* close the old one, if any */
ret = handle_close(&handle[j]);
if (ret == -1) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Close error. %s\n", i, disk->name, esc_tag(handle[j].file->sub, esc_buffer), strerror(errno));
log_fatal("DANGER! Unexpected close error in a data disk.\n");
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* if fixing, and the file is not excluded, we must open for writing */
if (fix && !file_flag_has(file, FILE_IS_EXCLUDED)) {
/* if fixing, create the file, open for writing and resize if required */
ret = handle_create(&handle[j], file, state->file_mode);
if (ret == -1) {
/* LCOV_EXCL_START */
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the file.\n");
} else {
log_fatal("DANGER! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* check if the file was just created */
if (handle[j].created != 0) {
/* if fragmented, it may be reopened, so remember that the file */
/* was originally missing */
file_flag_set(file, FILE_IS_CREATED);
}
} else {
/* open the file only for reading */
if (!file_flag_has(file, FILE_IS_MISSING))
ret = handle_open(&handle[j], file, state->file_mode,
log_error, state->opt.expected_missing ? log_expected : 0);
else
ret = -1; /* if the file is missing, we cannot open it */
if (ret == -1) {
/* save the failed block for the check/fix */
failed[failed_count].is_bad = 1;
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = file;
failed[failed_count].file_pos = file_pos;
failed[failed_count].handle = &handle[j];
++failed_count;
log_tag("error:%u:%s:%s: Open error at position %u\n", i, disk->name, esc_tag(file->sub, esc_buffer), file_pos);
++error;
/* mark the file as missing, to avoid to retry to open it again */
/* note that this can be done only if we are not fixing it */
/* otherwise, it could be recreated */
file_flag_set(file, FILE_IS_MISSING);
continue;
}
}
/* if it's the first open, and not excluded */
if (!file_flag_has(file, FILE_IS_OPENED)
&& !file_flag_has(file, FILE_IS_EXCLUDED)) {
/* check if the file is changed */
if (handle[j].st.st_size != file->size
|| handle[j].st.st_mtime != file->mtime_sec
|| STAT_NSEC(&handle[j].st) != file->mtime_nsec
/* don't check the inode to support file-system without persistent inodes */
) {
/* report that the file is not synced */
file_flag_set(file, FILE_IS_UNSYNCED);
}
}
/* if it's the first open, and not excluded and larger */
if (!file_flag_has(file, FILE_IS_OPENED)
&& !file_flag_has(file, FILE_IS_EXCLUDED)
&& !(state->opt.syncedonly && file_flag_has(file, FILE_IS_UNSYNCED))
&& handle[j].st.st_size > file->size
) {
log_error("File '%s' is larger than expected.\n", handle[j].path);
log_tag("error:%u:%s:%s: Size error\n", i, disk->name, esc_tag(file->sub, esc_buffer));
++error;
if (fix) {
ret = handle_truncate(&handle[j], file);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("DANGER! Unexpected truncate error in a data disk, it isn't possible to fix.\n");
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("fixed:%u:%s:%s: Fixed size\n", i, disk->name, esc_tag(file->sub, esc_buffer));
++recovered_error;
}
}
/* mark the file as opened at least one time */
/* this is used to avoid to check the unsynced and size */
/* more than one time, in case the file is reopened later */
file_flag_set(file, FILE_IS_OPENED);
}
/* read from the file */
read_size = handle_read(&handle[j], file_pos, buffer[j], state->block_size,
log_error, state->opt.expected_missing ? log_expected : 0);
if (read_size == -1) {
/* save the failed block for the check/fix */
failed[failed_count].is_bad = 1; /* it's bad because we cannot read it */
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = file;
failed[failed_count].file_pos = file_pos;
failed[failed_count].handle = &handle[j];
++failed_count;
log_tag("error:%u:%s:%s: Read error at position %u\n", i, disk->name, esc_tag(file->sub, esc_buffer), file_pos);
++error;
continue;
}
countsize += read_size;
/* always insert CHG blocks, the repair functions needs all of them */
/* because the parity may be still referring at the old state */
/* and the repair must be aware of it */
if (block_state == BLOCK_STATE_CHG) {
/* we DO NOT mark them as bad to avoid to overwrite them with wrong data. */
/* if we don't have a hash, we always assume the first read of the block correct. */
failed[failed_count].is_bad = 0; /* we assume the CHG block correct */
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = file;
failed[failed_count].file_pos = file_pos;
failed[failed_count].handle = &handle[j];
++failed_count;
continue;
}
assert(block_state == BLOCK_STATE_BLK || block_state == BLOCK_STATE_REP);
/* compute the hash of the block just read */
if (rehash) {
memhash(state->prevhash, state->prevhashseed, hash, buffer[j], read_size);
} else {
memhash(state->hash, state->hashseed, hash, buffer[j], read_size);
}
/* compare the hash */
if (memcmp(hash, block->hash, BLOCK_HASH_SIZE) != 0) {
unsigned diff = memdiff(hash, block->hash, BLOCK_HASH_SIZE);
/* save the failed block for the check/fix */
failed[failed_count].is_bad = 1; /* it's bad because the hash doesn't match */
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = file;
failed[failed_count].file_pos = file_pos;
failed[failed_count].handle = &handle[j];
++failed_count;
log_tag("error:%u:%s:%s: Data error at position %u, diff bits %u/%u\n", i, disk->name, esc_tag(file->sub, esc_buffer), file_pos, diff, BLOCK_HASH_SIZE * 8);
++error;
continue;
}
/* always insert REP blocks, the repair functions needs all of them */
/* because the parity may be still referring at the old state */
/* and the repair must be aware of it */
if (block_state == BLOCK_STATE_REP) {
failed[failed_count].is_bad = 0; /* it's not bad */
failed[failed_count].is_outofdate = 0;
failed[failed_count].index = j;
failed[failed_count].block = block;
failed[failed_count].disk = disk;
failed[failed_count].file = file;
failed[failed_count].file_pos = file_pos;
failed[failed_count].handle = &handle[j];
++failed_count;
continue;
}
}
/* now read and check the parity if requested */
if (!state->opt.auditonly) {
void* buffer_recov[LEV_MAX];
void* buffer_zero;
/* buffers for parity read and not computed */
for (l = 0; l < state->level; ++l)
buffer_recov[l] = buffer[diskmax + state->level + l];
for (; l < LEV_MAX; ++l)
buffer_recov[l] = 0;
/* the zero buffer is the last one */
buffer_zero = buffer[buffermax - 1];
/* read the parity */
for (l = 0; l < state->level; ++l) {
if (parity[l]) {
ret = parity_read(parity[l], i, buffer_recov[l], state->block_size, log_error);
if (ret == -1) {
buffer_recov[l] = 0; /* no parity to use */
log_tag("parity_error:%u:%s: Read error\n", i, lev_config_name(l));
++error;
}
} else {
buffer_recov[l] = 0;
}
}
/* try all the recovering strategies */
ret = repair(state, rehash, i, diskmax, failed, failed_map, failed_count, buffer, buffer_recov, buffer_zero);
if (ret != 0) {
/* increment the number of errors */
if (ret > 0)
error += ret;
++unrecoverable_error;
/* print a list of all the errors in files */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad)
log_tag("unrecoverable:%u:%s:%s: Unrecoverable error at position %u\n", i, failed[j].disk->name, esc_tag(failed[j].file->sub, esc_buffer), failed[j].file_pos);
}
/* keep track of damaged files */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad)
file_flag_set(failed[j].file, FILE_IS_DAMAGED);
}
} else {
/* now counts partial recovers */
/* note that this could happen only when we have an incomplete 'sync' */
/* and that we have recovered is the state before the 'sync' */
int partial_recover_error = 0;
/* print a list of all the errors in files */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad && failed[j].is_outofdate) {
++partial_recover_error;
log_tag("unrecoverable:%u:%s:%s: Unrecoverable unsynced error at position %u\n", i, failed[j].disk->name, esc_tag(failed[j].file->sub, esc_buffer), failed[j].file_pos);
}
}
if (partial_recover_error != 0) {
error += partial_recover_error;
++unrecoverable_error;
}
/*
* Check parities, but only if all the blocks have it computed and it's used.
*
* If you check/fix after a partial sync, it's OK to have parity errors
* on the blocks with invalid parity and doesn't make sense to try to fix it.
*
* It's also OK to have data errors on unused parity, because sync doesn't
* update it.
*/
if (used_parity && valid_parity) {
/* check the parity */
for (l = 0; l < state->level; ++l) {
if (buffer_recov[l] != 0 && memcmp(buffer_recov[l], buffer[diskmax + l], state->block_size) != 0) {
unsigned diff = memdiff(buffer_recov[l], buffer[diskmax + l], state->block_size);
/* mark that the read parity is wrong, setting ptr to 0 */
buffer_recov[l] = 0;
log_tag("parity_error:%u:%s: Data error, diff bits %u/%u\n", i, lev_config_name(l), diff, state->block_size * 8);
++error;
}
}
}
/* now write recovered files */
if (fix) {
/* update the fixed files */
for (j = 0; j < failed_count; ++j) {
/* nothing to do if it doesn't need recovering */
if (!failed[j].is_bad)
continue;
/* do not fix if the file is excluded */
if (file_flag_has(failed[j].file, FILE_IS_EXCLUDED)
|| (state->opt.syncedonly && file_flag_has(failed[j].file, FILE_IS_UNSYNCED)))
continue;
ret = handle_write(failed[j].handle, failed[j].file_pos, buffer[failed[j].index], state->block_size);
if (ret == -1) {
/* LCOV_EXCL_START */
/* mark the file as damaged */
file_flag_set(failed[j].file, FILE_IS_DAMAGED);
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the file.\n");
} else {
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* if we are not sure that the recovered content is uptodate */
if (failed[j].is_outofdate) {
/* mark the file as damaged */
file_flag_set(failed[j].file, FILE_IS_DAMAGED);
continue;
}
/* mark the file as containing some fixes */
/* note that it could be also marked as damaged in other iterations */
file_flag_set(failed[j].file, FILE_IS_FIXED);
log_tag("fixed:%u:%s:%s: Fixed data error at position %u\n", i, failed[j].disk->name, esc_tag(failed[j].file->sub, esc_buffer), failed[j].file_pos);
++recovered_error;
}
/*
* Update parity only if all the blocks have it computed and it's used.
*
* If you check/fix after a partial sync, you do not want to fix parity
* for blocks that are going to have it computed in the sync completion.
*
* For unused parity there is no need to write it, because when fixing
* we already have allocated space for it on parity file creation,
* and its content doesn't matter.
*/
if (used_parity && valid_parity) {
/* update the parity */
for (l = 0; l < state->level; ++l) {
/* if the parity on disk is wrong */
if (buffer_recov[l] == 0
/* and we have access at the parity */
&& parity[l] != 0
/* and the parity is not excluded */
&& !state->parity[l].is_excluded_by_filter
) {
ret = parity_write(parity[l], i, buffer[diskmax + l], state->block_size);
if (ret == -1) {
/* LCOV_EXCL_START */
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working %s disk, it isn't possible to fix errors on it.\n", lev_name(l));
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("parity_fixed:%u:%s: Fixed data error\n", i, lev_config_name(l));
++recovered_error;
}
}
}
} else {
/* if we are not fixing, we just set the FIXED flag */
/* meaning that we could fix this file if we try */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad) {
file_flag_set(failed[j].file, FILE_IS_FIXED);
}
}
}
}
} else {
/* if we are not checking, we just set the DAMAGED flag */
/* to report that the file is damaged, and we don't know if we can fix it */
for (j = 0; j < failed_count; ++j) {
if (failed[j].is_bad) {
file_flag_set(failed[j].file, FILE_IS_DAMAGED);
}
}
}
/* post process the files */
ret = file_post(state, fix, i, handle, diskmax);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("Stopping at block %u\n", i);
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* count the number of processed block */
++countpos;
/* progress */
if (state_progress(state, 0, i, countpos, countmax, countsize)) {
/* LCOV_EXCL_START */
break;
/* LCOV_EXCL_STOP */
}
/* thermal control */
if (state_thermal_alarm(state)) {
/* until now is misc */
state_usage_misc(state);
state_progress_stop(state);
state_thermal_cooldown(state);
state_progress_restart(state);
/* drop until now */
state_usage_waste(state);
}
}
/* for each disk, recover empty files, symlinks and empty dirs */
for (i = 0; i < diskmax; ++i) {
tommy_node* node;
struct snapraid_disk* disk;
if (!handle[i].disk)
continue;
/* for each empty file in the disk */
disk = handle[i].disk;
node = disk->filelist;
while (node) {
char path[PATH_MAX];
struct stat st;
struct snapraid_file* file;
int unsuccessful = 0;
file = node->data;
node = node->next; /* next node */
/* if not empty, it's already checked and continue to the next one */
if (file->size != 0) {
continue;
}
/* if excluded continue to the next one */
if (file_flag_has(file, FILE_IS_EXCLUDED)) {
continue;
}
/* stat the file */
pathprint(path, sizeof(path), "%s%s", disk->dir, file->sub);
ret = stat(path, &st);
if (ret == -1) {
unsuccessful = 1;
log_error("Error stating empty file '%s'. %s.\n", path, strerror(errno));
log_tag("error:%s:%s: Empty file stat error\n", disk->name, esc_tag(file->sub, esc_buffer));
++error;
} else if (!S_ISREG(st.st_mode)) {
unsuccessful = 1;
log_tag("error:%s:%s: Empty file error for not regular file\n", disk->name, esc_tag(file->sub, esc_buffer));
++error;
} else if (st.st_size != 0) {
unsuccessful = 1;
log_tag("error:%s:%s: Empty file error for size '%" PRIu64 "'\n", disk->name, esc_tag(file->sub, esc_buffer), (uint64_t)st.st_size);
++error;
}
if (fix && unsuccessful) {
int f;
/* create the ancestor directories */
ret = mkancestor(path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* create it */
/* O_NOFOLLOW: do not follow links to ensure to open the real file */
f = open(path, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY | O_NOFOLLOW, 0600);
if (f == -1) {
/* LCOV_EXCL_START */
log_fatal("Error creating empty file '%s'. %s.\n", path, strerror(errno));
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the file.\n");
} else {
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* set the original modification time */
ret = fmtime(f, file->mtime_sec, file->mtime_nsec);
if (ret != 0) {
/* LCOV_EXCL_START */
close(f);
log_fatal("Error timing file '%s'. %s.\n", file->sub, strerror(errno));
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* close it */
ret = close(f);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("fixed:%s:%s: Fixed empty file\n", disk->name, esc_tag(file->sub, esc_buffer));
++recovered_error;
log_tag("status:recovered:%s:%s\n", disk->name, esc_tag(file->sub, esc_buffer));
msg_info("recovered %s\n", fmt_term(disk, file->sub, esc_buffer));
}
}
/* for each link in the disk */
disk = handle[i].disk;
node = disk->linklist;
while (node) {
char path[PATH_MAX];
char pathto[PATH_MAX];
char linkto[PATH_MAX];
struct stat st;
struct stat stto;
struct snapraid_link* slink;
int unsuccessful = 0;
int unrecoverable = 0;
slink = node->data;
node = node->next; /* next node */
/* if excluded continue to the next one */
if (link_flag_has(slink, FILE_IS_EXCLUDED)) {
continue;
}
if (link_flag_has(slink, FILE_IS_HARDLINK)) {
/* stat the link */
pathprint(path, sizeof(path), "%s%s", disk->dir, slink->sub);
ret = stat(path, &st);
if (ret == -1) {
unsuccessful = 1;
log_error("Error stating hardlink '%s'. %s.\n", path, strerror(errno));
log_tag("hardlink_error:%s:%s:%s: Hardlink stat error\n", disk->name, esc_tag(slink->sub, esc_buffer), esc_tag(slink->linkto, esc_buffer_alt));
++error;
} else if (!S_ISREG(st.st_mode)) {
unsuccessful = 1;
log_tag("hardlink_error:%s:%s:%s: Hardlink error for not regular file\n", disk->name, esc_tag(slink->sub, esc_buffer), esc_tag(slink->linkto, esc_buffer_alt));
++error;
}
/* stat the "to" file */
pathprint(pathto, sizeof(pathto), "%s%s", disk->dir, slink->linkto);
ret = stat(pathto, &stto);
if (ret == -1) {
unsuccessful = 1;
if (errno == ENOENT) {
unrecoverable = 1;
if (fix) {
/* if the target doesn't exist, it's unrecoverable */
/* because we cannot create an hardlink of a file that */
/* doesn't exists */
++unrecoverable_error;
} else {
/* but in check, we can assume that fixing will recover */
/* such missing file, so we assume a less drastic error */
++error;
}
}
log_error("Error stating hardlink-to '%s'. %s.\n", pathto, strerror(errno));
log_tag("hardlink_error:%s:%s:%s: Hardlink to stat error\n", disk->name, esc_tag(slink->sub, esc_buffer), esc_tag(slink->linkto, esc_buffer_alt));
++error;
} else if (!S_ISREG(stto.st_mode)) {
unsuccessful = 1;
log_tag("hardlink_error:%s:%s:%s: Hardlink-to error for not regular file\n", disk->name, esc_tag(slink->sub, esc_buffer), esc_tag(slink->linkto, esc_buffer_alt));
++error;
} else if (!unsuccessful && st.st_ino != stto.st_ino) {
unsuccessful = 1;
log_error("Mismatch hardlink '%s' and '%s'. Different inode.\n", path, pathto);
log_tag("hardlink_error:%s:%s:%s: Hardlink mismatch for different inode\n", disk->name, esc_tag(slink->sub, esc_buffer), esc_tag(slink->linkto, esc_buffer_alt));
++error;
}
} else {
/* read the symlink */
pathprint(path, sizeof(path), "%s%s", disk->dir, slink->sub);
ret = readlink(path, linkto, sizeof(linkto));
if (ret < 0) {
unsuccessful = 1;
log_error("Error reading symlink '%s'. %s.\n", path, strerror(errno));
log_tag("symlink_error:%s:%s: Symlink read error\n", disk->name, esc_tag(slink->sub, esc_buffer));
++error;
} else if (ret >= PATH_MAX) {
unsuccessful = 1;
log_error("Error reading symlink '%s'. Symlink too long.\n", path);
log_tag("symlink_error:%s:%s: Symlink read error\n", disk->name, esc_tag(slink->sub, esc_buffer));
++error;
} else {
linkto[ret] = 0;
if (strcmp(linkto, slink->linkto) != 0) {
unsuccessful = 1;
log_tag("symlink_error:%s:%s: Symlink data error '%s' instead of '%s'\n", disk->name, esc_tag(slink->sub, esc_buffer), linkto, slink->linkto);
++error;
}
}
}
if (fix && unsuccessful && !unrecoverable) {
/* create the ancestor directories */
ret = mkancestor(path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* if it exists, it must be deleted before recreating */
ret = remove(path);
if (ret != 0 && errno != ENOENT) {
/* LCOV_EXCL_START */
log_fatal("Error removing '%s'. %s.\n", path, strerror(errno));
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* create it */
if (link_flag_has(slink, FILE_IS_HARDLINK)) {
ret = hardlink(pathto, path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("Error writing hardlink '%s' to '%s'. %s.\n", path, pathto, strerror(errno));
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the hardlink.\n");
} else {
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("hardlink_fixed:%s:%s: Fixed hardlink error\n", disk->name, esc_tag(slink->sub, esc_buffer));
++recovered_error;
} else {
ret = symlink(slink->linkto, path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("Error writing symlink '%s' to '%s'. %s.\n", path, slink->linkto, strerror(errno));
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the symlink.\n");
} else {
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("symlink_fixed:%s:%s: Fixed symlink error\n", disk->name, esc_tag(slink->sub, esc_buffer));
++recovered_error;
}
log_tag("status:recovered:%s:%s\n", disk->name, esc_tag(slink->sub, esc_buffer));
msg_info("recovered %s\n", fmt_term(disk, slink->sub, esc_buffer));
}
}
/* for each dir in the disk */
disk = handle[i].disk;
node = disk->dirlist;
while (node) {
char path[PATH_MAX];
struct stat st;
struct snapraid_dir* dir;
int unsuccessful = 0;
dir = node->data;
node = node->next; /* next node */
/* if excluded continue to the next one */
if (dir_flag_has(dir, FILE_IS_EXCLUDED)) {
continue;
}
/* stat the dir */
pathprint(path, sizeof(path), "%s%s", disk->dir, dir->sub);
ret = stat(path, &st);
if (ret == -1) {
unsuccessful = 1;
log_error("Error stating dir '%s'. %s.\n", path, strerror(errno));
log_tag("dir_error:%s:%s: Dir stat error\n", disk->name, esc_tag(dir->sub, esc_buffer));
++error;
} else if (!S_ISDIR(st.st_mode)) {
unsuccessful = 1;
log_tag("dir_error:%s:%s: Dir error for not directory\n", disk->name, esc_tag(dir->sub, esc_buffer));
++error;
}
if (fix && unsuccessful) {
/* create the ancestor directories */
ret = mkancestor(path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
/* create it */
ret = mkdir(path, S_IRWXU | S_IRGRP | S_IXGRP | S_IROTH | S_IXOTH);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("Error creating dir '%s'. %s.\n", path, strerror(errno));
if (errno == EACCES) {
log_fatal("WARNING! Please give write permission to the dir.\n");
} else {
/* we do not use DANGER because it could be ENOSPC which is not always correctly reported */
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
}
log_fatal("Stopping\n");
++unrecoverable_error;
goto bail;
/* LCOV_EXCL_STOP */
}
log_tag("dir_fixed:%s:%s: Fixed dir error\n", disk->name, esc_tag(dir->sub, esc_buffer));
++recovered_error;
log_tag("status:recovered:%s:%s\n", disk->name, esc_tag(dir->sub, esc_buffer));
msg_info("recovered %s\n", fmt_term(disk, dir->sub, esc_buffer));
}
}
}
end:
state_progress_end(state, countpos, countmax, countsize, "Nothing to check.\n");
bail:
/* close all the files left open */
for (j = 0; j < diskmax; ++j) {
struct snapraid_file* file = handle[j].file;
struct snapraid_disk* disk = handle[j].disk;
ret = handle_close(&handle[j]);
if (ret == -1) {
/* LCOV_EXCL_START */
log_tag("error:%u:%s:%s: Close error. %s\n", blockmax, disk->name, esc_tag(file->sub, esc_buffer), strerror(errno));
log_fatal("DANGER! Unexpected close error in a data disk.\n");
++unrecoverable_error;
/* continue, as we are already exiting */
/* LCOV_EXCL_STOP */
}
}
/* remove all the files created from scratch that have not finished the processing */
/* it happens only when aborting pressing Ctrl+C or other reason. */
if (fix) {
/* for each disk */
for (i = 0; i < diskmax; ++i) {
tommy_node* node;
struct snapraid_disk* disk;
if (!handle[i].disk)
continue;
/* for each file in the disk */
disk = handle[i].disk;
node = disk->filelist;
while (node) {
char path[PATH_MAX];
struct snapraid_file* file;
file = node->data;
node = node->next; /* next node */
/* if the file was not created, meaning that it was already existing */
if (!file_flag_has(file, FILE_IS_CREATED)) {
/* nothing to do */
continue;
}
/* if processing was finished */
if (file_flag_has(file, FILE_IS_FINISHED)) {
/* nothing to do */
continue;
}
/* if the file was originally missing, and processing not yet finished */
/* we have to throw it away to ensure that at the next run we will retry */
/* to fix it, in case we select to undelete missing files */
pathprint(path, sizeof(path), "%s%s", disk->dir, file->sub);
ret = remove(path);
if (ret != 0) {
/* LCOV_EXCL_START */
log_fatal("Error removing '%s'. %s.\n", path, strerror(errno));
log_fatal("WARNING! Without a working data disk, it isn't possible to fix errors on it.\n");
++unrecoverable_error;
/* continue, as we are already exiting */
/* LCOV_EXCL_STOP */
}
}
}
}
if (error || recovered_error || unrecoverable_error) {
msg_status("\n");
msg_status("%8u errors\n", error);
if (fix) {
msg_status("%8u recovered errors\n", recovered_error);
}
if (unrecoverable_error) {
msg_status("%8u UNRECOVERABLE errors\n", unrecoverable_error);
} else {
/* without checking, we don't know if they are really recoverable or not */
if (!state->opt.auditonly)
msg_status("%8u unrecoverable errors\n", unrecoverable_error);
if (fix)
msg_status("Everything OK\n");
}
} else {
msg_status("Everything OK\n");
}
if (error && !fix)
log_fatal("WARNING! There are errors!\n");
if (unrecoverable_error)
log_fatal("DANGER! Unrecoverable errors detected!\n");
log_tag("summary:error:%u\n", error);
if (fix)
log_tag("summary:error_recovered:%u\n", recovered_error);
if (!state->opt.auditonly)
log_tag("summary:error_unrecoverable:%u\n", unrecoverable_error);
if (fix) {
if (error + recovered_error + unrecoverable_error == 0)
log_tag("summary:exit:ok\n");
else if (unrecoverable_error == 0)
log_tag("summary:exit:recovered\n");
else
log_tag("summary:exit:unrecoverable\n");
} else if (!state->opt.auditonly) {
if (error + unrecoverable_error == 0)
log_tag("summary:exit:ok\n");
else if (unrecoverable_error == 0)
log_tag("summary:exit:recoverable\n");
else
log_tag("summary:exit:unrecoverable\n");
} else { /* audit only */
if (error == 0)
log_tag("summary:exit:ok\n");
else
log_tag("summary:exit:error\n");
}
log_flush();
free(failed);
free(failed_map);
free(block_enabled);
free(handle);
free(buffer_alloc);
free(buffer);
/* fail if some error are present after the run */
if (fix) {
if (state->opt.expect_unrecoverable) {
if (unrecoverable_error == 0)
return -1;
} else {
if (unrecoverable_error != 0)
return -1;
}
} else {
if (state->opt.expect_unrecoverable) {
if (unrecoverable_error == 0)
return -1;
} else if (state->opt.expect_recoverable) {
if (unrecoverable_error != 0 || error == 0)
return -1;
} else {
if (error != 0 || unrecoverable_error != 0)
return -1;
}
}
return 0;
}
int state_check(struct snapraid_state* state, int fix, block_off_t blockstart, block_off_t blockcount)
{
block_off_t blockmax;
data_off_t size;
int ret;
struct snapraid_parity_handle parity[LEV_MAX];
struct snapraid_parity_handle* parity_ptr[LEV_MAX];
unsigned error;
unsigned l;
msg_progress("Initializing...\n");
blockmax = parity_allocated_size(state);
size = blockmax * (data_off_t)state->block_size;
if (blockstart > blockmax) {
/* LCOV_EXCL_START */
log_fatal("Error in the specified starting block %u. It's larger than the parity size %u.\n", blockstart, blockmax);
exit(EXIT_FAILURE);
/* LCOV_EXCL_STOP */
}
/* adjust the number of block to process */
if (blockcount != 0 && blockstart + blockcount < blockmax) {
blockmax = blockstart + blockcount;
}
if (fix) {
/* if fixing, create the file and open for writing */
/* if it fails, we cannot continue */
for (l = 0; l < state->level; ++l) {
/* skip parity disks that are not accessible */
if (state->parity[l].skip_access) {
parity_ptr[l] = 0;
continue;
}
parity_ptr[l] = &parity[l];
/* if the parity is excluded */
if (state->parity[l].is_excluded_by_filter) {
/* open for reading, and ignore error */
ret = parity_open(parity_ptr[l], &state->parity[l], l, state->file_mode, state->block_size, state->opt.parity_limit_size);
if (ret == -1) {
/* continue anyway */
parity_ptr[l] = 0;
}
} else {
/* open for writing */
ret = parity_create(parity_ptr[l], &state->parity[l], l, state->file_mode, state->block_size, state->opt.parity_limit_size);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without an accessible %s file, it isn't possible to fix any error.\n", lev_name(l));
exit(EXIT_FAILURE);
/* LCOV_EXCL_STOP */
}
ret = parity_chsize(parity_ptr[l], &state->parity[l], 0, size, state->block_size, state->opt.skip_fallocate, state->opt.skip_space_holder);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("WARNING! Without an accessible %s file, it isn't possible to sync.\n", lev_name(l));
exit(EXIT_FAILURE);
/* LCOV_EXCL_STOP */
}
}
}
} else if (!state->opt.auditonly) {
/* if checking, open the file for reading */
/* it may fail if the file doesn't exist, in this case we continue to check the files */
for (l = 0; l < state->level; ++l) {
parity_ptr[l] = &parity[l];
ret = parity_open(parity_ptr[l], &state->parity[l], l, state->file_mode, state->block_size, state->opt.parity_limit_size);
if (ret == -1) {
msg_status("No accessible %s file, only files will be checked.\n", lev_name(l));
/* continue anyway */
parity_ptr[l] = 0;
}
}
} else {
/* otherwise don't use any parity */
for (l = 0; l < state->level; ++l)
parity_ptr[l] = 0;
}
error = 0;
/* skip degenerated cases of empty parity, or skipping all */
if (blockstart < blockmax) {
ret = state_check_process(state, fix, parity_ptr, blockstart, blockmax);
if (ret == -1) {
/* LCOV_EXCL_START */
++error;
/* continue, as we are already exiting */
/* LCOV_EXCL_STOP */
}
}
/* try to close only if opened */
for (l = 0; l < state->level; ++l) {
if (parity_ptr[l]) {
/* if fixing and not excluded, truncate parity not valid */
if (fix && !state->parity[l].is_excluded_by_filter) {
ret = parity_truncate(parity_ptr[l]);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("DANGER! Unexpected truncate error in %s disk.\n", lev_name(l));
++error;
/* continue, as we are already exiting */
/* LCOV_EXCL_STOP */
}
}
ret = parity_close(parity_ptr[l]);
if (ret == -1) {
/* LCOV_EXCL_START */
log_fatal("DANGER! Unexpected close error in %s disk.\n", lev_name(l));
++error;
/* continue, as we are already exiting */
/* LCOV_EXCL_STOP */
}
}
}
/* abort if error are present */
if (error != 0)
return -1;
return 0;
}
|