X-Git-Url: http://git.ipfire.org/?p=thirdparty%2Fmdadm.git;a=blobdiff_plain;f=mdadm.c;h=5afe4155d7011f2f3b653ff162dedeb940cdd911;hp=222ba6c4ba13e4c9d04c49a3623bd2e0838c8d88;hb=1b7eb962db2cf9179d097e06cce74b84ac80e49d;hpb=32754b7d8474b34fc5ab0e629e43f30e3a50d408 diff --git a/mdadm.c b/mdadm.c index 222ba6c4..5afe4155 100644 --- a/mdadm.c +++ b/mdadm.c @@ -1,7 +1,7 @@ /* * mdadm - manage Linux "md" devices aka RAID arrays. * - * Copyright (C) 2001-2012 Neil Brown + * Copyright (C) 2001-2013 Neil Brown * * * This program is free software; you can redistribute it and/or modify @@ -29,7 +29,6 @@ #include "md_p.h" #include - static int scan_assemble(struct supertype *ss, struct context *c, struct mddev_ident *ident); @@ -37,8 +36,9 @@ static int misc_scan(char devmode, struct context *c); static int stop_scan(int verbose); static int misc_list(struct mddev_dev *devlist, struct mddev_ident *ident, + char *dump_directory, struct supertype *ss, struct context *c); - +const char Name[] = "mdadm"; int main(int argc, char *argv[]) { @@ -49,6 +49,7 @@ int main(int argc, char *argv[]) int i; unsigned long long array_size = 0; + unsigned long long data_offset = INVALID_SECTORS; struct mddev_ident ident; char *configfile = NULL; int devmode = 0; @@ -56,6 +57,7 @@ int main(int argc, char *argv[]) struct mddev_dev *devlist = NULL; struct mddev_dev **devlistend = & devlist; struct mddev_dev *dv; + mdu_array_info_t array; int devs_found = 0; char *symlinks = NULL; int grow_continue = 0; @@ -73,9 +75,11 @@ int main(int argc, char *argv[]) .require_homehost = 1, }; struct shape s = { + .journaldisks = 0, .level = UnSet, .layout = UnSet, .bitmap_chunk = UnSet, + .consistency_policy = CONSISTENCY_POLICY_UNKNOWN, }; char sys_hostname[256]; @@ -87,25 +91,28 @@ int main(int argc, char *argv[]) int oneshot = 0; int spare_sharing = 1; struct supertype *ss = NULL; - int writemostly = 0; + enum flag_mode writemostly = FlagDefault; + enum flag_mode failfast = FlagDefault; char *shortopt = short_options; int dosyslog = 0; int rebuild_map = 0; char *remove_path = NULL; char *udev_filename = NULL; + char *dump_directory = NULL; int print_help = 0; FILE *outf; int mdfd = -1; + int locked = 0; srandom(time(0) ^ getpid()); - ident.uuid_set=0; + ident.uuid_set = 0; ident.level = UnSet; ident.raid_disks = UnSet; - ident.super_minor= UnSet; - ident.devices=0; + ident.super_minor = UnSet; + ident.devices = 0; ident.spare_group = NULL; ident.autof = 0; ident.st = NULL; @@ -115,10 +122,14 @@ int main(int argc, char *argv[]) ident.container = NULL; ident.member = NULL; - while ((option_index = -1) , - (opt=getopt_long(argc, argv, - shortopt, long_options, - &option_index)) != -1) { + if (get_linux_version() < 2006015) { + pr_err("This version of mdadm does not support kernels older than 2.6.15\n"); + exit(1); + } + + while ((option_index = -1), + (opt = getopt_long(argc, argv, shortopt, long_options, + &option_index)) != -1) { int newmode = mode; /* firstly, some mode-independent options */ switch(opt) { @@ -140,9 +151,9 @@ int main(int argc, char *argv[]) continue; case 'b': - if (mode == ASSEMBLE || mode == BUILD || mode == CREATE - || mode == GROW || mode == INCREMENTAL - || mode == MANAGE) + if (mode == ASSEMBLE || mode == BUILD || + mode == CREATE || mode == GROW || + mode == INCREMENTAL || mode == MANAGE) break; /* b means bitmap */ case Brief: c.brief = 1; @@ -158,14 +169,8 @@ int main(int argc, char *argv[]) c.homehost = optarg; continue; - /* - * --offroot sets first char of argv[0] to @. This is used - * by systemd to signal that the task was launched from - * initrd/initramfs and should be preserved during shutdown - */ case OffRootOpt: - argv[0][0] = '@'; - __offroot = 1; + /* Silently ignore old option */ continue; case Prefer: @@ -192,11 +197,16 @@ int main(int argc, char *argv[]) break; case 'a': case Add: + case AddSpare: + case AddJournal: case 'r': case Remove: + case Replace: + case With: case 'f': case Fail: case ReAdd: /* re-add */ + case ClusterConfirm: if (!mode) { newmode = MANAGE; shortopt = short_bitmap_options; @@ -229,6 +239,10 @@ int main(int argc, char *argv[]) case 'E': case 'X': case 'Q': + case ExamineBB: + case Dump: + case Restore: + case Action: newmode = MISC; break; @@ -270,14 +284,14 @@ int main(int argc, char *argv[]) if (mode == MISC && devs_found) { pr_err("No action given for %s in --misc mode\n", devlist->devname); - fprintf(stderr," Action options must come before device names\n"); + cont_err("Action options must come before device names\n"); exit(2); } } else { /* special case of -c --help */ if ((opt == 'c' || opt == ConfigFile) && - ( strncmp(optarg, "--h", 3)==0 || - strncmp(optarg, "-h", 2)==0)) { + (strncmp(optarg, "--h", 3) == 0 || + strncmp(optarg, "-h", 2) == 0)) { fputs(Help_config, stdout); exit(0); } @@ -289,6 +303,7 @@ int main(int argc, char *argv[]) dv->devname = optarg; dv->disposition = devmode; dv->writemostly = writemostly; + dv->failfast = failfast; dv->used = 0; dv->next = NULL; *devlistend = dv; @@ -324,22 +339,28 @@ int main(int argc, char *argv[]) continue; } if (opt == 1) { - /* an undecorated option - must be a device name. + /* an undecorated option - must be a device name. */ + + if (devs_found > 0 && devmode == DetailPlatform) { + pr_err("controller may only be specified once. %s ignored\n", + optarg); + continue; + } + if (devs_found > 0 && mode == MANAGE && !devmode) { - pr_err("Must give one of -a/-r/-f" - " for subsequent devices at %s\n", optarg); + pr_err("Must give one of -a/-r/-f for subsequent devices at %s\n", optarg); exit(2); } if (devs_found > 0 && mode == GROW && !devmode) { - pr_err("Must give -a/--add for" - " devices to add: %s\n", optarg); + pr_err("Must give -a/--add for devices to add: %s\n", optarg); exit(2); } dv = xmalloc(sizeof(*dv)); dv->devname = optarg; dv->disposition = devmode; dv->writemostly = writemostly; + dv->failfast = failfast; dv->used = 0; dv->next = NULL; *devlistend = dv; @@ -360,12 +381,12 @@ int main(int argc, char *argv[]) case O(BUILD,'c'): /* chunk or rounding */ case O(BUILD,ChunkSize): /* chunk or rounding */ if (s.chunk) { - pr_err("chunk/rounding may only be specified once. " - "Second value is %s.\n", optarg); + pr_err("chunk/rounding may only be specified once. Second value is %s.\n", optarg); exit(2); } s.chunk = parse_size(optarg); - if (s.chunk < 8 || (s.chunk&1)) { + if (s.chunk == INVALID_SECTORS || + s.chunk < 8 || (s.chunk&1)) { pr_err("invalid chunk/rounding value: %s\n", optarg); exit(2); @@ -382,7 +403,7 @@ int main(int argc, char *argv[]) pr_err("metadata information already given\n"); exit(2); } - for(i=0; !ss && superlist[i]; i++) + for(i = 0; !ss && superlist[i]; i++) ss = superlist[i]->match_metadata_desc(optarg); if (!ss) { @@ -398,30 +419,35 @@ int main(int argc, char *argv[]) case O(CREATE,'W'): case O(CREATE,WriteMostly): /* set write-mostly for following devices */ - writemostly = 1; + writemostly = FlagSet; continue; case O(MANAGE,'w'): /* clear write-mostly for following devices */ - writemostly = 2; + writemostly = FlagClear; continue; + case O(MANAGE,FailFast): + case O(CREATE,FailFast): + failfast = FlagSet; + continue; + case O(MANAGE,NoFailFast): + failfast = FlagClear; + continue; case O(GROW,'z'): case O(CREATE,'z'): case O(BUILD,'z'): /* size */ if (s.size > 0) { - pr_err("size may only be specified once. " - "Second value is %s.\n", optarg); + pr_err("size may only be specified once. Second value is %s.\n", optarg); exit(2); } - if (strcmp(optarg, "max")==0) + if (strcmp(optarg, "max") == 0) s.size = MAX_SIZE; else { s.size = parse_size(optarg); - if (s.size < 8) { - pr_err("invalid size: %s\n", - optarg); + if (s.size == INVALID_SECTORS || s.size < 8) { + pr_err("invalid size: %s\n", optarg); exit(2); } /* convert sectors to K */ @@ -431,15 +457,15 @@ int main(int argc, char *argv[]) case O(GROW,'Z'): /* array size */ if (array_size > 0) { - pr_err("array-size may only be specified once. " - "Second value is %s.\n", optarg); + pr_err("array-size may only be specified once. Second value is %s.\n", optarg); exit(2); } if (strcmp(optarg, "max") == 0) array_size = MAX_SIZE; else { array_size = parse_size(optarg); - if (array_size <= 0) { + if (array_size == 0 || + array_size == INVALID_SECTORS) { pr_err("invalid array size: %s\n", optarg); exit(2); @@ -447,12 +473,28 @@ int main(int argc, char *argv[]) } continue; + case O(CREATE,DataOffset): + case O(GROW,DataOffset): + if (data_offset != INVALID_SECTORS) { + pr_err("data-offset may only be specified one. Second value is %s.\n", optarg); + exit(2); + } + if (mode == CREATE && strcmp(optarg, "variable") == 0) + data_offset = VARIABLE_OFFSET; + else + data_offset = parse_size(optarg); + if (data_offset == INVALID_SECTORS) { + pr_err("invalid data-offset: %s\n", + optarg); + exit(2); + } + continue; + case O(GROW,'l'): case O(CREATE,'l'): case O(BUILD,'l'): /* set raid level*/ if (s.level != UnSet) { - pr_err("raid level may only be set once. " - "Second value is %s.\n", optarg); + pr_err("raid level may only be set once. Second value is %s.\n", optarg); exit(2); } s.level = map_name(pers, optarg); @@ -461,9 +503,9 @@ int main(int argc, char *argv[]) optarg); exit(2); } - if (s.level != 0 && s.level != LEVEL_LINEAR && s.level != 1 && - s.level != LEVEL_MULTIPATH && s.level != LEVEL_FAULTY && - s.level != 10 && + if (s.level != 0 && s.level != LEVEL_LINEAR && + s.level != 1 && s.level != LEVEL_MULTIPATH && + s.level != LEVEL_FAULTY && s.level != 10 && mode == BUILD) { pr_err("Raid level %s not permitted with --build.\n", optarg); @@ -480,8 +522,7 @@ int main(int argc, char *argv[]) case O(GROW, 'p'): /* new layout */ case O(GROW, Layout): if (s.layout_str) { - pr_err("layout may only be sent once. " - "Second value was %s\n", optarg); + pr_err("layout may only be sent once. Second value was %s\n", optarg); exit(2); } s.layout_str = optarg; @@ -493,8 +534,7 @@ int main(int argc, char *argv[]) case O(BUILD,'p'): /* faulty layout */ case O(BUILD,Layout): if (s.layout != UnSet) { - pr_err("layout may only be sent once. " - "Second value was %s\n", optarg); + pr_err("layout may only be sent once. Second value was %s\n", optarg); exit(2); } switch(s.level) { @@ -508,7 +548,7 @@ int main(int argc, char *argv[]) case 5: s.layout = map_name(r5layout, optarg); - if (s.layout==UnSet) { + if (s.layout == UnSet) { pr_err("layout %s not understood for raid5.\n", optarg); exit(2); @@ -516,7 +556,7 @@ int main(int argc, char *argv[]) break; case 6: s.layout = map_name(r6layout, optarg); - if (s.layout==UnSet) { + if (s.layout == UnSet) { pr_err("layout %s not understood for raid6.\n", optarg); exit(2); @@ -566,7 +606,24 @@ int main(int argc, char *argv[]) } ident.raid_disks = s.raiddisks; continue; - + case O(ASSEMBLE, Nodes): + case O(GROW, Nodes): + case O(CREATE, Nodes): + c.nodes = parse_num(optarg); + if (c.nodes < 2) { + pr_err("clustered array needs two nodes at least: %s\n", + optarg); + exit(2); + } + continue; + case O(CREATE, ClusterName): + case O(ASSEMBLE, ClusterName): + c.homecluster = optarg; + if (strlen(c.homecluster) > 64) { + pr_err("Cluster name too big.\n"); + exit(2); + } + continue; case O(CREATE,'x'): /* number of spare (eXtra) disks */ if (s.sparedisks) { pr_err("spare-devices set twice: %d and %s\n", @@ -614,7 +671,7 @@ int main(int argc, char *argv[]) case O(MISC,'f'): /* force zero */ case O(MISC,Force): /* force zero */ case O(MANAGE,Force): /* add device which is too large */ - c.force=1; + c.force = 1; continue; /* now for the Assemble options */ case O(ASSEMBLE, FreezeReshape): /* Freeze reshape during @@ -625,8 +682,7 @@ int main(int argc, char *argv[]) case O(CREATE,'u'): /* uuid of array */ case O(ASSEMBLE,'u'): /* uuid of array */ if (ident.uuid_set) { - pr_err("uuid cannot be set twice. " - "Second value %s.\n", optarg); + pr_err("uuid cannot be set twice. Second value %s.\n", optarg); exit(2); } if (parse_uuid(optarg, ident.uuid)) @@ -641,8 +697,7 @@ int main(int argc, char *argv[]) case O(ASSEMBLE,'N'): case O(MISC,'N'): if (ident.name[0]) { - pr_err("name cannot be set twice. " - "Second value %s.\n", optarg); + pr_err("name cannot be set twice. Second value %s.\n", optarg); exit(2); } if (mode == MISC && !c.subarray) { @@ -660,11 +715,10 @@ int main(int argc, char *argv[]) case O(ASSEMBLE,'m'): /* super-minor for array */ case O(ASSEMBLE,SuperMinor): if (ident.super_minor != UnSet) { - pr_err("super-minor cannot be set twice. " - "Second value: %s.\n", optarg); + pr_err("super-minor cannot be set twice. Second value: %s.\n", optarg); exit(2); } - if (strcmp(optarg, "dev")==0) + if (strcmp(optarg, "dev") == 0) ident.super_minor = -2; else { ident.super_minor = parse_num(optarg); @@ -684,47 +738,61 @@ int main(int argc, char *argv[]) case O(ASSEMBLE,'U'): /* update the superblock */ case O(MISC,'U'): if (c.update) { - pr_err("Can only update one aspect" - " of superblock, both %s and %s given.\n", + pr_err("Can only update one aspect of superblock, both %s and %s given.\n", c.update, optarg); exit(2); } if (mode == MISC && !c.subarray) { - pr_err("Only subarrays can be" - " updated in misc mode\n"); + pr_err("Only subarrays can be updated in misc mode\n"); exit(2); } c.update = optarg; - if (strcmp(c.update, "sparc2.2")==0) + if (strcmp(c.update, "sparc2.2") == 0) continue; if (strcmp(c.update, "super-minor") == 0) continue; - if (strcmp(c.update, "summaries")==0) + if (strcmp(c.update, "summaries") == 0) + continue; + if (strcmp(c.update, "resync") == 0) continue; - if (strcmp(c.update, "resync")==0) + if (strcmp(c.update, "uuid") == 0) continue; - if (strcmp(c.update, "uuid")==0) + if (strcmp(c.update, "name") == 0) continue; - if (strcmp(c.update, "name")==0) + if (strcmp(c.update, "homehost") == 0) continue; - if (strcmp(c.update, "homehost")==0) + if (strcmp(c.update, "home-cluster") == 0) continue; - if (strcmp(c.update, "devicesize")==0) + if (strcmp(c.update, "nodes") == 0) continue; - if (strcmp(c.update, "no-bitmap")==0) + if (strcmp(c.update, "devicesize") == 0) continue; - if (strcmp(c.update, "byteorder")==0) { + if (strcmp(c.update, "no-bitmap") == 0) + continue; + if (strcmp(c.update, "bbl") == 0) + continue; + if (strcmp(c.update, "no-bbl") == 0) + continue; + if (strcmp(c.update, "force-no-bbl") == 0) + continue; + if (strcmp(c.update, "ppl") == 0) + continue; + if (strcmp(c.update, "no-ppl") == 0) + continue; + if (strcmp(c.update, "metadata") == 0) + continue; + if (strcmp(c.update, "revert-reshape") == 0) + continue; + if (strcmp(c.update, "byteorder") == 0) { if (ss) { - pr_err("must not set metadata" - " type with --update=byteorder.\n"); + pr_err("must not set metadata type with --update=byteorder.\n"); exit(2); } - for(i=0; !ss && superlist[i]; i++) + for(i = 0; !ss && superlist[i]; i++) ss = superlist[i]->match_metadata_desc( "0.swap"); if (!ss) { - pr_err("INTERNAL ERROR" - " cannot find 0.swap\n"); + pr_err("INTERNAL ERROR cannot find 0.swap\n"); exit(2); } @@ -733,36 +801,38 @@ int main(int argc, char *argv[]) if (strcmp(c.update,"?") == 0 || strcmp(c.update, "help") == 0) { outf = stdout; - fprintf(outf, Name ": "); + fprintf(outf, "%s: ", Name); } else { outf = stderr; fprintf(outf, - Name ": '--update=%s' is invalid. ", - c.update); + "%s: '--update=%s' is invalid. ", + Name, c.update); } fprintf(outf, "Valid --update options are:\n" - " 'sparc2.2', 'super-minor', 'uuid', 'name', 'resync',\n" - " 'summaries', 'homehost', 'byteorder', 'devicesize',\n" - " 'no-bitmap'\n"); + " 'sparc2.2', 'super-minor', 'uuid', 'name', 'nodes', 'resync',\n" + " 'summaries', 'homehost', 'home-cluster', 'byteorder', 'devicesize',\n" + " 'no-bitmap', 'metadata', 'revert-reshape'\n" + " 'bbl', 'no-bbl', 'force-no-bbl', 'ppl', 'no-ppl'\n" + ); exit(outf == stdout ? 0 : 2); case O(MANAGE,'U'): /* update=devicesize is allowed with --re-add */ if (devmode != 'A') { - pr_err("--update in Manage mode only" - " allowed with --re-add.\n"); + pr_err("--update in Manage mode only allowed with --re-add.\n"); exit(1); } if (c.update) { - pr_err("Can only update one aspect" - " of superblock, both %s and %s given.\n", + pr_err("Can only update one aspect of superblock, both %s and %s given.\n", c.update, optarg); exit(2); } c.update = optarg; - if (strcmp(c.update, "devicesize") != 0) { - pr_err("only 'devicesize' can be" - " updated with --re-add\n"); + if (strcmp(c.update, "devicesize") != 0 && + strcmp(c.update, "bbl") != 0 && + strcmp(c.update, "force-no-bbl") != 0 && + strcmp(c.update, "no-bbl") != 0) { + pr_err("only 'devicesize', 'bbl', 'no-bbl', and 'force-no-bbl' can be updated with --re-add\n"); exit(2); } continue; @@ -782,9 +852,9 @@ int main(int argc, char *argv[]) case O(MISC, ConfigFile): case O(MONITOR,'c'): case O(MONITOR,ConfigFile): + case O(CREATE,ConfigFile): if (configfile) { - pr_err("configfile cannot be set twice. " - "Second value is %s.\n", optarg); + pr_err("configfile cannot be set twice. Second value is %s.\n", optarg); exit(2); } configfile = optarg; @@ -876,6 +946,16 @@ int main(int argc, char *argv[]) case O(MANAGE,Add): /* add a drive */ devmode = 'a'; continue; + case O(MANAGE,AddSpare): /* add drive - never re-add */ + devmode = 'S'; + continue; + case O(MANAGE,AddJournal): /* add journal */ + if (s.journaldisks && (s.level < 4 || s.level > 6)) { + pr_err("--add-journal is only supported for RAID level 4/5/6.\n"); + exit(2); + } + devmode = 'j'; + continue; case O(MANAGE,ReAdd): devmode = 'A'; continue; @@ -892,6 +972,21 @@ int main(int argc, char *argv[]) * remove the device */ devmode = 'f'; continue; + case O(MANAGE, ClusterConfirm): + devmode = 'c'; + continue; + case O(MANAGE,Replace): + /* Mark these devices for replacement */ + devmode = 'R'; + continue; + case O(MANAGE,With): + /* These are the replacements to use */ + if (devmode != 'R') { + pr_err("--with must follow --replace\n"); + exit(2); + } + devmode = 'W'; + continue; case O(INCREMENTAL,'R'): case O(MANAGE,'R'): case O(ASSEMBLE,'R'): @@ -921,6 +1016,7 @@ int main(int argc, char *argv[]) case O(MISC,'R'): case O(MISC,'S'): case O(MISC,'X'): + case O(MISC, ExamineBB): case O(MISC,'o'): case O(MISC,'w'): case O(MISC,'W'): @@ -929,16 +1025,34 @@ int main(int argc, char *argv[]) case O(MISC, DetailPlatform): case O(MISC, KillSubarray): case O(MISC, UpdateSubarray): + case O(MISC, Dump): + case O(MISC, Restore): + case O(MISC ,Action): if (opt == KillSubarray || opt == UpdateSubarray) { if (c.subarray) { - pr_err("subarray can only" - " be specified once\n"); + pr_err("subarray can only be specified once\n"); exit(2); } c.subarray = optarg; } + if (opt == Action) { + if (c.action) { + pr_err("Only one --action can be specified\n"); + exit(2); + } + if (strcmp(optarg, "idle") == 0 || + strcmp(optarg, "frozen") == 0 || + strcmp(optarg, "check") == 0 || + strcmp(optarg, "repair") == 0) + c.action = optarg; + else { + pr_err("action must be one of idle, frozen, check, repair\n"); + exit(2); + } + } if (devmode && devmode != opt && - (devmode == 'E' || (opt == 'E' && devmode != 'Q'))) { + (devmode == 'E' || + (opt == 'E' && devmode != 'Q'))) { pr_err("--examine/-E cannot be given with "); if (devmode == 'E') { if (option_index >= 0) @@ -953,15 +1067,21 @@ int main(int argc, char *argv[]) exit(2); } devmode = opt; + if (opt == Dump || opt == Restore) { + if (dump_directory != NULL) { + pr_err("dump/restore directory specified twice: %s and %s\n", + dump_directory, optarg); + exit(2); + } + dump_directory = optarg; + } continue; case O(MISC, UdevRules): if (devmode && devmode != opt) { - pr_err("--udev-rules must" - " be the only option.\n"); + pr_err("--udev-rules must be the only option.\n"); } else { if (udev_filename) - pr_err("only specify one udev " - "rule filename. %s ignored.\n", + pr_err("only specify one udev rule filename. %s ignored.\n", optarg); else udev_filename = optarg; @@ -986,8 +1106,10 @@ int main(int argc, char *argv[]) pr_err("bitmap file needed with -b in --assemble mode\n"); exit(2); } - if (strcmp(optarg, "internal")==0) { - pr_err("there is no need to specify --bitmap when assembling arrays with internal bitmaps\n"); + if (strcmp(optarg, "internal") == 0 || + strcmp(optarg, "clustered") == 0) { + pr_err("no need to specify --bitmap when assembling" + " arrays with internal or clustered bitmap\n"); continue; } bitmap_fd = open(optarg, O_RDWR); @@ -1026,22 +1148,29 @@ int main(int argc, char *argv[]) case O(BUILD,Bitmap): case O(CREATE,'b'): case O(CREATE,Bitmap): /* here we create the bitmap */ - if (strcmp(optarg, "none") == 0) { - pr_err("'--bitmap none' only" - " supported for --grow\n"); - exit(2); - } - /* FALL THROUGH */ case O(GROW,'b'): case O(GROW,Bitmap): - if (strcmp(optarg, "internal")== 0 || - strcmp(optarg, "none")== 0 || + if (s.bitmap_file) { + pr_err("bitmap cannot be set twice. Second value: %s.\n", optarg); + exit(2); + } + if (strcmp(optarg, "internal") == 0 || + strcmp(optarg, "none") == 0 || strchr(optarg, '/') != NULL) { s.bitmap_file = optarg; continue; } + if (strcmp(optarg, "clustered") == 0) { + s.bitmap_file = optarg; + /* Set the default number of cluster nodes + * to 4 if not already set by user + */ + if (c.nodes < 1) + c.nodes = 4; + continue; + } /* probable typo */ - pr_err("bitmap file must contain a '/', or be 'internal', or 'none'\n" + pr_err("bitmap file must contain a '/', or be 'internal', or be 'clustered', or 'none'\n" " not '%s'\n", optarg); exit(2); @@ -1049,7 +1178,8 @@ int main(int argc, char *argv[]) case O(BUILD,BitmapChunk): case O(CREATE,BitmapChunk): /* bitmap chunksize */ s.bitmap_chunk = parse_size(optarg); - if (s.bitmap_chunk <= 0 || + if (s.bitmap_chunk == 0 || + s.bitmap_chunk == INVALID_SECTORS || s.bitmap_chunk & (s.bitmap_chunk - 1)) { pr_err("invalid bitmap chunksize: %s\n", optarg); @@ -1079,6 +1209,33 @@ int main(int argc, char *argv[]) case O(INCREMENTAL, IncrementalPath): remove_path = optarg; continue; + case O(CREATE, WriteJournal): + if (s.journaldisks) { + pr_err("Please specify only one journal device for the array.\n"); + pr_err("Ignoring --write-journal %s...\n", optarg); + continue; + } + dv = xmalloc(sizeof(*dv)); + dv->devname = optarg; + dv->disposition = 'j'; /* WriteJournal */ + dv->used = 0; + dv->next = NULL; + *devlistend = dv; + devlistend = &dv->next; + devs_found++; + + s.journaldisks = 1; + continue; + case O(CREATE, 'k'): + case O(GROW, 'k'): + s.consistency_policy = map_name(consistency_policies, + optarg); + if (s.consistency_policy < CONSISTENCY_POLICY_RESYNC) { + pr_err("Invalid consistency policy: %s\n", + optarg); + exit(2); + } + continue; } /* We have now processed all the valid options. Anything else is * an error @@ -1106,6 +1263,50 @@ int main(int argc, char *argv[]) exit(0); } + if (s.journaldisks) { + if (s.level < 4 || s.level > 6) { + pr_err("--write-journal is only supported for RAID level 4/5/6.\n"); + exit(2); + } + if (s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN && + s.consistency_policy != CONSISTENCY_POLICY_JOURNAL) { + pr_err("--write-journal is not supported with consistency policy: %s\n", + map_num(consistency_policies, s.consistency_policy)); + exit(2); + } + } + + if (mode == CREATE && + s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN) { + if (s.level <= 0) { + pr_err("--consistency-policy not meaningful with level %s.\n", + map_num(pers, s.level)); + exit(2); + } else if (s.consistency_policy == CONSISTENCY_POLICY_JOURNAL && + !s.journaldisks) { + pr_err("--write-journal is required for consistency policy: %s\n", + map_num(consistency_policies, s.consistency_policy)); + exit(2); + } else if (s.consistency_policy == CONSISTENCY_POLICY_PPL && + s.level != 5) { + pr_err("PPL consistency policy is only supported for RAID level 5.\n"); + exit(2); + } else if (s.consistency_policy == CONSISTENCY_POLICY_BITMAP && + (!s.bitmap_file || + strcmp(s.bitmap_file, "none") == 0)) { + pr_err("--bitmap is required for consistency policy: %s\n", + map_num(consistency_policies, s.consistency_policy)); + exit(2); + } else if (s.bitmap_file && + strcmp(s.bitmap_file, "none") != 0 && + s.consistency_policy != CONSISTENCY_POLICY_BITMAP && + s.consistency_policy != CONSISTENCY_POLICY_JOURNAL) { + pr_err("--bitmap is not compatible with consistency policy: %s\n", + map_num(consistency_policies, s.consistency_policy)); + exit(2); + } + } + if (!mode && devs_found) { mode = MISC; devmode = 'Q'; @@ -1133,15 +1334,14 @@ int main(int argc, char *argv[]) * hopefully it's mostly right but there might be some stuff * missing * - * That is mosty checked in the per-mode stuff but... + * That is mostly checked in the per-mode stuff but... * * For @,B,C and A without -s, the first device listed must be * an md device. We check that here and open it. */ - if (mode == MANAGE || mode == BUILD || mode == CREATE - || mode == GROW - || (mode == ASSEMBLE && ! c.scan)) { + if (mode == MANAGE || mode == BUILD || mode == CREATE || + mode == GROW || (mode == ASSEMBLE && ! c.scan)) { if (devs_found < 1) { pr_err("an md device must be given in this mode\n"); exit(2); @@ -1158,15 +1358,13 @@ int main(int argc, char *argv[]) /* non-existent device is OK */ mdfd = open_mddev(devlist->devname, 0); if (mdfd == -2) { - pr_err("device %s exists but is not an " - "md array.\n", devlist->devname); + pr_err("device %s exists but is not an md array.\n", devlist->devname); exit(1); } if ((int)ident.super_minor == -2) { struct stat stb; if (mdfd < 0) { - pr_err("--super-minor=dev given, and " - "listed device %s doesn't exist.\n", + pr_err("--super-minor=dev given, and listed device %s doesn't exist.\n", devlist->devname); exit(1); } @@ -1191,21 +1389,41 @@ int main(int argc, char *argv[]) } } - if (c.homehost == NULL) + if (c.homehost == NULL && c.require_homehost) c.homehost = conf_get_homehost(&c.require_homehost); - if (c.homehost == NULL || strcasecmp(c.homehost, "")==0) { + if (c.homehost == NULL || strcasecmp(c.homehost, "") == 0) { if (gethostname(sys_hostname, sizeof(sys_hostname)) == 0) { sys_hostname[sizeof(sys_hostname)-1] = 0; c.homehost = sys_hostname; } } - if (c.homehost && (!c.homehost[0] || strcasecmp(c.homehost, "") == 0)) { + if (c.homehost && + (!c.homehost[0] || strcasecmp(c.homehost, "") == 0)) { c.homehost = NULL; c.require_homehost = 0; } - if ((mode == MISC && devmode == 'E') - || (mode == MONITOR && spare_sharing == 0)) + rv = 0; + + set_hooks(); /* set hooks from libs */ + + if (c.homecluster == NULL && (c.nodes > 0)) { + c.homecluster = conf_get_homecluster(); + if (c.homecluster == NULL) + rv = get_cluster_name(&c.homecluster); + if (rv) { + pr_err("The md can't get cluster name\n"); + exit(1); + } + } + + if (c.backup_file && data_offset != INVALID_SECTORS) { + pr_err("--backup-file and --data-offset are incompatible\n"); + exit(2); + } + + if ((mode == MISC && devmode == 'E') || + (mode == MONITOR && spare_sharing == 0)) /* Anyone may try this */; else if (geteuid() != 0) { pr_err("must be super-user to perform this action\n"); @@ -1218,7 +1436,22 @@ int main(int argc, char *argv[]) /* --scan implied --brief unless -vv */ c.brief = 1; - rv = 0; + if (mode == CREATE) { + if (s.bitmap_file && strcmp(s.bitmap_file, "clustered") == 0) { + locked = cluster_get_dlmlock(); + if (locked != 1) + exit(1); + } + } else if (mode == MANAGE || mode == GROW || mode == INCREMENTAL) { + if (!md_get_array_info(mdfd, &array) && (devmode != 'c')) { + if (array.state & (1 << MD_SB_CLUSTERED)) { + locked = cluster_get_dlmlock(); + if (locked != 1) + exit(1); + } + } + } + switch(mode) { case MANAGE: /* readonly, add/remove, readwrite, runstop */ @@ -1230,12 +1463,15 @@ int main(int argc, char *argv[]) c.update, c.force); if (!rv && c.readonly < 0) rv = Manage_ro(devlist->devname, mdfd, c.readonly); - if (!rv && c.runstop) - rv = Manage_runstop(devlist->devname, mdfd, c.runstop, c.verbose, 0); + if (!rv && c.runstop > 0) + rv = Manage_run(devlist->devname, mdfd, &c); + if (!rv && c.runstop < 0) + rv = Manage_stop(devlist->devname, mdfd, c.verbose, 0); break; case ASSEMBLE: if (devs_found == 1 && ident.uuid_set == 0 && - ident.super_minor == UnSet && ident.name[0] == 0 && !c.scan ) { + ident.super_minor == UnSet && ident.name[0] == 0 && + !c.scan ) { /* Only a device has been given, so get details from config file */ struct mddev_ident *array_ident = conf_get_ident(devlist->devname); if (array_ident == NULL) { @@ -1262,7 +1498,7 @@ int main(int argc, char *argv[]) pr_err("can only assemble a single array when providing a backup file.\n"); exit(1); } - for (dv = devlist ; dv ; dv=dv->next) { + for (dv = devlist; dv; dv = dv->next) { struct mddev_ident *array_ident = conf_get_ident(dv->devname); if (array_ident == NULL) { pr_err("%s not identified in config file.\n", @@ -1303,8 +1539,9 @@ int main(int argc, char *argv[]) } if (s.bitmap_file) { - if (strcmp(s.bitmap_file, "internal")==0) { - pr_err("'internal' bitmaps not supported with --build\n"); + if (strcmp(s.bitmap_file, "internal") == 0 || + strcmp(s.bitmap_file, "clustered") == 0) { + pr_err("'internal' and 'clustered' bitmaps not supported with --build\n"); rv |= 1; break; } @@ -1314,6 +1551,27 @@ int main(int argc, char *argv[]) case CREATE: if (c.delay == 0) c.delay = DEFAULT_BITMAP_DELAY; + + if (c.nodes) { + if (!s.bitmap_file || + strcmp(s.bitmap_file, "clustered") != 0) { + pr_err("--nodes argument only compatible with --bitmap=clustered\n"); + rv = 1; + break; + } + + if (s.level != 1 && s.level != 10) { + pr_err("--bitmap=clustered is currently supported with raid1/10 only\n"); + rv = 1; + break; + } + if (s.level == 10 && !(is_near_layout_10(s.layout) || s.layout == UnSet)) { + pr_err("only near layout is supported with clustered raid10\n"); + rv = 1; + break; + } + } + if (s.write_behind && !s.bitmap_file) { pr_err("write-behind mode requires a bitmap.\n"); rv = 1; @@ -1328,7 +1586,7 @@ int main(int argc, char *argv[]) rv = Create(ss, devlist->devname, ident.name, ident.uuid_set ? ident.uuid : NULL, devs_found-1, devlist->next, - &s, &c); + &s, &c, data_offset); break; case MISC: if (devmode == 'E') { @@ -1344,11 +1602,14 @@ int main(int argc, char *argv[]) } rv = Examine(devlist, &c, ss); } else if (devmode == DetailPlatform) { - rv = Detail_Platform(ss ? ss->ss : NULL, ss ? c.scan : 1, c.verbose); + rv = Detail_Platform(ss ? ss->ss : NULL, ss ? c.scan : 1, + c.verbose, c.export, + devlist ? devlist->devname : NULL); } else if (devlist == NULL) { if (devmode == 'S' && c.scan) rv = stop_scan(c.verbose); - else if ((devmode == 'D' || devmode == Waitclean) && c.scan) + else if ((devmode == 'D' || devmode == Waitclean) && + c.scan) rv = misc_scan(devmode, &c); else if (devmode == UdevRules) rv = Write_rules(udev_filename); @@ -1357,7 +1618,7 @@ int main(int argc, char *argv[]) exit(2); } } else - rv = misc_list(devlist, &ident, ss, &c); + rv = misc_list(devlist, &ident, dump_directory, ss, &c); break; case MONITOR: if (!devlist && !c.scan) { @@ -1377,12 +1638,10 @@ int main(int argc, char *argv[]) else c.delay = 60; } - if (c.delay == 0) - c.delay = 60; - rv= Monitor(devlist, mailaddr, program, - &c, daemonise, oneshot, - dosyslog, pidfile, increments, - spare_sharing); + rv = Monitor(devlist, mailaddr, program, + &c, daemonise, oneshot, + dosyslog, pidfile, increments, + spare_sharing); break; case GROW: @@ -1395,37 +1654,37 @@ int main(int argc, char *argv[]) struct mdinfo sra; int err; if (s.raiddisks || s.level != UnSet) { - pr_err("cannot change array size in same operation " - "as changing raiddisks or level.\n" + pr_err("cannot change array size in same operation as changing raiddisks or level.\n" " Change size first, then check that data is still intact.\n"); rv = 1; break; } - sysfs_init(&sra, mdfd, 0); + if (sysfs_init(&sra, mdfd, NULL)) { + rv = 1; + break; + } if (array_size == MAX_SIZE) err = sysfs_set_str(&sra, NULL, "array_size", "default"); else err = sysfs_set_num(&sra, NULL, "array_size", array_size / 2); if (err < 0) { if (errno == E2BIG) - pr_err("--array-size setting" - " is too large.\n"); + pr_err("--array-size setting is too large.\n"); else - pr_err("current kernel does" - " not support setting --array-size\n"); + pr_err("current kernel does not support setting --array-size\n"); rv = 1; break; } } - if (devs_found > 1 && s.raiddisks == 0) { + if (devs_found > 1 && s.raiddisks == 0 && s.level == UnSet) { /* must be '-a'. */ - if (s.size > 0 || s.chunk || s.layout_str != NULL || s.bitmap_file) { - pr_err("--add cannot be used with " - "other geometry changes in --grow mode\n"); + if (s.size > 0 || s.chunk || + s.layout_str || s.bitmap_file) { + pr_err("--add cannot be used with other geometry changes in --grow mode\n"); rv = 1; break; } - for (dv=devlist->next; dv ; dv=dv->next) { + for (dv = devlist->next; dv; dv = dv->next) { rv = Grow_Add_device(devlist->devname, mdfd, dv->devname); if (rv) @@ -1433,25 +1692,26 @@ int main(int argc, char *argv[]) } } else if (s.bitmap_file) { if (s.size > 0 || s.raiddisks || s.chunk || - s.layout_str != NULL || devs_found > 1) { - pr_err("--bitmap changes cannot be " - "used with other geometry changes " - "in --grow mode\n"); + s.layout_str || devs_found > 1) { + pr_err("--bitmap changes cannot be used with other geometry changes in --grow mode\n"); rv = 1; break; } if (c.delay == 0) c.delay = DEFAULT_BITMAP_DELAY; - rv = Grow_addbitmap(devlist->devname, mdfd, s.bitmap_file, - s.bitmap_chunk, c.delay, s.write_behind, c.force); + rv = Grow_addbitmap(devlist->devname, mdfd, &c, &s); } else if (grow_continue) rv = Grow_continue_command(devlist->devname, mdfd, c.backup_file, c.verbose); - else if (s.size > 0 || s.raiddisks || s.layout_str != NULL - || s.chunk != 0 || s.level != UnSet) { + else if (s.size > 0 || s.raiddisks || s.layout_str || + s.chunk != 0 || s.level != UnSet || + data_offset != INVALID_SECTORS) { rv = Grow_reshape(devlist->devname, mdfd, - devlist->next, &c, &s); + devlist->next, + data_offset, &c, &s); + } else if (s.consistency_policy != CONSISTENCY_POLICY_UNKNOWN) { + rv = Grow_consistency_policy(devlist->devname, mdfd, &c, &s); } else if (array_size == 0) pr_err("no changes to --grow\n"); break; @@ -1460,6 +1720,11 @@ int main(int argc, char *argv[]) RebuildMap(); } if (c.scan) { + rv = 1; + if (devlist) { + pr_err("In --incremental mode, a device cannot be given with --scan.\n"); + break; + } if (c.runstop <= 0) { pr_err("--incremental --scan meaningless without --run.\n"); break; @@ -1468,7 +1733,7 @@ int main(int argc, char *argv[]) pr_err("--incremental --scan --fail not supported.\n"); break; } - rv = IncrementalScan(c.verbose); + rv = IncrementalScan(&c, NULL); } if (!devlist) { if (!rebuild_map && !c.scan) { @@ -1477,21 +1742,25 @@ int main(int argc, char *argv[]) } break; } - if (devlist->next) { - pr_err("--incremental can only handle one device.\n"); - rv = 1; - break; - } - if (devmode == 'f') + if (devmode == 'f') { + if (devlist->next) { + pr_err("'--incremental --fail' can only handle one device.\n"); + rv = 1; + break; + } rv = IncrementalRemove(devlist->devname, remove_path, c.verbose); - else - rv = Incremental(devlist->devname, &c, ss); + } else + rv = Incremental(devlist, &c, ss); break; case AUTODETECT: autodetect(); break; } + if (locked) + cluster_release_dlmlock(); + if (mdfd > 0) + close(mdfd); exit(rv); } @@ -1507,35 +1776,32 @@ static int scan_assemble(struct supertype *ss, int failures, successes; if (conf_verify_devnames(array_list)) { - pr_err("Duplicate MD device names in " - "conf file were found.\n"); + pr_err("Duplicate MD device names in conf file were found.\n"); return 1; } if (devlist == NULL) { pr_err("No devices listed in conf file were found.\n"); return 1; } - for (a = array_list; a ; a = a->next) { + for (a = array_list; a; a = a->next) { a->assembled = 0; if (a->autof == 0) a->autof = c->autof; } if (map_lock(&map)) - pr_err("%s: failed to get " - "exclusive lock on mapfile\n", - __func__); + pr_err("failed to get exclusive lock on mapfile\n"); do { failures = 0; successes = 0; rv = 0; - for (a = array_list; a ; a = a->next) { + for (a = array_list; a; a = a->next) { int r; if (a->assembled) continue; if (a->devname && strcasecmp(a->devname, "") == 0) continue; - + r = Assemble(ss, a->devname, a, NULL, c); if (r == 0) { @@ -1562,11 +1828,11 @@ static int scan_assemble(struct supertype *ss, rv2 = Assemble(ss, NULL, ident, devlist, c); - if (rv2==0) { + if (rv2 == 0) { cnt++; acnt++; } - } while (rv2!=2); + } while (rv2 != 2); /* Incase there are stacked devices, we need to go around again */ } while (acnt); if (cnt == 0 && rv == 0) { @@ -1594,30 +1860,30 @@ static int misc_scan(char devmode, struct context *c) int rv = 0; for (members = 0; members <= 1; members++) { - for (e=ms ; e ; e=e->next) { - char *name; + for (e = ms; e; e = e->next) { + char *name = NULL; struct map_ent *me; + struct stat stb; int member = e->metadata_version && strncmp(e->metadata_version, "external:/", 10) == 0; if (members != member) continue; - me = map_by_devnum(&map, e->devnum); - if (me && me->path - && strcmp(me->path, "/unknown") != 0) + me = map_by_devnm(&map, e->devnm); + if (me && me->path && strcmp(me->path, "/unknown") != 0) name = me->path; - else - name = get_md_name(e->devnum); + if (name == NULL || stat(name, &stb) != 0) + name = get_md_name(e->devnm); if (!name) { pr_err("cannot find device file for %s\n", - e->dev); + e->devnm); continue; } if (devmode == 'D') rv |= Detail(name, c); else - rv |= WaitClean(name, -1, c->verbose); + rv |= WaitClean(name, c->verbose); put_md_name(name); } } @@ -1631,7 +1897,7 @@ static int stop_scan(int verbose) /* Due to possible stacking of devices, repeat until * nothing more can be stopped */ - int progress=1, err; + int progress = 1, err; int last = 0; int rv = 0; do { @@ -1640,18 +1906,18 @@ static int stop_scan(int verbose) if (!progress) last = 1; progress = 0; err = 0; - for (e=ms ; e ; e=e->next) { - char *name = get_md_name(e->devnum); + for (e = ms; e; e = e->next) { + char *name = get_md_name(e->devnm); int mdfd; if (!name) { pr_err("cannot find device file for %s\n", - e->dev); + e->devnm); continue; } mdfd = open_mddev(name, 1); if (mdfd >= 0) { - if (Manage_runstop(name, mdfd, -1, verbose, !last)) + if (Manage_stop(name, mdfd, verbose, !last)) err = 1; else progress = 1; @@ -1669,13 +1935,14 @@ static int stop_scan(int verbose) static int misc_list(struct mddev_dev *devlist, struct mddev_ident *ident, + char *dump_directory, struct supertype *ss, struct context *c) { struct mddev_dev *dv; int rv = 0; - for (dv=devlist ; dv; dv=dv->next) { - int mdfd; + for (dv = devlist; dv; dv = (rv & 16) ? NULL : dv->next) { + int mdfd = -1; switch(dv->disposition) { case 'D': @@ -1694,14 +1961,21 @@ static int misc_list(struct mddev_dev *devlist, } continue; case 'Q': - rv |= Query(dv->devname); continue; + rv |= Query(dv->devname); + continue; case 'X': - rv |= ExamineBitmap(dv->devname, c->brief, ss); continue; + rv |= ExamineBitmap(dv->devname, c->brief, ss); + continue; + case ExamineBB: + rv |= ExamineBadblocks(dv->devname, c->brief, ss); + continue; case 'W': case WaitOpt: - rv |= Wait(dv->devname); continue; + rv |= Wait(dv->devname); + continue; case Waitclean: - rv |= WaitClean(dv->devname, -1, c->verbose); continue; + rv |= WaitClean(dv->devname, c->verbose); + continue; case KillSubarray: rv |= Kill_subarray(dv->devname, c->subarray, c->verbose); continue; @@ -1714,18 +1988,38 @@ static int misc_list(struct mddev_dev *devlist, rv |= Update_subarray(dv->devname, c->subarray, c->update, ident, c->verbose); continue; + case Dump: + rv |= Dump_metadata(dv->devname, dump_directory, c, ss); + continue; + case Restore: + rv |= Restore_metadata(dv->devname, dump_directory, c, ss, + (dv == devlist && dv->next == NULL)); + continue; + case Action: + rv |= SetAction(dv->devname, c->action); + continue; } - mdfd = open_mddev(dv->devname, 1); - if (mdfd>=0) { + + if (dv->devname[0] != '/') + mdfd = open_dev(dv->devname); + if (dv->devname[0] == '/' || mdfd < 0) + mdfd = open_mddev(dv->devname, 1); + + if (mdfd >= 0) { switch(dv->disposition) { case 'R': - rv |= Manage_runstop(dv->devname, mdfd, 1, c->verbose, 0); break; + c->runstop = 1; + rv |= Manage_run(dv->devname, mdfd, c); + break; case 'S': - rv |= Manage_runstop(dv->devname, mdfd, -1, c->verbose, 0); break; + rv |= Manage_stop(dv->devname, mdfd, c->verbose, 0); + break; case 'o': - rv |= Manage_ro(dv->devname, mdfd, 1); break; + rv |= Manage_ro(dv->devname, mdfd, 1); + break; case 'w': - rv |= Manage_ro(dv->devname, mdfd, -1); break; + rv |= Manage_ro(dv->devname, mdfd, -1); + break; } close(mdfd); } else @@ -1733,3 +2027,28 @@ static int misc_list(struct mddev_dev *devlist, } return rv; } + +int SetAction(char *dev, char *action) +{ + int fd = open(dev, O_RDONLY); + struct mdinfo mdi; + int retval; + + if (fd < 0) { + pr_err("Couldn't open %s: %s\n", dev, strerror(errno)); + return 1; + } + retval = sysfs_init(&mdi, fd, NULL); + close(fd); + if (retval) { + pr_err("%s is no an md array\n", dev); + return 1; + } + + if (sysfs_set_str(&mdi, NULL, "sync_action", action) < 0) { + pr_err("Count not set action for %s to %s: %s\n", + dev, action, strerror(errno)); + return 1; + } + return 0; +}