ctdb/server/ctdb_recover.c

   1 /*
   2    ctdb recovery code
   3
   4    Copyright (C) Andrew Tridgell  2007
   5    Copyright (C) Ronnie Sahlberg  2007
   6
   7    This program is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 3 of the License, or
  10    (at your option) any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program; if not, see <http://www.gnu.org/licenses/>.
  19 */
  20 #include "replace.h"
  21 #include "system/time.h"
  22 #include "system/network.h"
  23 #include "system/filesys.h"
  24 #include "system/wait.h"
  25
  26 #include <talloc.h>
  27 #include <tevent.h>
  28 #include <tdb.h>
  29
  30 #include "lib/tdb_wrap/tdb_wrap.h"
  31 #include "lib/util/dlinklist.h"
  32 #include "lib/util/debug.h"
  33 #include "lib/util/time.h"
  34 #include "lib/util/util_process.h"
  35
  36 #include "ctdb_private.h"
  37 #include "ctdb_client.h"
  38
  39 #include "common/system.h"
  40 #include "common/common.h"
  41 #include "common/logging.h"
  42
  43 #include "ctdb_cluster_mutex.h"
  44
  45 int
  46 ctdb_control_getvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  47 {
  48         struct ctdb_vnn_map_wire *map;
  49         size_t len;
  50
  51         CHECK_CONTROL_DATA_SIZE(0);
  52
  53         len = offsetof(struct ctdb_vnn_map_wire, map) + sizeof(uint32_t)*ctdb->vnn_map->size;
  54         map = talloc_size(outdata, len);
  55         CTDB_NO_MEMORY(ctdb, map);
  56
  57         map->generation = ctdb->vnn_map->generation;
  58         map->size = ctdb->vnn_map->size;
  59         memcpy(map->map, ctdb->vnn_map->map, sizeof(uint32_t)*map->size);
  60
  61         outdata->dsize = len;
  62         outdata->dptr  = (uint8_t *)map;
  63
  64         return 0;
  65 }
  66
  67 int
  68 ctdb_control_setvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  69 {
  70         struct ctdb_vnn_map_wire *map = (struct ctdb_vnn_map_wire *)indata.dptr;
  71
  72         if (ctdb->recovery_mode != CTDB_RECOVERY_ACTIVE) {
  73                 DEBUG(DEBUG_ERR, ("Attempt to set vnnmap when not in recovery\n"));
  74                 return -1;
  75         }
  76
  77         talloc_free(ctdb->vnn_map);
  78
  79         ctdb->vnn_map = talloc(ctdb, struct ctdb_vnn_map);
  80         CTDB_NO_MEMORY(ctdb, ctdb->vnn_map);
  81
  82         ctdb->vnn_map->generation = map->generation;
  83         ctdb->vnn_map->size       = map->size;
  84         ctdb->vnn_map->map = talloc_array(ctdb->vnn_map, uint32_t, map->size);
  85         CTDB_NO_MEMORY(ctdb, ctdb->vnn_map->map);
  86
  87         memcpy(ctdb->vnn_map->map, map->map, sizeof(uint32_t)*map->size);
  88
  89         return 0;
  90 }
  91
  92 int
  93 ctdb_control_getdbmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  94 {
  95         uint32_t i, len;
  96         struct ctdb_db_context *ctdb_db;
  97         struct ctdb_dbid_map_old *dbid_map;
  98
  99         CHECK_CONTROL_DATA_SIZE(0);
 100
 101         len = 0;
 102         for(ctdb_db=ctdb->db_list;ctdb_db;ctdb_db=ctdb_db->next){
 103                 len++;
 104         }
 105
 106
 107         outdata->dsize = offsetof(struct ctdb_dbid_map_old, dbs) + sizeof(dbid_map->dbs[0])*len;
 108         outdata->dptr  = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
 109         if (!outdata->dptr) {
 110                 DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate dbmap array\n"));
 111                 exit(1);
 112         }
 113
 114         dbid_map = (struct ctdb_dbid_map_old *)outdata->dptr;
 115         dbid_map->num = len;
 116         for (i=0,ctdb_db=ctdb->db_list;ctdb_db;i++,ctdb_db=ctdb_db->next){
 117                 dbid_map->dbs[i].db_id       = ctdb_db->db_id;
 118                 dbid_map->dbs[i].flags       = ctdb_db->db_flags;
 119         }
 120
 121         return 0;
 122 }
 123
 124 int
 125 ctdb_control_getnodemap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
 126 {
 127         CHECK_CONTROL_DATA_SIZE(0);
 128
 129         outdata->dptr  = (unsigned char *)ctdb_node_list_to_map(ctdb->nodes,
 130                                                                 ctdb->num_nodes,
 131                                                                 outdata);
 132         if (outdata->dptr == NULL) {
 133                 return -1;
 134         }
 135
 136         outdata->dsize = talloc_get_size(outdata->dptr);
 137
 138         return 0;
 139 }
 140
 141 /*
 142   reload the nodes file
 143 */
 144 int
 145 ctdb_control_reload_nodes_file(struct ctdb_context *ctdb, uint32_t opcode)
 146 {
 147         unsigned int i, num_nodes;
 148         TALLOC_CTX *tmp_ctx;
 149         struct ctdb_node **nodes;
 150
 151         tmp_ctx = talloc_new(ctdb);
 152
 153         /* steal the old nodes file for a while */
 154         talloc_steal(tmp_ctx, ctdb->nodes);
 155         nodes = ctdb->nodes;
 156         ctdb->nodes = NULL;
 157         num_nodes = ctdb->num_nodes;
 158         ctdb->num_nodes = 0;
 159
 160         /* load the new nodes file */
 161         ctdb_load_nodes_file(ctdb);
 162
 163         for (i=0; i<ctdb->num_nodes; i++) {
 164                 /* keep any identical pre-existing nodes and connections */
 165                 if ((i < num_nodes) && ctdb_same_address(&ctdb->nodes[i]->address, &nodes[i]->address)) {
 166                         talloc_free(ctdb->nodes[i]);
 167                         ctdb->nodes[i] = talloc_steal(ctdb->nodes, nodes[i]);
 168                         continue;
 169                 }
 170
 171                 if (ctdb->nodes[i]->flags & NODE_FLAGS_DELETED) {
 172                         continue;
 173                 }
 174
 175                 /* any new or different nodes must be added */
 176                 if (ctdb->methods->add_node(ctdb->nodes[i]) != 0) {
 177                         DEBUG(DEBUG_CRIT, (__location__ " methods->add_node failed at %d\n", i));
 178                         ctdb_fatal(ctdb, "failed to add node. shutting down\n");
 179                 }
 180                 if (ctdb->methods->connect_node(ctdb->nodes[i]) != 0) {
 181                         DEBUG(DEBUG_CRIT, (__location__ " methods->add_connect failed at %d\n", i));
 182                         ctdb_fatal(ctdb, "failed to connect to node. shutting down\n");
 183                 }
 184         }
 185
 186         /* tell the recovery daemon to reaload the nodes file too */
 187         ctdb_daemon_send_message(ctdb, ctdb->pnn, CTDB_SRVID_RELOAD_NODES, tdb_null);
 188
 189         talloc_free(tmp_ctx);
 190
 191         return 0;
 192 }
 193
 194 /*
 195    a traverse function for pulling all relevent records from pulldb
 196  */
 197 struct pulldb_data {
 198         struct ctdb_context *ctdb;
 199         struct ctdb_db_context *ctdb_db;
 200         struct ctdb_marshall_buffer *pulldata;
 201         uint32_t len;
 202         uint32_t allocated_len;
 203         bool failed;
 204 };
 205
 206 static int traverse_pulldb(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data, void *p)
 207 {
 208         struct pulldb_data *params = (struct pulldb_data *)p;
 209         struct ctdb_rec_data_old *rec;
 210         struct ctdb_context *ctdb = params->ctdb;
 211         struct ctdb_db_context *ctdb_db = params->ctdb_db;
 212
 213         /* add the record to the blob */
 214         rec = ctdb_marshall_record(params->pulldata, 0, key, NULL, data);
 215         if (rec == NULL) {
 216                 params->failed = true;
 217                 return -1;
 218         }
 219         if (params->len + rec->length >= params->allocated_len) {
 220                 params->allocated_len = rec->length + params->len + ctdb->tunable.pulldb_preallocation_size;
 221                 params->pulldata = talloc_realloc_size(NULL, params->pulldata, params->allocated_len);
 222         }
 223         if (params->pulldata == NULL) {
 224                 DEBUG(DEBUG_CRIT,(__location__ " Failed to expand pulldb_data to %u\n", rec->length + params->len));
 225                 ctdb_fatal(params->ctdb, "failed to allocate memory for recovery. shutting down\n");
 226         }
 227         params->pulldata->count++;
 228         memcpy(params->len+(uint8_t *)params->pulldata, rec, rec->length);
 229         params->len += rec->length;
 230
 231         if (ctdb->tunable.db_record_size_warn != 0 && rec->length > ctdb->tunable.db_record_size_warn) {
 232                 DEBUG(DEBUG_ERR,("Data record in %s is big. Record size is %d bytes\n", ctdb_db->db_name, (int)rec->length));
 233         }
 234
 235         talloc_free(rec);
 236
 237         return 0;
 238 }
 239
 240 /*
 241   pull a bunch of records from a ltdb, filtering by lmaster
 242  */
 243 int32_t ctdb_control_pull_db(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
 244 {
 245         struct ctdb_pulldb *pull;
 246         struct ctdb_db_context *ctdb_db;
 247         struct pulldb_data params;
 248         struct ctdb_marshall_buffer *reply;
 249
 250         pull = (struct ctdb_pulldb *)indata.dptr;
 251
 252         ctdb_db = find_ctdb_db(ctdb, pull->db_id);
 253         if (!ctdb_db) {
 254                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", pull->db_id));
 255                 return -1;
 256         }
 257
 258         if (!ctdb_db_frozen(ctdb_db)) {
 259                 DEBUG(DEBUG_ERR,
 260                       ("rejecting ctdb_control_pull_db when not frozen\n"));
 261                 return -1;
 262         }
 263
 264         reply = talloc_zero(outdata, struct ctdb_marshall_buffer);
 265         CTDB_NO_MEMORY(ctdb, reply);
 266
 267         reply->db_id = pull->db_id;
 268
 269         params.ctdb = ctdb;
 270         params.ctdb_db = ctdb_db;
 271         params.pulldata = reply;
 272         params.len = offsetof(struct ctdb_marshall_buffer, data);
 273         params.allocated_len = params.len;
 274         params.failed = false;
 275
 276         if (ctdb_db->unhealthy_reason) {
 277                 /* this is just a warning, as the tdb should be empty anyway */
 278                 DEBUG(DEBUG_WARNING,("db(%s) unhealty in ctdb_control_pull_db: %s\n",
 279                                      ctdb_db->db_name, ctdb_db->unhealthy_reason));
 280         }
 281
 282         /* If the records are invalid, we are done */
 283         if (ctdb_db->invalid_records) {
 284                 goto done;
 285         }
 286
 287         if (ctdb_lockdb_mark(ctdb_db) != 0) {
 288                 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entire db - failing\n"));
 289                 return -1;
 290         }
 291
 292         if (tdb_traverse_read(ctdb_db->ltdb->tdb, traverse_pulldb, &params) == -1) {
 293                 DEBUG(DEBUG_ERR,(__location__ " Failed to get traverse db '%s'\n", ctdb_db->db_name));
 294                 ctdb_lockdb_unmark(ctdb_db);
 295                 talloc_free(params.pulldata);
 296                 return -1;
 297         }
 298
 299         ctdb_lockdb_unmark(ctdb_db);
 300
 301 done:
 302         outdata->dptr = (uint8_t *)params.pulldata;
 303         outdata->dsize = params.len;
 304
 305         if (ctdb->tunable.db_record_count_warn != 0 && params.pulldata->count > ctdb->tunable.db_record_count_warn) {
 306                 DEBUG(DEBUG_ERR,("Database %s is big. Contains %d records\n", ctdb_db->db_name, params.pulldata->count));
 307         }
 308         if (ctdb->tunable.db_size_warn != 0 && outdata->dsize > ctdb->tunable.db_size_warn) {
 309                 DEBUG(DEBUG_ERR,("Database %s is big. Contains %d bytes\n", ctdb_db->db_name, (int)outdata->dsize));
 310         }
 311
 312
 313         return 0;
 314 }
 315
 316 struct db_pull_state {
 317         struct ctdb_context *ctdb;
 318         struct ctdb_db_context *ctdb_db;
 319         struct ctdb_marshall_buffer *recs;
 320         uint32_t pnn;
 321         uint64_t srvid;
 322         uint32_t num_records;
 323 };
 324
 325 static int traverse_db_pull(struct tdb_context *tdb, TDB_DATA key,
 326                             TDB_DATA data, void *private_data)
 327 {
 328         struct db_pull_state *state = (struct db_pull_state *)private_data;
 329         struct ctdb_marshall_buffer *recs;
 330
 331         recs = ctdb_marshall_add(state->ctdb, state->recs,
 332                                  state->ctdb_db->db_id, 0, key, NULL, data);
 333         if (recs == NULL) {
 334                 TALLOC_FREE(state->recs);
 335                 return -1;
 336         }
 337         state->recs = recs;
 338
 339         if (talloc_get_size(state->recs) >=
 340                         state->ctdb->tunable.rec_buffer_size_limit) {
 341                 TDB_DATA buffer;
 342                 int ret;
 343
 344                 buffer = ctdb_marshall_finish(state->recs);
 345                 ret = ctdb_daemon_send_message(state->ctdb, state->pnn,
 346                                                state->srvid, buffer);
 347                 if (ret != 0) {
 348                         TALLOC_FREE(state->recs);
 349                         return -1;
 350                 }
 351
 352                 state->num_records += state->recs->count;
 353                 TALLOC_FREE(state->recs);
 354         }
 355
 356         return 0;
 357 }
 358
 359 int32_t ctdb_control_db_pull(struct ctdb_context *ctdb,
 360                              struct ctdb_req_control_old *c,
 361                              TDB_DATA indata, TDB_DATA *outdata)
 362 {
 363         struct ctdb_pulldb_ext *pulldb_ext;
 364         struct ctdb_db_context *ctdb_db;
 365         struct db_pull_state state;
 366         int ret;
 367
 368         pulldb_ext = (struct ctdb_pulldb_ext *)indata.dptr;
 369
 370         ctdb_db = find_ctdb_db(ctdb, pulldb_ext->db_id);
 371         if (ctdb_db == NULL) {
 372                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n",
 373                                  pulldb_ext->db_id));
 374                 return -1;
 375         }
 376
 377         if (!ctdb_db_frozen(ctdb_db)) {
 378                 DEBUG(DEBUG_ERR,
 379                       ("rejecting ctdb_control_pull_db when not frozen\n"));
 380                 return -1;
 381         }
 382
 383         if (ctdb_db->unhealthy_reason) {
 384                 /* this is just a warning, as the tdb should be empty anyway */
 385                 DEBUG(DEBUG_WARNING,
 386                       ("db(%s) unhealty in ctdb_control_db_pull: %s\n",
 387                        ctdb_db->db_name, ctdb_db->unhealthy_reason));
 388         }
 389
 390         state.ctdb = ctdb;
 391         state.ctdb_db = ctdb_db;
 392         state.recs = NULL;
 393         state.pnn = c->hdr.srcnode;
 394         state.srvid = pulldb_ext->srvid;
 395         state.num_records = 0;
 396
 397         /* If the records are invalid, we are done */
 398         if (ctdb_db->invalid_records) {
 399                 goto done;
 400         }
 401
 402         if (ctdb_lockdb_mark(ctdb_db) != 0) {
 403                 DEBUG(DEBUG_ERR,
 404                       (__location__ " Failed to get lock on entire db - failing\n"));
 405                 return -1;
 406         }
 407
 408         ret = tdb_traverse_read(ctdb_db->ltdb->tdb, traverse_db_pull, &state);
 409         if (ret == -1) {
 410                 DEBUG(DEBUG_ERR,
 411                       (__location__ " Failed to get traverse db '%s'\n",
 412                        ctdb_db->db_name));
 413                 ctdb_lockdb_unmark(ctdb_db);
 414                 return -1;
 415         }
 416
 417         /* Last few records */
 418         if (state.recs != NULL) {
 419                 TDB_DATA buffer;
 420
 421                 buffer = ctdb_marshall_finish(state.recs);
 422                 ret = ctdb_daemon_send_message(state.ctdb, state.pnn,
 423                                                state.srvid, buffer);
 424                 if (ret != 0) {
 425                         TALLOC_FREE(state.recs);
 426                         ctdb_lockdb_unmark(ctdb_db);
 427                         return -1;
 428                 }
 429
 430                 state.num_records += state.recs->count;
 431                 TALLOC_FREE(state.recs);
 432         }
 433
 434         ctdb_lockdb_unmark(ctdb_db);
 435
 436 done:
 437         outdata->dptr = talloc_size(outdata, sizeof(uint32_t));
 438         if (outdata->dptr == NULL) {
 439                 DEBUG(DEBUG_ERR, (__location__ " Memory allocation error\n"));
 440                 return -1;
 441         }
 442
 443         memcpy(outdata->dptr, (uint8_t *)&state.num_records, sizeof(uint32_t));
 444         outdata->dsize = sizeof(uint32_t);
 445
 446         return 0;
 447 }
 448
 449 /*
 450   push a bunch of records into a ltdb, filtering by rsn
 451  */
 452 int32_t ctdb_control_push_db(struct ctdb_context *ctdb, TDB_DATA indata)
 453 {
 454         struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
 455         struct ctdb_db_context *ctdb_db;
 456         unsigned int i;
 457         int ret;
 458         struct ctdb_rec_data_old *rec;
 459
 460         if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
 461                 DEBUG(DEBUG_ERR,(__location__ " invalid data in pulldb reply\n"));
 462                 return -1;
 463         }
 464
 465         ctdb_db = find_ctdb_db(ctdb, reply->db_id);
 466         if (!ctdb_db) {
 467                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", reply->db_id));
 468                 return -1;
 469         }
 470
 471         if (!ctdb_db_frozen(ctdb_db)) {
 472                 DEBUG(DEBUG_ERR,
 473                       ("rejecting ctdb_control_push_db when not frozen\n"));
 474                 return -1;
 475         }
 476
 477         if (ctdb_lockdb_mark(ctdb_db) != 0) {
 478                 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entire db - failing\n"));
 479                 return -1;
 480         }
 481
 482         rec = (struct ctdb_rec_data_old *)&reply->data[0];
 483
 484         DEBUG(DEBUG_INFO,("starting push of %u records for dbid 0x%x\n",
 485                  reply->count, reply->db_id));
 486
 487         for (i=0;i<reply->count;i++) {
 488                 TDB_DATA key, data;
 489                 struct ctdb_ltdb_header *hdr;
 490
 491                 key.dptr = &rec->data[0];
 492                 key.dsize = rec->keylen;
 493                 data.dptr = &rec->data[key.dsize];
 494                 data.dsize = rec->datalen;
 495
 496                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
 497                         DEBUG(DEBUG_CRIT,(__location__ " bad ltdb record\n"));
 498                         goto failed;
 499                 }
 500                 hdr = (struct ctdb_ltdb_header *)data.dptr;
 501                 /* strip off any read only record flags. All readonly records
 502                    are revoked implicitely by a recovery
 503                 */
 504                 hdr->flags &= ~CTDB_REC_RO_FLAGS;
 505
 506                 data.dptr += sizeof(*hdr);
 507                 data.dsize -= sizeof(*hdr);
 508
 509                 ret = ctdb_ltdb_store(ctdb_db, key, hdr, data);
 510                 if (ret != 0) {
 511                         DEBUG(DEBUG_CRIT, (__location__ " Unable to store record\n"));
 512                         goto failed;
 513                 }
 514
 515                 rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
 516         }
 517
 518         DEBUG(DEBUG_DEBUG,("finished push of %u records for dbid 0x%x\n",
 519                  reply->count, reply->db_id));
 520
 521         if (ctdb_db_readonly(ctdb_db)) {
 522                 DEBUG(DEBUG_CRIT,("Clearing the tracking database for dbid 0x%x\n",
 523                                   ctdb_db->db_id));
 524                 if (tdb_wipe_all(ctdb_db->rottdb) != 0) {
 525                         DEBUG(DEBUG_ERR,("Failed to wipe tracking database for 0x%x. Dropping read-only delegation support\n", ctdb_db->db_id));
 526                         tdb_close(ctdb_db->rottdb);
 527                         ctdb_db->rottdb = NULL;
 528                         ctdb_db_reset_readonly(ctdb_db);
 529                 }
 530                 while (ctdb_db->revokechild_active != NULL) {
 531                         talloc_free(ctdb_db->revokechild_active);
 532                 }
 533         }
 534
 535         ctdb_lockdb_unmark(ctdb_db);
 536         return 0;
 537
 538 failed:
 539         ctdb_lockdb_unmark(ctdb_db);
 540         return -1;
 541 }
 542
 543 struct db_push_state {
 544         struct ctdb_context *ctdb;
 545         struct ctdb_db_context *ctdb_db;
 546         uint64_t srvid;
 547         uint32_t num_records;
 548         bool failed;
 549 };
 550
 551 static void db_push_msg_handler(uint64_t srvid, TDB_DATA indata,
 552                                 void *private_data)
 553 {
 554         struct db_push_state *state = talloc_get_type(
 555                 private_data, struct db_push_state);
 556         struct ctdb_marshall_buffer *recs;
 557         struct ctdb_rec_data_old *rec;
 558         unsigned int i;
 559         int ret;
 560
 561         if (state->failed) {
 562                 return;
 563         }
 564
 565         recs = (struct ctdb_marshall_buffer *)indata.dptr;
 566         rec = (struct ctdb_rec_data_old *)&recs->data[0];
 567
 568         DEBUG(DEBUG_INFO, ("starting push of %u records for dbid 0x%x\n",
 569                            recs->count, recs->db_id));
 570
 571         for (i=0; i<recs->count; i++) {
 572                 TDB_DATA key, data;
 573                 struct ctdb_ltdb_header *hdr;
 574
 575                 key.dptr = &rec->data[0];
 576                 key.dsize = rec->keylen;
 577                 data.dptr = &rec->data[key.dsize];
 578                 data.dsize = rec->datalen;
 579
 580                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
 581                         DEBUG(DEBUG_CRIT,(__location__ " bad ltdb record\n"));
 582                         goto failed;
 583                 }
 584
 585                 hdr = (struct ctdb_ltdb_header *)data.dptr;
 586                 /* Strip off any read only record flags.
 587                  * All readonly records are revoked implicitely by a recovery.
 588                  */
 589                 hdr->flags &= ~CTDB_REC_RO_FLAGS;
 590
 591                 data.dptr += sizeof(*hdr);
 592                 data.dsize -= sizeof(*hdr);
 593
 594                 ret = ctdb_ltdb_store(state->ctdb_db, key, hdr, data);
 595                 if (ret != 0) {
 596                         DEBUG(DEBUG_ERR,
 597                               (__location__ " Unable to store record\n"));
 598                         goto failed;
 599                 }
 600
 601                 rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
 602         }
 603
 604         DEBUG(DEBUG_DEBUG, ("finished push of %u records for dbid 0x%x\n",
 605                             recs->count, recs->db_id));
 606
 607         state->num_records += recs->count;
 608         return;
 609
 610 failed:
 611         state->failed = true;
 612 }
 613
 614 int32_t ctdb_control_db_push_start(struct ctdb_context *ctdb, TDB_DATA indata)
 615 {
 616         struct ctdb_pulldb_ext *pulldb_ext;
 617         struct ctdb_db_context *ctdb_db;
 618         struct db_push_state *state;
 619         int ret;
 620
 621         pulldb_ext = (struct ctdb_pulldb_ext *)indata.dptr;
 622
 623         ctdb_db = find_ctdb_db(ctdb, pulldb_ext->db_id);
 624         if (ctdb_db == NULL) {
 625                 DEBUG(DEBUG_ERR,
 626                       (__location__ " Unknown db 0x%08x\n", pulldb_ext->db_id));
 627                 return -1;
 628         }
 629
 630         if (!ctdb_db_frozen(ctdb_db)) {
 631                 DEBUG(DEBUG_ERR,
 632                       ("rejecting ctdb_control_db_push_start when not frozen\n"));
 633                 return -1;
 634         }
 635
 636         if (ctdb_db->push_started) {
 637                 DEBUG(DEBUG_WARNING,
 638                       (__location__ " DB push already started for %s\n",
 639                        ctdb_db->db_name));
 640
 641                 /* De-register old state */
 642                 state = (struct db_push_state *)ctdb_db->push_state;
 643                 if (state != NULL) {
 644                         srvid_deregister(ctdb->srv, state->srvid, state);
 645                         talloc_free(state);
 646                         ctdb_db->push_state = NULL;
 647                 }
 648         }
 649
 650         state = talloc_zero(ctdb_db, struct db_push_state);
 651         if (state == NULL) {
 652                 DEBUG(DEBUG_ERR, (__location__ " Memory allocation error\n"));
 653                 return -1;
 654         }
 655
 656         state->ctdb = ctdb;
 657         state->ctdb_db = ctdb_db;
 658         state->srvid = pulldb_ext->srvid;
 659         state->failed = false;
 660
 661         ret = srvid_register(ctdb->srv, state, state->srvid,
 662                              db_push_msg_handler, state);
 663         if (ret != 0) {
 664                 DEBUG(DEBUG_ERR,
 665                       (__location__ " Failed to register srvid for db push\n"));
 666                 talloc_free(state);
 667                 return -1;
 668         }
 669
 670         if (ctdb_lockdb_mark(ctdb_db) != 0) {
 671                 DEBUG(DEBUG_ERR,
 672                       (__location__ " Failed to get lock on entire db - failing\n"));
 673                 srvid_deregister(ctdb->srv, state->srvid, state);
 674                 talloc_free(state);
 675                 return -1;
 676         }
 677
 678         ctdb_db->push_started = true;
 679         ctdb_db->push_state = state;
 680
 681         return 0;
 682 }
 683
 684 int32_t ctdb_control_db_push_confirm(struct ctdb_context *ctdb,
 685                                      TDB_DATA indata, TDB_DATA *outdata)
 686 {
 687         uint32_t db_id;
 688         struct ctdb_db_context *ctdb_db;
 689         struct db_push_state *state;
 690
 691         db_id = *(uint32_t *)indata.dptr;
 692
 693         ctdb_db = find_ctdb_db(ctdb, db_id);
 694         if (ctdb_db == NULL) {
 695                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", db_id));
 696                 return -1;
 697         }
 698
 699         if (!ctdb_db_frozen(ctdb_db)) {
 700                 DEBUG(DEBUG_ERR,
 701                       ("rejecting ctdb_control_db_push_confirm when not frozen\n"));
 702                 return -1;
 703         }
 704
 705         if (!ctdb_db->push_started) {
 706                 DEBUG(DEBUG_ERR, (__location__ " DB push not started\n"));
 707                 return -1;
 708         }
 709
 710         if (ctdb_db_readonly(ctdb_db)) {
 711                 DEBUG(DEBUG_ERR,
 712                       ("Clearing the tracking database for dbid 0x%x\n",
 713                        ctdb_db->db_id));
 714                 if (tdb_wipe_all(ctdb_db->rottdb) != 0) {
 715                         DEBUG(DEBUG_ERR,
 716                               ("Failed to wipe tracking database for 0x%x."
 717                                " Dropping read-only delegation support\n",
 718                                ctdb_db->db_id));
 719                         tdb_close(ctdb_db->rottdb);
 720                         ctdb_db->rottdb = NULL;
 721                         ctdb_db_reset_readonly(ctdb_db);
 722                 }
 723
 724                 while (ctdb_db->revokechild_active != NULL) {
 725                         talloc_free(ctdb_db->revokechild_active);
 726                 }
 727         }
 728
 729         ctdb_lockdb_unmark(ctdb_db);
 730
 731         state = (struct db_push_state *)ctdb_db->push_state;
 732         if (state == NULL) {
 733                 DEBUG(DEBUG_ERR, (__location__ " Missing push db state\n"));
 734                 return -1;
 735         }
 736
 737         srvid_deregister(ctdb->srv, state->srvid, state);
 738
 739         outdata->dptr = talloc_size(outdata, sizeof(uint32_t));
 740         if (outdata->dptr == NULL) {
 741                 DEBUG(DEBUG_ERR, (__location__ " Memory allocation error\n"));
 742                 talloc_free(state);
 743                 ctdb_db->push_state = NULL;
 744                 return -1;
 745         }
 746
 747         memcpy(outdata->dptr, (uint8_t *)&state->num_records, sizeof(uint32_t));
 748         outdata->dsize = sizeof(uint32_t);
 749
 750         talloc_free(state);
 751         ctdb_db->push_started = false;
 752         ctdb_db->push_state = NULL;
 753
 754         return 0;
 755 }
 756
 757 struct set_recmode_state {
 758         struct ctdb_context *ctdb;
 759         struct ctdb_req_control_old *c;
 760 };
 761
 762 static void set_recmode_handler(char status,
 763                                 double latency,
 764                                 void *private_data)
 765 {
 766         struct set_recmode_state *state = talloc_get_type_abort(
 767                 private_data, struct set_recmode_state);
 768         int s = 0;
 769         const char *err = NULL;
 770
 771         switch (status) {
 772         case '0':
 773                 /* Mutex taken */
 774                 DEBUG(DEBUG_ERR,
 775                       ("ERROR: Daemon able to take recovery lock on \"%s\" during recovery\n",
 776                        state->ctdb->recovery_lock));
 777                 s = -1;
 778                 err = "Took recovery lock from daemon during recovery - probably a cluster filesystem lock coherence problem";
 779                 break;
 780
 781         case '1':
 782                 /* Contention */
 783                 DEBUG(DEBUG_DEBUG, (__location__ " Recovery lock check OK\n"));
 784                 state->ctdb->recovery_mode = CTDB_RECOVERY_NORMAL;
 785                 ctdb_process_deferred_attach(state->ctdb);
 786
 787                 s = 0;
 788
 789                 CTDB_UPDATE_RECLOCK_LATENCY(state->ctdb, "daemon reclock",
 790                                             reclock.ctdbd, latency);
 791                 break;
 792
 793         case '2':
 794                 /* Timeout.  Consider this a success, not a failure,
 795                  * as we failed to set the recovery lock which is what
 796                  * we wanted.  This can be caused by the cluster
 797                  * filesystem being very slow to arbitrate locks
 798                  * immediately after a node failure. */
 799                 DEBUG(DEBUG_WARNING,
 800                       (__location__
 801                        "Time out getting recovery lock, allowing recmode set anyway\n"));
 802                 state->ctdb->recovery_mode = CTDB_RECOVERY_NORMAL;
 803                 ctdb_process_deferred_attach(state->ctdb);
 804
 805                 s = 0;
 806                 break;
 807
 808         default:
 809                 DEBUG(DEBUG_ERR,
 810                       ("Unexpected error when testing recovery lock\n"));
 811                 s = -1;
 812                 err = "Unexpected error when testing recovery lock";
 813         }
 814
 815         ctdb_request_control_reply(state->ctdb, state->c, NULL, s, err);
 816         talloc_free(state);
 817 }
 818
 819 static void
 820 ctdb_drop_all_ips_event(struct tevent_context *ev, struct tevent_timer *te,
 821                         struct timeval t, void *private_data)
 822 {
 823         struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
 824
 825         DEBUG(DEBUG_ERR,(__location__ " Been in recovery mode for too long. Dropping all IPS\n"));
 826         talloc_free(ctdb->release_ips_ctx);
 827         ctdb->release_ips_ctx = NULL;
 828
 829         ctdb_release_all_ips(ctdb);
 830 }
 831
 832 /*
 833  * Set up an event to drop all public ips if we remain in recovery for too
 834  * long
 835  */
 836 int ctdb_deferred_drop_all_ips(struct ctdb_context *ctdb)
 837 {
 838         if (ctdb->release_ips_ctx != NULL) {
 839                 talloc_free(ctdb->release_ips_ctx);
 840         }
 841         ctdb->release_ips_ctx = talloc_new(ctdb);
 842         CTDB_NO_MEMORY(ctdb, ctdb->release_ips_ctx);
 843
 844         tevent_add_timer(ctdb->ev, ctdb->release_ips_ctx,
 845                          timeval_current_ofs(ctdb->tunable.recovery_drop_all_ips, 0),
 846                          ctdb_drop_all_ips_event, ctdb);
 847         return 0;
 848 }
 849
 850 /*
 851   set the recovery mode
 852  */
 853 int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
 854                                  struct ctdb_req_control_old *c,
 855                                  TDB_DATA indata, bool *async_reply,
 856                                  const char **errormsg)
 857 {
 858         uint32_t recmode = *(uint32_t *)indata.dptr;
 859         struct ctdb_db_context *ctdb_db;
 860         struct set_recmode_state *state;
 861         struct ctdb_cluster_mutex_handle *h;
 862
 863         if (recmode == ctdb->recovery_mode) {
 864                 D_INFO("Recovery mode already set to %s\n",
 865                        recmode == CTDB_RECOVERY_NORMAL ? "NORMAL" : "ACTIVE");
 866                 return 0;
 867         }
 868
 869         D_NOTICE("Recovery mode set to %s\n",
 870                  recmode == CTDB_RECOVERY_NORMAL ? "NORMAL" : "ACTIVE");
 871
 872         /* if we enter recovery but stay in recovery for too long
 873            we will eventually drop all our ip addresses
 874         */
 875         if (recmode == CTDB_RECOVERY_ACTIVE) {
 876                 if (ctdb_deferred_drop_all_ips(ctdb) != 0) {
 877                         D_ERR("Failed to set up deferred drop all ips\n");
 878                 }
 879
 880                 ctdb->recovery_mode = CTDB_RECOVERY_ACTIVE;
 881                 return 0;
 882         }
 883
 884         /* From this point: recmode == CTDB_RECOVERY_NORMAL
 885          *
 886          * Therefore, what follows is special handling when setting
 887          * recovery mode back to normal */
 888
 889         TALLOC_FREE(ctdb->release_ips_ctx);
 890
 891         for (ctdb_db = ctdb->db_list; ctdb_db != NULL; ctdb_db = ctdb_db->next) {
 892                 if (ctdb_db->generation != ctdb->vnn_map->generation) {
 893                         DEBUG(DEBUG_ERR,
 894                               ("Inconsistent DB generation %u for %s\n",
 895                                ctdb_db->generation, ctdb_db->db_name));
 896                         DEBUG(DEBUG_ERR, ("Recovery mode set to ACTIVE\n"));
 897                         return -1;
 898                 }
 899         }
 900
 901         /* force the databases to thaw */
 902         if (ctdb_db_all_frozen(ctdb)) {
 903                 ctdb_control_thaw(ctdb, false);
 904         }
 905
 906         if (ctdb->recovery_lock == NULL) {
 907                 /* Not using recovery lock file */
 908                 ctdb->recovery_mode = CTDB_RECOVERY_NORMAL;
 909                 ctdb_process_deferred_attach(ctdb);
 910                 return 0;
 911         }
 912
 913         state = talloc_zero(ctdb, struct set_recmode_state);
 914         if (state == NULL) {
 915                 DEBUG(DEBUG_ERR, (__location__ " out of memory\n"));
 916                 return -1;
 917         }
 918         state->ctdb = ctdb;
 919         state->c = NULL;
 920
 921         h = ctdb_cluster_mutex(state, ctdb, ctdb->recovery_lock, 5,
 922                                set_recmode_handler, state, NULL, NULL);
 923         if (h == NULL) {
 924                 talloc_free(state);
 925                 return -1;
 926         }
 927
 928         state->c = talloc_steal(state, c);
 929         *async_reply = true;
 930
 931         return 0;
 932 }
 933
 934
 935 /*
 936   delete a record as part of the vacuum process
 937   only delete if we are not lmaster or dmaster, and our rsn is <= the provided rsn
 938   use non-blocking locks
 939
 940   return 0 if the record was successfully deleted (i.e. it does not exist
 941   when the function returns)
 942   or !0 is the record still exists in the tdb after returning.
 943  */
 944 static int delete_tdb_record(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, struct ctdb_rec_data_old *rec)
 945 {
 946         TDB_DATA key, data, data2;
 947         struct ctdb_ltdb_header *hdr, *hdr2;
 948
 949         /* these are really internal tdb functions - but we need them here for
 950            non-blocking lock of the freelist */
 951         int tdb_lock_nonblock(struct tdb_context *tdb, int list, int ltype);
 952         int tdb_unlock(struct tdb_context *tdb, int list, int ltype);
 953
 954
 955         key.dsize = rec->keylen;
 956         key.dptr  = &rec->data[0];
 957         data.dsize = rec->datalen;
 958         data.dptr = &rec->data[rec->keylen];
 959
 960         if (ctdb_lmaster(ctdb, &key) == ctdb->pnn) {
 961                 DBG_INFO("Called delete on record where we are lmaster\n");
 962                 return -1;
 963         }
 964
 965         if (data.dsize != sizeof(struct ctdb_ltdb_header)) {
 966                 DBG_ERR("Bad record size\n");
 967                 return -1;
 968         }
 969
 970         hdr = (struct ctdb_ltdb_header *)data.dptr;
 971
 972         /* use a non-blocking lock */
 973         if (tdb_chainlock_nonblock(ctdb_db->ltdb->tdb, key) != 0) {
 974                 DBG_INFO("Failed to get non-blocking chain lock\n");
 975                 return -1;
 976         }
 977
 978         data2 = tdb_fetch(ctdb_db->ltdb->tdb, key);
 979         if (data2.dptr == NULL) {
 980                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 981                 return 0;
 982         }
 983
 984         if (data2.dsize < sizeof(struct ctdb_ltdb_header)) {
 985                 if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) == 0) {
 986                         if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
 987                                 DBG_ERR("Failed to delete corrupt record\n");
 988                         }
 989                         tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
 990                         DBG_ERR("Deleted corrupt record\n");
 991                 }
 992                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 993                 free(data2.dptr);
 994                 return 0;
 995         }
 996
 997         hdr2 = (struct ctdb_ltdb_header *)data2.dptr;
 998
 999         if (hdr2->rsn > hdr->rsn) {
1000                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1001                 DBG_INFO("Skipping record with rsn=%llu - called with rsn=%llu\n",
1002                          (unsigned long long)hdr2->rsn,
1003                          (unsigned long long)hdr->rsn);
1004                 free(data2.dptr);
1005                 return -1;
1006         }
1007
1008         /* do not allow deleting record that have readonly flags set. */
1009         if (hdr->flags & CTDB_REC_RO_FLAGS) {
1010                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1011                 DBG_INFO("Skipping record with readonly flags set\n");
1012                 free(data2.dptr);
1013                 return -1;
1014         }
1015         if (hdr2->flags & CTDB_REC_RO_FLAGS) {
1016                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1017                 DBG_INFO("Skipping record with readonly flags set locally\n");
1018                 free(data2.dptr);
1019                 return -1;
1020         }
1021
1022         if (hdr2->dmaster == ctdb->pnn) {
1023                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1024                 DBG_INFO("Attempted delete record where we are the dmaster\n");
1025                 free(data2.dptr);
1026                 return -1;
1027         }
1028
1029         if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) != 0) {
1030                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1031                 DBG_INFO("Failed to get non-blocking freelist lock\n");
1032                 free(data2.dptr);
1033                 return -1;
1034         }
1035
1036         if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
1037                 tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
1038                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1039                 DBG_INFO("Failed to delete record\n");
1040                 free(data2.dptr);
1041                 return -1;
1042         }
1043
1044         tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
1045         tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1046         free(data2.dptr);
1047         return 0;
1048 }
1049
1050
1051
1052 struct recovery_callback_state {
1053         struct ctdb_req_control_old *c;
1054 };
1055
1056
1057 /*
1058   called when the 'recovered' event script has finished
1059  */
1060 static void ctdb_end_recovery_callback(struct ctdb_context *ctdb, int status, void *p)
1061 {
1062         struct recovery_callback_state *state = talloc_get_type(p, struct recovery_callback_state);
1063
1064         CTDB_INCREMENT_STAT(ctdb, num_recoveries);
1065
1066         if (status != 0) {
1067                 DEBUG(DEBUG_ERR,(__location__ " recovered event script failed (status %d)\n", status));
1068                 if (status == -ETIMEDOUT) {
1069                         ctdb_ban_self(ctdb);
1070                 }
1071         }
1072
1073         ctdb_request_control_reply(ctdb, state->c, NULL, status, NULL);
1074         talloc_free(state);
1075
1076         gettimeofday(&ctdb->last_recovery_finished, NULL);
1077
1078         if (ctdb->runstate == CTDB_RUNSTATE_FIRST_RECOVERY) {
1079                 ctdb_set_runstate(ctdb, CTDB_RUNSTATE_STARTUP);
1080         }
1081 }
1082
1083 /*
1084   recovery has finished
1085  */
1086 int32_t ctdb_control_end_recovery(struct ctdb_context *ctdb,
1087                                 struct ctdb_req_control_old *c,
1088                                 bool *async_reply)
1089 {
1090         int ret;
1091         struct recovery_callback_state *state;
1092
1093         DEBUG(DEBUG_ERR,("Recovery has finished\n"));
1094
1095         ctdb_persistent_finish_trans3_commits(ctdb);
1096
1097         state = talloc(ctdb, struct recovery_callback_state);
1098         CTDB_NO_MEMORY(ctdb, state);
1099
1100         state->c    = c;
1101
1102         ret = ctdb_event_script_callback(ctdb, state,
1103                                          ctdb_end_recovery_callback,
1104                                          state,
1105                                          CTDB_EVENT_RECOVERED, "%s", "");
1106
1107         if (ret != 0) {
1108                 DEBUG(DEBUG_ERR,(__location__ " Failed to end recovery\n"));
1109                 talloc_free(state);
1110                 return -1;
1111         }
1112
1113         /* tell the control that we will be reply asynchronously */
1114         state->c    = talloc_steal(state, c);
1115         *async_reply = true;
1116         return 0;
1117 }
1118
1119 /*
1120   called when the 'startrecovery' event script has finished
1121  */
1122 static void ctdb_start_recovery_callback(struct ctdb_context *ctdb, int status, void *p)
1123 {
1124         struct recovery_callback_state *state = talloc_get_type(p, struct recovery_callback_state);
1125
1126         if (status != 0) {
1127                 DEBUG(DEBUG_ERR,(__location__ " startrecovery event script failed (status %d)\n", status));
1128         }
1129
1130         ctdb_request_control_reply(ctdb, state->c, NULL, status, NULL);
1131         talloc_free(state);
1132 }
1133
1134 static void run_start_recovery_event(struct ctdb_context *ctdb,
1135                                      struct recovery_callback_state *state)
1136 {
1137         int ret;
1138
1139         ret = ctdb_event_script_callback(ctdb, state,
1140                                          ctdb_start_recovery_callback,
1141                                          state,
1142                                          CTDB_EVENT_START_RECOVERY,
1143                                          "%s", "");
1144
1145         if (ret != 0) {
1146                 DEBUG(DEBUG_ERR,("Unable to run startrecovery event\n"));
1147                 ctdb_request_control_reply(ctdb, state->c, NULL, -1, NULL);
1148                 talloc_free(state);
1149                 return;
1150         }
1151
1152         return;
1153 }
1154
1155 static bool reclock_strings_equal(const char *a, const char *b)
1156 {
1157         return (a == NULL && b == NULL) ||
1158                 (a != NULL && b != NULL && strcmp(a, b) == 0);
1159 }
1160
1161 static void start_recovery_reclock_callback(struct ctdb_context *ctdb,
1162                                                 int32_t status,
1163                                                 TDB_DATA data,
1164                                                 const char *errormsg,
1165                                                 void *private_data)
1166 {
1167         struct recovery_callback_state *state = talloc_get_type_abort(
1168                 private_data, struct recovery_callback_state);
1169         const char *local = ctdb->recovery_lock;
1170         const char *remote = NULL;
1171
1172         if (status != 0) {
1173                 DEBUG(DEBUG_ERR, (__location__ " GET_RECLOCK failed\n"));
1174                 ctdb_request_control_reply(ctdb, state->c, NULL,
1175                                            status, errormsg);
1176                 talloc_free(state);
1177                 return;
1178         }
1179
1180         /* Check reclock consistency */
1181         if (data.dsize > 0) {
1182                 /* Ensure NUL-termination */
1183                 data.dptr[data.dsize-1] = '\0';
1184                 remote = (const char *)data.dptr;
1185         }
1186         if (! reclock_strings_equal(local, remote)) {
1187                 /* Inconsistent */
1188                 ctdb_request_control_reply(ctdb, state->c, NULL, -1, NULL);
1189                 DEBUG(DEBUG_ERR,
1190                       ("Recovery lock configuration inconsistent: "
1191                        "recmaster has %s, this node has %s, shutting down\n",
1192                        remote == NULL ? "NULL" : remote,
1193                        local == NULL ? "NULL" : local));
1194                 talloc_free(state);
1195                 ctdb_shutdown_sequence(ctdb, 1);
1196         }
1197         DEBUG(DEBUG_INFO,
1198               ("Recovery lock consistency check successful\n"));
1199
1200         run_start_recovery_event(ctdb, state);
1201 }
1202
1203 /* Check recovery lock consistency and run eventscripts for the
1204  * "startrecovery" event */
1205 int32_t ctdb_control_start_recovery(struct ctdb_context *ctdb,
1206                                     struct ctdb_req_control_old *c,
1207                                     bool *async_reply)
1208 {
1209         int ret;
1210         struct recovery_callback_state *state;
1211         uint32_t recmaster = c->hdr.srcnode;
1212
1213         DEBUG(DEBUG_ERR, ("Recovery has started\n"));
1214         gettimeofday(&ctdb->last_recovery_started, NULL);
1215
1216         state = talloc(ctdb, struct recovery_callback_state);
1217         CTDB_NO_MEMORY(ctdb, state);
1218
1219         state->c = c;
1220
1221         /* Although the recovery master sent this node a start
1222          * recovery control, this node might still think the recovery
1223          * master is disconnected.  In this case defer the recovery
1224          * lock consistency check. */
1225         if (ctdb->nodes[recmaster]->flags & NODE_FLAGS_DISCONNECTED) {
1226                 run_start_recovery_event(ctdb, state);
1227         } else {
1228                 /* Ask the recovery master about its reclock setting */
1229                 ret = ctdb_daemon_send_control(ctdb,
1230                                                recmaster,
1231                                                0,
1232                                                CTDB_CONTROL_GET_RECLOCK_FILE,
1233                                                0, 0,
1234                                                tdb_null,
1235                                                start_recovery_reclock_callback,
1236                                                state);
1237
1238                 if (ret != 0) {
1239                         DEBUG(DEBUG_ERR, (__location__ " GET_RECLOCK failed\n"));
1240                         talloc_free(state);
1241                         return -1;
1242                 }
1243         }
1244
1245         /* tell the control that we will be reply asynchronously */
1246         state->c = talloc_steal(state, c);
1247         *async_reply = true;
1248
1249         return 0;
1250 }
1251
1252 /*
1253  try to delete all these records as part of the vacuuming process
1254  and return the records we failed to delete
1255 */
1256 int32_t ctdb_control_try_delete_records(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
1257 {
1258         struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
1259         struct ctdb_db_context *ctdb_db;
1260         unsigned int i;
1261         struct ctdb_rec_data_old *rec;
1262         struct ctdb_marshall_buffer *records;
1263
1264         if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
1265                 DEBUG(DEBUG_ERR,(__location__ " invalid data in try_delete_records\n"));
1266                 return -1;
1267         }
1268
1269         ctdb_db = find_ctdb_db(ctdb, reply->db_id);
1270         if (!ctdb_db) {
1271                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", reply->db_id));
1272                 return -1;
1273         }
1274
1275
1276         DEBUG(DEBUG_DEBUG,("starting try_delete_records of %u records for dbid 0x%x\n",
1277                  reply->count, reply->db_id));
1278
1279
1280         /* create a blob to send back the records we couldnt delete */
1281         records = (struct ctdb_marshall_buffer *)
1282                         talloc_zero_size(outdata,
1283                                     offsetof(struct ctdb_marshall_buffer, data));
1284         if (records == NULL) {
1285                 DEBUG(DEBUG_ERR,(__location__ " Out of memory\n"));
1286                 return -1;
1287         }
1288         records->db_id = ctdb_db->db_id;
1289
1290
1291         rec = (struct ctdb_rec_data_old *)&reply->data[0];
1292         for (i=0;i<reply->count;i++) {
1293                 TDB_DATA key, data;
1294
1295                 key.dptr = &rec->data[0];
1296                 key.dsize = rec->keylen;
1297                 data.dptr = &rec->data[key.dsize];
1298                 data.dsize = rec->datalen;
1299
1300                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
1301                         DEBUG(DEBUG_CRIT,(__location__ " bad ltdb record in indata\n"));
1302                         talloc_free(records);
1303                         return -1;
1304                 }
1305
1306                 /* If we cant delete the record we must add it to the reply
1307                    so the lmaster knows it may not purge this record
1308                 */
1309                 if (delete_tdb_record(ctdb, ctdb_db, rec) != 0) {
1310                         size_t old_size;
1311                         struct ctdb_ltdb_header *hdr;
1312
1313                         hdr = (struct ctdb_ltdb_header *)data.dptr;
1314                         data.dptr += sizeof(*hdr);
1315                         data.dsize -= sizeof(*hdr);
1316
1317                         DEBUG(DEBUG_INFO, (__location__ " Failed to vacuum delete record with hash 0x%08x\n", ctdb_hash(&key)));
1318
1319                         old_size = talloc_get_size(records);
1320                         records = talloc_realloc_size(outdata, records, old_size + rec->length);
1321                         if (records == NULL) {
1322                                 DEBUG(DEBUG_ERR,(__location__ " Failed to expand\n"));
1323                                 return -1;
1324                         }
1325                         records->count++;
1326                         memcpy(old_size+(uint8_t *)records, rec, rec->length);
1327                 }
1328
1329                 rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
1330         }
1331
1332
1333         *outdata = ctdb_marshall_finish(records);
1334
1335         return 0;
1336 }
1337
1338 /*
1339   report capabilities
1340  */
1341 int32_t ctdb_control_get_capabilities(struct ctdb_context *ctdb, TDB_DATA *outdata)
1342 {
1343         uint32_t *capabilities = NULL;
1344
1345         capabilities = talloc(outdata, uint32_t);
1346         CTDB_NO_MEMORY(ctdb, capabilities);
1347         *capabilities = ctdb->capabilities;
1348
1349         outdata->dsize = sizeof(uint32_t);
1350         outdata->dptr = (uint8_t *)capabilities;
1351
1352         return 0;
1353 }
1354
1355 /* The recovery daemon will ping us at regular intervals.
1356    If we havent been pinged for a while we assume the recovery
1357    daemon is inoperable and we restart.
1358 */
1359 static void ctdb_recd_ping_timeout(struct tevent_context *ev,
1360                                    struct tevent_timer *te,
1361                                    struct timeval t, void *p)
1362 {
1363         struct ctdb_context *ctdb = talloc_get_type(p, struct ctdb_context);
1364         uint32_t *count = talloc_get_type(ctdb->recd_ping_count, uint32_t);
1365
1366         DEBUG(DEBUG_ERR, ("Recovery daemon ping timeout. Count : %u\n", *count));
1367
1368         if (*count < ctdb->tunable.recd_ping_failcount) {
1369                 (*count)++;
1370                 tevent_add_timer(ctdb->ev, ctdb->recd_ping_count,
1371                                  timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
1372                                  ctdb_recd_ping_timeout, ctdb);
1373                 return;
1374         }
1375
1376         DEBUG(DEBUG_ERR, ("Final timeout for recovery daemon ping. Restarting recovery daemon. (This can be caused if the cluster filesystem has hung)\n"));
1377
1378         ctdb_stop_recoverd(ctdb);
1379         ctdb_start_recoverd(ctdb);
1380 }
1381
1382 int32_t ctdb_control_recd_ping(struct ctdb_context *ctdb)
1383 {
1384         talloc_free(ctdb->recd_ping_count);
1385
1386         ctdb->recd_ping_count = talloc_zero(ctdb, uint32_t);
1387         CTDB_NO_MEMORY(ctdb, ctdb->recd_ping_count);
1388
1389         if (ctdb->tunable.recd_ping_timeout != 0) {
1390                 tevent_add_timer(ctdb->ev, ctdb->recd_ping_count,
1391                                  timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
1392                                  ctdb_recd_ping_timeout, ctdb);
1393         }
1394
1395         return 0;
1396 }
1397
1398
1399
1400 int32_t ctdb_control_set_recmaster(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata)
1401 {
1402         uint32_t new_recmaster;
1403
1404         CHECK_CONTROL_DATA_SIZE(sizeof(uint32_t));
1405         new_recmaster = ((uint32_t *)(&indata.dptr[0]))[0];
1406
1407         if (ctdb->pnn != new_recmaster && ctdb->recovery_master == ctdb->pnn) {
1408                 DEBUG(DEBUG_ERR,
1409                       ("Remote node (%u) is now the recovery master\n",
1410                        new_recmaster));
1411         }
1412
1413         if (ctdb->pnn == new_recmaster && ctdb->recovery_master != new_recmaster) {
1414                 DEBUG(DEBUG_ERR,
1415                       ("This node (%u) is now the recovery master\n",
1416                        ctdb->pnn));
1417         }
1418
1419         ctdb->recovery_master = new_recmaster;
1420         return 0;
1421 }
1422
1423
1424 int32_t ctdb_control_stop_node(struct ctdb_context *ctdb)
1425 {
1426         DEBUG(DEBUG_ERR, ("Stopping node\n"));
1427         ctdb->nodes[ctdb->pnn]->flags |= NODE_FLAGS_STOPPED;
1428
1429         return 0;
1430 }
1431
1432 int32_t ctdb_control_continue_node(struct ctdb_context *ctdb)
1433 {
1434         DEBUG(DEBUG_ERR, ("Continue node\n"));
1435         ctdb->nodes[ctdb->pnn]->flags &= ~NODE_FLAGS_STOPPED;
1436
1437         return 0;
1438 }
1439