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 "includes.h"
  21 #include "tdb.h"
  22 #include "system/time.h"
  23 #include "system/network.h"
  24 #include "system/filesys.h"
  25 #include "system/wait.h"
  26 #include "../include/ctdb_private.h"
  27 #include "lib/util/dlinklist.h"
  28 #include "lib/tdb_wrap/tdb_wrap.h"
  29
  30
  31 int
  32 ctdb_control_getvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  33 {
  34         struct ctdb_vnn_map_wire *map;
  35         size_t len;
  36
  37         CHECK_CONTROL_DATA_SIZE(0);
  38
  39         len = offsetof(struct ctdb_vnn_map_wire, map) + sizeof(uint32_t)*ctdb->vnn_map->size;
  40         map = talloc_size(outdata, len);
  41         CTDB_NO_MEMORY(ctdb, map);
  42
  43         map->generation = ctdb->vnn_map->generation;
  44         map->size = ctdb->vnn_map->size;
  45         memcpy(map->map, ctdb->vnn_map->map, sizeof(uint32_t)*map->size);
  46
  47         outdata->dsize = len;
  48         outdata->dptr  = (uint8_t *)map;
  49
  50         return 0;
  51 }
  52
  53 int
  54 ctdb_control_setvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  55 {
  56         struct ctdb_vnn_map_wire *map = (struct ctdb_vnn_map_wire *)indata.dptr;
  57         int i;
  58
  59         for(i=1; i<=NUM_DB_PRIORITIES; i++) {
  60                 if (ctdb->freeze_mode[i] != CTDB_FREEZE_FROZEN) {
  61                         DEBUG(DEBUG_ERR,("Attempt to set vnnmap when not frozen\n"));
  62                         return -1;
  63                 }
  64         }
  65
  66         talloc_free(ctdb->vnn_map);
  67
  68         ctdb->vnn_map = talloc(ctdb, struct ctdb_vnn_map);
  69         CTDB_NO_MEMORY(ctdb, ctdb->vnn_map);
  70
  71         ctdb->vnn_map->generation = map->generation;
  72         ctdb->vnn_map->size       = map->size;
  73         ctdb->vnn_map->map = talloc_array(ctdb->vnn_map, uint32_t, map->size);
  74         CTDB_NO_MEMORY(ctdb, ctdb->vnn_map->map);
  75
  76         memcpy(ctdb->vnn_map->map, map->map, sizeof(uint32_t)*map->size);
  77
  78         return 0;
  79 }
  80
  81 int
  82 ctdb_control_getdbmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
  83 {
  84         uint32_t i, len;
  85         struct ctdb_db_context *ctdb_db;
  86         struct ctdb_dbid_map *dbid_map;
  87
  88         CHECK_CONTROL_DATA_SIZE(0);
  89
  90         len = 0;
  91         for(ctdb_db=ctdb->db_list;ctdb_db;ctdb_db=ctdb_db->next){
  92                 len++;
  93         }
  94
  95
  96         outdata->dsize = offsetof(struct ctdb_dbid_map, dbs) + sizeof(dbid_map->dbs[0])*len;
  97         outdata->dptr  = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
  98         if (!outdata->dptr) {
  99                 DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate dbmap array\n"));
 100                 exit(1);
 101         }
 102
 103         dbid_map = (struct ctdb_dbid_map *)outdata->dptr;
 104         dbid_map->num = len;
 105         for (i=0,ctdb_db=ctdb->db_list;ctdb_db;i++,ctdb_db=ctdb_db->next){
 106                 dbid_map->dbs[i].dbid       = ctdb_db->db_id;
 107                 if (ctdb_db->persistent != 0) {
 108                         dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_PERSISTENT;
 109                 }
 110                 if (ctdb_db->readonly != 0) {
 111                         dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_READONLY;
 112                 }
 113                 if (ctdb_db->sticky != 0) {
 114                         dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_STICKY;
 115                 }
 116         }
 117
 118         return 0;
 119 }
 120
 121 int
 122 ctdb_control_getnodemap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
 123 {
 124         uint32_t i, num_nodes;
 125         struct ctdb_node_map *node_map;
 126
 127         CHECK_CONTROL_DATA_SIZE(0);
 128
 129         num_nodes = ctdb->num_nodes;
 130
 131         outdata->dsize = offsetof(struct ctdb_node_map, nodes) + num_nodes*sizeof(struct ctdb_node_and_flags);
 132         outdata->dptr  = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
 133         if (!outdata->dptr) {
 134                 DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate nodemap array\n"));
 135                 exit(1);
 136         }
 137
 138         node_map = (struct ctdb_node_map *)outdata->dptr;
 139         node_map->num = num_nodes;
 140         for (i=0; i<num_nodes; i++) {
 141                 if (parse_ip(ctdb->nodes[i]->address.address,
 142                              NULL, /* TODO: pass in the correct interface here*/
 143                              0,
 144                              &node_map->nodes[i].addr) == 0)
 145                 {
 146                         DEBUG(DEBUG_ERR, (__location__ " Failed to parse %s into a sockaddr\n", ctdb->nodes[i]->address.address));
 147                 }
 148
 149                 node_map->nodes[i].pnn   = ctdb->nodes[i]->pnn;
 150                 node_map->nodes[i].flags = ctdb->nodes[i]->flags;
 151         }
 152
 153         return 0;
 154 }
 155
 156 /*
 157    get an old style ipv4-only nodemap
 158 */
 159 int
 160 ctdb_control_getnodemapv4(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
 161 {
 162         uint32_t i, num_nodes;
 163         struct ctdb_node_mapv4 *node_map;
 164
 165         CHECK_CONTROL_DATA_SIZE(0);
 166
 167         num_nodes = ctdb->num_nodes;
 168
 169         outdata->dsize = offsetof(struct ctdb_node_mapv4, nodes) + num_nodes*sizeof(struct ctdb_node_and_flagsv4);
 170         outdata->dptr  = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
 171         if (!outdata->dptr) {
 172                 DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate nodemap array\n"));
 173                 exit(1);
 174         }
 175
 176         node_map = (struct ctdb_node_mapv4 *)outdata->dptr;
 177         node_map->num = num_nodes;
 178         for (i=0; i<num_nodes; i++) {
 179                 if (parse_ipv4(ctdb->nodes[i]->address.address, 0, &node_map->nodes[i].sin) == 0) {
 180                         DEBUG(DEBUG_ERR, (__location__ " Failed to parse %s into a sockaddr\n", ctdb->nodes[i]->address.address));
 181                         return -1;
 182                 }
 183
 184                 node_map->nodes[i].pnn   = ctdb->nodes[i]->pnn;
 185                 node_map->nodes[i].flags = ctdb->nodes[i]->flags;
 186         }
 187
 188         return 0;
 189 }
 190
 191 static void
 192 ctdb_reload_nodes_event(struct event_context *ev, struct timed_event *te,
 193                                struct timeval t, void *private_data)
 194 {
 195         int i, num_nodes;
 196         struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
 197         TALLOC_CTX *tmp_ctx;
 198         struct ctdb_node **nodes;
 199
 200         tmp_ctx = talloc_new(ctdb);
 201
 202         /* steal the old nodes file for a while */
 203         talloc_steal(tmp_ctx, ctdb->nodes);
 204         nodes = ctdb->nodes;
 205         ctdb->nodes = NULL;
 206         num_nodes = ctdb->num_nodes;
 207         ctdb->num_nodes = 0;
 208
 209         /* load the new nodes file */
 210         ctdb_load_nodes_file(ctdb);
 211
 212         for (i=0; i<ctdb->num_nodes; i++) {
 213                 /* keep any identical pre-existing nodes and connections */
 214                 if ((i < num_nodes) && ctdb_same_address(&ctdb->nodes[i]->address, &nodes[i]->address)) {
 215                         talloc_free(ctdb->nodes[i]);
 216                         ctdb->nodes[i] = talloc_steal(ctdb->nodes, nodes[i]);
 217                         continue;
 218                 }
 219
 220                 if (ctdb->nodes[i]->flags & NODE_FLAGS_DELETED) {
 221                         continue;
 222                 }
 223
 224                 /* any new or different nodes must be added */
 225                 if (ctdb->methods->add_node(ctdb->nodes[i]) != 0) {
 226                         DEBUG(DEBUG_CRIT, (__location__ " methods->add_node failed at %d\n", i));
 227                         ctdb_fatal(ctdb, "failed to add node. shutting down\n");
 228                 }
 229                 if (ctdb->methods->connect_node(ctdb->nodes[i]) != 0) {
 230                         DEBUG(DEBUG_CRIT, (__location__ " methods->add_connect failed at %d\n", i));
 231                         ctdb_fatal(ctdb, "failed to connect to node. shutting down\n");
 232                 }
 233         }
 234
 235         /* tell the recovery daemon to reaload the nodes file too */
 236         ctdb_daemon_send_message(ctdb, ctdb->pnn, CTDB_SRVID_RELOAD_NODES, tdb_null);
 237
 238         talloc_free(tmp_ctx);
 239         return;
 240 }
 241
 242 /*
 243   reload the nodes file after a short delay (so that we can send the response
 244   back first
 245 */
 246 int
 247 ctdb_control_reload_nodes_file(struct ctdb_context *ctdb, uint32_t opcode)
 248 {
 249         event_add_timed(ctdb->ev, ctdb, timeval_current_ofs(1,0), ctdb_reload_nodes_event, ctdb);
 250
 251         return 0;
 252 }
 253
 254 /*
 255    a traverse function for pulling all relevent records from pulldb
 256  */
 257 struct pulldb_data {
 258         struct ctdb_context *ctdb;
 259         struct ctdb_db_context *ctdb_db;
 260         struct ctdb_marshall_buffer *pulldata;
 261         uint32_t len;
 262         uint32_t allocated_len;
 263         bool failed;
 264 };
 265
 266 static int traverse_pulldb(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data, void *p)
 267 {
 268         struct pulldb_data *params = (struct pulldb_data *)p;
 269         struct ctdb_rec_data *rec;
 270         struct ctdb_context *ctdb = params->ctdb;
 271         struct ctdb_db_context *ctdb_db = params->ctdb_db;
 272
 273         /* add the record to the blob */
 274         rec = ctdb_marshall_record(params->pulldata, 0, key, NULL, data);
 275         if (rec == NULL) {
 276                 params->failed = true;
 277                 return -1;
 278         }
 279         if (params->len + rec->length >= params->allocated_len) {
 280                 params->allocated_len = rec->length + params->len + ctdb->tunable.pulldb_preallocation_size;
 281                 params->pulldata = talloc_realloc_size(NULL, params->pulldata, params->allocated_len);
 282         }
 283         if (params->pulldata == NULL) {
 284                 DEBUG(DEBUG_CRIT,(__location__ " Failed to expand pulldb_data to %u\n", rec->length + params->len));
 285                 ctdb_fatal(params->ctdb, "failed to allocate memory for recovery. shutting down\n");
 286         }
 287         params->pulldata->count++;
 288         memcpy(params->len+(uint8_t *)params->pulldata, rec, rec->length);
 289         params->len += rec->length;
 290
 291         if (ctdb->tunable.db_record_size_warn != 0 && rec->length > ctdb->tunable.db_record_size_warn) {
 292                 DEBUG(DEBUG_ERR,("Data record in %s is big. Record size is %d bytes\n", ctdb_db->db_name, (int)rec->length));
 293         }
 294
 295         talloc_free(rec);
 296
 297         return 0;
 298 }
 299
 300 /*
 301   pull a bunch of records from a ltdb, filtering by lmaster
 302  */
 303 int32_t ctdb_control_pull_db(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
 304 {
 305         struct ctdb_control_pulldb *pull;
 306         struct ctdb_db_context *ctdb_db;
 307         struct pulldb_data params;
 308         struct ctdb_marshall_buffer *reply;
 309
 310         pull = (struct ctdb_control_pulldb *)indata.dptr;
 311
 312         ctdb_db = find_ctdb_db(ctdb, pull->db_id);
 313         if (!ctdb_db) {
 314                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", pull->db_id));
 315                 return -1;
 316         }
 317
 318         if (ctdb->freeze_mode[ctdb_db->priority] != CTDB_FREEZE_FROZEN) {
 319                 DEBUG(DEBUG_DEBUG,("rejecting ctdb_control_pull_db when not frozen\n"));
 320                 return -1;
 321         }
 322
 323         reply = talloc_zero(outdata, struct ctdb_marshall_buffer);
 324         CTDB_NO_MEMORY(ctdb, reply);
 325
 326         reply->db_id = pull->db_id;
 327
 328         params.ctdb = ctdb;
 329         params.ctdb_db = ctdb_db;
 330         params.pulldata = reply;
 331         params.len = offsetof(struct ctdb_marshall_buffer, data);
 332         params.allocated_len = params.len;
 333         params.failed = false;
 334
 335         if (ctdb_db->unhealthy_reason) {
 336                 /* this is just a warning, as the tdb should be empty anyway */
 337                 DEBUG(DEBUG_WARNING,("db(%s) unhealty in ctdb_control_pull_db: %s\n",
 338                                      ctdb_db->db_name, ctdb_db->unhealthy_reason));
 339         }
 340
 341         if (ctdb_lockall_mark_prio(ctdb, ctdb_db->priority) != 0) {
 342                 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entired db - failing\n"));
 343                 return -1;
 344         }
 345
 346         if (tdb_traverse_read(ctdb_db->ltdb->tdb, traverse_pulldb, &params) == -1) {
 347                 DEBUG(DEBUG_ERR,(__location__ " Failed to get traverse db '%s'\n", ctdb_db->db_name));
 348                 ctdb_lockall_unmark_prio(ctdb, ctdb_db->priority);
 349                 talloc_free(params.pulldata);
 350                 return -1;
 351         }
 352
 353         ctdb_lockall_unmark_prio(ctdb, ctdb_db->priority);
 354
 355         outdata->dptr = (uint8_t *)params.pulldata;
 356         outdata->dsize = params.len;
 357
 358         if (ctdb->tunable.db_record_count_warn != 0 && params.pulldata->count > ctdb->tunable.db_record_count_warn) {
 359                 DEBUG(DEBUG_ERR,("Database %s is big. Contains %d records\n", ctdb_db->db_name, params.pulldata->count));
 360         }
 361         if (ctdb->tunable.db_size_warn != 0 && outdata->dsize > ctdb->tunable.db_size_warn) {
 362                 DEBUG(DEBUG_ERR,("Database %s is big. Contains %d bytes\n", ctdb_db->db_name, (int)outdata->dsize));
 363         }
 364
 365
 366         return 0;
 367 }
 368
 369 /*
 370   push a bunch of records into a ltdb, filtering by rsn
 371  */
 372 int32_t ctdb_control_push_db(struct ctdb_context *ctdb, TDB_DATA indata)
 373 {
 374         struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
 375         struct ctdb_db_context *ctdb_db;
 376         int i, ret;
 377         struct ctdb_rec_data *rec;
 378
 379         if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
 380                 DEBUG(DEBUG_ERR,(__location__ " invalid data in pulldb reply\n"));
 381                 return -1;
 382         }
 383
 384         ctdb_db = find_ctdb_db(ctdb, reply->db_id);
 385         if (!ctdb_db) {
 386                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", reply->db_id));
 387                 return -1;
 388         }
 389
 390         if (ctdb->freeze_mode[ctdb_db->priority] != CTDB_FREEZE_FROZEN) {
 391                 DEBUG(DEBUG_DEBUG,("rejecting ctdb_control_push_db when not frozen\n"));
 392                 return -1;
 393         }
 394
 395         if (ctdb_lockall_mark_prio(ctdb, ctdb_db->priority) != 0) {
 396                 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entired db - failing\n"));
 397                 return -1;
 398         }
 399
 400         rec = (struct ctdb_rec_data *)&reply->data[0];
 401
 402         DEBUG(DEBUG_INFO,("starting push of %u records for dbid 0x%x\n",
 403                  reply->count, reply->db_id));
 404
 405         for (i=0;i<reply->count;i++) {
 406                 TDB_DATA key, data;
 407                 struct ctdb_ltdb_header *hdr;
 408
 409                 key.dptr = &rec->data[0];
 410                 key.dsize = rec->keylen;
 411                 data.dptr = &rec->data[key.dsize];
 412                 data.dsize = rec->datalen;
 413
 414                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
 415                         DEBUG(DEBUG_CRIT,(__location__ " bad ltdb record\n"));
 416                         goto failed;
 417                 }
 418                 hdr = (struct ctdb_ltdb_header *)data.dptr;
 419                 /* strip off any read only record flags. All readonly records
 420                    are revoked implicitely by a recovery
 421                 */
 422                 hdr->flags &= ~CTDB_REC_RO_FLAGS;
 423
 424                 data.dptr += sizeof(*hdr);
 425                 data.dsize -= sizeof(*hdr);
 426
 427                 ret = ctdb_ltdb_store(ctdb_db, key, hdr, data);
 428                 if (ret != 0) {
 429                         DEBUG(DEBUG_CRIT, (__location__ " Unable to store record\n"));
 430                         goto failed;
 431                 }
 432
 433                 rec = (struct ctdb_rec_data *)(rec->length + (uint8_t *)rec);
 434         }
 435
 436         DEBUG(DEBUG_DEBUG,("finished push of %u records for dbid 0x%x\n",
 437                  reply->count, reply->db_id));
 438
 439         if (ctdb_db->readonly) {
 440                 DEBUG(DEBUG_CRIT,("Clearing the tracking database for dbid 0x%x\n",
 441                                   ctdb_db->db_id));
 442                 if (tdb_wipe_all(ctdb_db->rottdb) != 0) {
 443                         DEBUG(DEBUG_ERR,("Failed to wipe tracking database for 0x%x. Dropping read-only delegation support\n", ctdb_db->db_id));
 444                         ctdb_db->readonly = false;
 445                         tdb_close(ctdb_db->rottdb);
 446                         ctdb_db->rottdb = NULL;
 447                         ctdb_db->readonly = false;
 448                 }
 449                 while (ctdb_db->revokechild_active != NULL) {
 450                         talloc_free(ctdb_db->revokechild_active);
 451                 }
 452         }
 453
 454         ctdb_lockall_unmark_prio(ctdb, ctdb_db->priority);
 455         return 0;
 456
 457 failed:
 458         ctdb_lockall_unmark_prio(ctdb, ctdb_db->priority);
 459         return -1;
 460 }
 461
 462 struct ctdb_set_recmode_state {
 463         struct ctdb_context *ctdb;
 464         struct ctdb_req_control *c;
 465         uint32_t recmode;
 466         int fd[2];
 467         struct timed_event *te;
 468         struct fd_event *fde;
 469         pid_t child;
 470         struct timeval start_time;
 471 };
 472
 473 /*
 474   called if our set_recmode child times out. this would happen if
 475   ctdb_recovery_lock() would block.
 476  */
 477 static void ctdb_set_recmode_timeout(struct event_context *ev, struct timed_event *te,
 478                                          struct timeval t, void *private_data)
 479 {
 480         struct ctdb_set_recmode_state *state = talloc_get_type(private_data,
 481                                            struct ctdb_set_recmode_state);
 482
 483         /* we consider this a success, not a failure, as we failed to
 484            set the recovery lock which is what we wanted.  This can be
 485            caused by the cluster filesystem being very slow to
 486            arbitrate locks immediately after a node failure.
 487          */
 488         DEBUG(DEBUG_ERR,(__location__ " set_recmode child process hung/timedout CFS slow to grant locks? (allowing recmode set anyway)\n"));
 489         state->ctdb->recovery_mode = state->recmode;
 490         ctdb_request_control_reply(state->ctdb, state->c, NULL, 0, NULL);
 491         talloc_free(state);
 492 }
 493
 494
 495 /* when we free the recmode state we must kill any child process.
 496 */
 497 static int set_recmode_destructor(struct ctdb_set_recmode_state *state)
 498 {
 499         double l = timeval_elapsed(&state->start_time);
 500
 501         CTDB_UPDATE_RECLOCK_LATENCY(state->ctdb, "daemon reclock", reclock.ctdbd, l);
 502
 503         if (state->fd[0] != -1) {
 504                 state->fd[0] = -1;
 505         }
 506         if (state->fd[1] != -1) {
 507                 state->fd[1] = -1;
 508         }
 509         ctdb_kill(state->ctdb, state->child, SIGKILL);
 510         return 0;
 511 }
 512
 513 /* this is called when the client process has completed ctdb_recovery_lock()
 514    and has written data back to us through the pipe.
 515 */
 516 static void set_recmode_handler(struct event_context *ev, struct fd_event *fde,
 517                              uint16_t flags, void *private_data)
 518 {
 519         struct ctdb_set_recmode_state *state= talloc_get_type(private_data,
 520                                              struct ctdb_set_recmode_state);
 521         char c = 0;
 522         int ret;
 523
 524         /* we got a response from our child process so we can abort the
 525            timeout.
 526         */
 527         talloc_free(state->te);
 528         state->te = NULL;
 529
 530
 531         /* read the childs status when trying to lock the reclock file.
 532            child wrote 0 if everything is fine and 1 if it did manage
 533            to lock the file, which would be a problem since that means
 534            we got a request to exit from recovery but we could still lock
 535            the file   which at this time SHOULD be locked by the recovery
 536            daemon on the recmaster
 537         */
 538         ret = sys_read(state->fd[0], &c, 1);
 539         if (ret != 1 || c != 0) {
 540                 ctdb_request_control_reply(state->ctdb, state->c, NULL, -1, "managed to lock reclock file from inside daemon");
 541                 talloc_free(state);
 542                 return;
 543         }
 544
 545         state->ctdb->recovery_mode = state->recmode;
 546
 547         /* release any deferred attach calls from clients */
 548         if (state->recmode == CTDB_RECOVERY_NORMAL) {
 549                 ctdb_process_deferred_attach(state->ctdb);
 550         }
 551
 552         ctdb_request_control_reply(state->ctdb, state->c, NULL, 0, NULL);
 553         talloc_free(state);
 554         return;
 555 }
 556
 557 static void
 558 ctdb_drop_all_ips_event(struct event_context *ev, struct timed_event *te,
 559                                struct timeval t, void *private_data)
 560 {
 561         struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
 562
 563         DEBUG(DEBUG_ERR,(__location__ " Been in recovery mode for too long. Dropping all IPS\n"));
 564         talloc_free(ctdb->release_ips_ctx);
 565         ctdb->release_ips_ctx = NULL;
 566
 567         ctdb_release_all_ips(ctdb);
 568 }
 569
 570 /*
 571  * Set up an event to drop all public ips if we remain in recovery for too
 572  * long
 573  */
 574 int ctdb_deferred_drop_all_ips(struct ctdb_context *ctdb)
 575 {
 576         if (ctdb->release_ips_ctx != NULL) {
 577                 talloc_free(ctdb->release_ips_ctx);
 578         }
 579         ctdb->release_ips_ctx = talloc_new(ctdb);
 580         CTDB_NO_MEMORY(ctdb, ctdb->release_ips_ctx);
 581
 582         event_add_timed(ctdb->ev, ctdb->release_ips_ctx, timeval_current_ofs(ctdb->tunable.recovery_drop_all_ips, 0), ctdb_drop_all_ips_event, ctdb);
 583         return 0;
 584 }
 585
 586 /*
 587   set the recovery mode
 588  */
 589 int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
 590                                  struct ctdb_req_control *c,
 591                                  TDB_DATA indata, bool *async_reply,
 592                                  const char **errormsg)
 593 {
 594         uint32_t recmode = *(uint32_t *)indata.dptr;
 595         int i, ret;
 596         struct ctdb_set_recmode_state *state;
 597         pid_t parent = getpid();
 598
 599         /* if we enter recovery but stay in recovery for too long
 600            we will eventually drop all our ip addresses
 601         */
 602         if (recmode == CTDB_RECOVERY_NORMAL) {
 603                 talloc_free(ctdb->release_ips_ctx);
 604                 ctdb->release_ips_ctx = NULL;
 605         } else {
 606                 if (ctdb_deferred_drop_all_ips(ctdb) != 0) {
 607                         DEBUG(DEBUG_ERR,("Failed to set up deferred drop all ips\n"));
 608                 }
 609         }
 610
 611         if (recmode != ctdb->recovery_mode) {
 612                 DEBUG(DEBUG_NOTICE,(__location__ " Recovery mode set to %s\n",
 613                          recmode==CTDB_RECOVERY_NORMAL?"NORMAL":"ACTIVE"));
 614         }
 615
 616         if (recmode != CTDB_RECOVERY_NORMAL ||
 617             ctdb->recovery_mode != CTDB_RECOVERY_ACTIVE) {
 618                 ctdb->recovery_mode = recmode;
 619                 return 0;
 620         }
 621
 622         /* some special handling when ending recovery mode */
 623
 624         /* force the databases to thaw */
 625         for (i=1; i<=NUM_DB_PRIORITIES; i++) {
 626                 if (ctdb->freeze_handles[i] != NULL) {
 627                         ctdb_control_thaw(ctdb, i, false);
 628                 }
 629         }
 630
 631         state = talloc(ctdb, struct ctdb_set_recmode_state);
 632         CTDB_NO_MEMORY(ctdb, state);
 633
 634         state->start_time = timeval_current();
 635         state->fd[0] = -1;
 636         state->fd[1] = -1;
 637
 638         /* release any deferred attach calls from clients */
 639         if (recmode == CTDB_RECOVERY_NORMAL) {
 640                 ctdb_process_deferred_attach(ctdb);
 641         }
 642
 643         if (ctdb->tunable.verify_recovery_lock == 0) {
 644                 /* dont need to verify the reclock file */
 645                 ctdb->recovery_mode = recmode;
 646                 return 0;
 647         }
 648
 649         /* For the rest of what needs to be done, we need to do this in
 650            a child process since
 651            1, the call to ctdb_recovery_lock() can block if the cluster
 652               filesystem is in the process of recovery.
 653         */
 654         ret = pipe(state->fd);
 655         if (ret != 0) {
 656                 talloc_free(state);
 657                 DEBUG(DEBUG_CRIT,(__location__ " Failed to open pipe for set_recmode child\n"));
 658                 return -1;
 659         }
 660
 661         state->child = ctdb_fork(ctdb);
 662         if (state->child == (pid_t)-1) {
 663                 close(state->fd[0]);
 664                 close(state->fd[1]);
 665                 talloc_free(state);
 666                 return -1;
 667         }
 668
 669         if (state->child == 0) {
 670                 char cc = 0;
 671                 close(state->fd[0]);
 672
 673                 ctdb_set_process_name("ctdb_recmode");
 674                 debug_extra = talloc_asprintf(NULL, "set_recmode:");
 675                 /* we should not be able to get the lock on the reclock file,
 676                   as it should  be held by the recovery master
 677                 */
 678                 if (ctdb_recovery_lock(ctdb, false)) {
 679                         DEBUG(DEBUG_CRIT,("ERROR: recovery lock file %s not locked when recovering!\n", ctdb->recovery_lock_file));
 680                         cc = 1;
 681                 }
 682
 683                 sys_write(state->fd[1], &cc, 1);
 684                 /* make sure we die when our parent dies */
 685                 while (ctdb_kill(ctdb, parent, 0) == 0 || errno != ESRCH) {
 686                         sleep(5);
 687                         sys_write(state->fd[1], &cc, 1);
 688                 }
 689                 _exit(0);
 690         }
 691         close(state->fd[1]);
 692         set_close_on_exec(state->fd[0]);
 693
 694         state->fd[1] = -1;
 695
 696         talloc_set_destructor(state, set_recmode_destructor);
 697
 698         DEBUG(DEBUG_DEBUG, (__location__ " Created PIPE FD:%d for setrecmode\n", state->fd[0]));
 699
 700         state->te = event_add_timed(ctdb->ev, state, timeval_current_ofs(5, 0),
 701                                     ctdb_set_recmode_timeout, state);
 702
 703         state->fde = event_add_fd(ctdb->ev, state, state->fd[0],
 704                                 EVENT_FD_READ,
 705                                 set_recmode_handler,
 706                                 (void *)state);
 707
 708         if (state->fde == NULL) {
 709                 talloc_free(state);
 710                 return -1;
 711         }
 712         tevent_fd_set_auto_close(state->fde);
 713
 714         state->ctdb    = ctdb;
 715         state->recmode = recmode;
 716         state->c       = talloc_steal(state, c);
 717
 718         *async_reply = true;
 719
 720         return 0;
 721 }
 722
 723
 724 /*
 725   try and get the recovery lock in shared storage - should only work
 726   on the recovery master recovery daemon. Anywhere else is a bug
 727  */
 728 bool ctdb_recovery_lock(struct ctdb_context *ctdb, bool keep)
 729 {
 730         struct flock lock;
 731
 732         if (keep) {
 733                 DEBUG(DEBUG_ERR, ("Take the recovery lock\n"));
 734         }
 735         if (ctdb->recovery_lock_fd != -1) {
 736                 close(ctdb->recovery_lock_fd);
 737                 ctdb->recovery_lock_fd = -1;
 738         }
 739
 740         ctdb->recovery_lock_fd = open(ctdb->recovery_lock_file, O_RDWR|O_CREAT, 0600);
 741         if (ctdb->recovery_lock_fd == -1) {
 742                 DEBUG(DEBUG_ERR,("ctdb_recovery_lock: Unable to open %s - (%s)\n",
 743                          ctdb->recovery_lock_file, strerror(errno)));
 744                 return false;
 745         }
 746
 747         set_close_on_exec(ctdb->recovery_lock_fd);
 748
 749         lock.l_type = F_WRLCK;
 750         lock.l_whence = SEEK_SET;
 751         lock.l_start = 0;
 752         lock.l_len = 1;
 753         lock.l_pid = 0;
 754
 755         if (fcntl(ctdb->recovery_lock_fd, F_SETLK, &lock) != 0) {
 756                 close(ctdb->recovery_lock_fd);
 757                 ctdb->recovery_lock_fd = -1;
 758                 if (keep) {
 759                         DEBUG(DEBUG_CRIT,("ctdb_recovery_lock: Failed to get recovery lock on '%s'\n", ctdb->recovery_lock_file));
 760                 }
 761                 return false;
 762         }
 763
 764         if (!keep) {
 765                 close(ctdb->recovery_lock_fd);
 766                 ctdb->recovery_lock_fd = -1;
 767         }
 768
 769         if (keep) {
 770                 DEBUG(DEBUG_NOTICE, ("Recovery lock taken successfully\n"));
 771         }
 772
 773         DEBUG(DEBUG_NOTICE,("ctdb_recovery_lock: Got recovery lock on '%s'\n", ctdb->recovery_lock_file));
 774
 775         return true;
 776 }
 777
 778 /*
 779   delete a record as part of the vacuum process
 780   only delete if we are not lmaster or dmaster, and our rsn is <= the provided rsn
 781   use non-blocking locks
 782
 783   return 0 if the record was successfully deleted (i.e. it does not exist
 784   when the function returns)
 785   or !0 is the record still exists in the tdb after returning.
 786  */
 787 static int delete_tdb_record(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, struct ctdb_rec_data *rec)
 788 {
 789         TDB_DATA key, data, data2;
 790         struct ctdb_ltdb_header *hdr, *hdr2;
 791
 792         /* these are really internal tdb functions - but we need them here for
 793            non-blocking lock of the freelist */
 794         int tdb_lock_nonblock(struct tdb_context *tdb, int list, int ltype);
 795         int tdb_unlock(struct tdb_context *tdb, int list, int ltype);
 796
 797
 798         key.dsize = rec->keylen;
 799         key.dptr  = &rec->data[0];
 800         data.dsize = rec->datalen;
 801         data.dptr = &rec->data[rec->keylen];
 802
 803         if (ctdb_lmaster(ctdb, &key) == ctdb->pnn) {
 804                 DEBUG(DEBUG_INFO,(__location__ " Called delete on record where we are lmaster\n"));
 805                 return -1;
 806         }
 807
 808         if (data.dsize != sizeof(struct ctdb_ltdb_header)) {
 809                 DEBUG(DEBUG_ERR,(__location__ " Bad record size\n"));
 810                 return -1;
 811         }
 812
 813         hdr = (struct ctdb_ltdb_header *)data.dptr;
 814
 815         /* use a non-blocking lock */
 816         if (tdb_chainlock_nonblock(ctdb_db->ltdb->tdb, key) != 0) {
 817                 return -1;
 818         }
 819
 820         data2 = tdb_fetch(ctdb_db->ltdb->tdb, key);
 821         if (data2.dptr == NULL) {
 822                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 823                 return 0;
 824         }
 825
 826         if (data2.dsize < sizeof(struct ctdb_ltdb_header)) {
 827                 if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) == 0) {
 828                         if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
 829                                 DEBUG(DEBUG_CRIT,(__location__ " Failed to delete corrupt record\n"));
 830                         }
 831                         tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
 832                         DEBUG(DEBUG_CRIT,(__location__ " Deleted corrupt record\n"));
 833                 }
 834                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 835                 free(data2.dptr);
 836                 return 0;
 837         }
 838
 839         hdr2 = (struct ctdb_ltdb_header *)data2.dptr;
 840
 841         if (hdr2->rsn > hdr->rsn) {
 842                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 843                 DEBUG(DEBUG_INFO,(__location__ " Skipping record with rsn=%llu - called with rsn=%llu\n",
 844                          (unsigned long long)hdr2->rsn, (unsigned long long)hdr->rsn));
 845                 free(data2.dptr);
 846                 return -1;
 847         }
 848
 849         /* do not allow deleting record that have readonly flags set. */
 850         if (hdr->flags & CTDB_REC_RO_FLAGS) {
 851                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 852                 DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly flags set\n"));
 853                 free(data2.dptr);
 854                 return -1;
 855         }
 856         if (hdr2->flags & CTDB_REC_RO_FLAGS) {
 857                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 858                 DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly flags set\n"));
 859                 free(data2.dptr);
 860                 return -1;
 861         }
 862
 863         if (hdr2->dmaster == ctdb->pnn) {
 864                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 865                 DEBUG(DEBUG_INFO,(__location__ " Attempted delete record where we are the dmaster\n"));
 866                 free(data2.dptr);
 867                 return -1;
 868         }
 869
 870         if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) != 0) {
 871                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 872                 free(data2.dptr);
 873                 return -1;
 874         }
 875
 876         if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
 877                 tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
 878                 tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 879                 DEBUG(DEBUG_INFO,(__location__ " Failed to delete record\n"));
 880                 free(data2.dptr);
 881                 return -1;
 882         }
 883
 884         tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
 885         tdb_chainunlock(ctdb_db->ltdb->tdb, key);
 886         free(data2.dptr);
 887         return 0;
 888 }
 889
 890
 891
 892 struct recovery_callback_state {
 893         struct ctdb_req_control *c;
 894 };
 895
 896
 897 /*
 898   called when the 'recovered' event script has finished
 899  */
 900 static void ctdb_end_recovery_callback(struct ctdb_context *ctdb, int status, void *p)
 901 {
 902         struct recovery_callback_state *state = talloc_get_type(p, struct recovery_callback_state);
 903
 904         ctdb_enable_monitoring(ctdb);
 905         CTDB_INCREMENT_STAT(ctdb, num_recoveries);
 906
 907         if (status != 0) {
 908                 DEBUG(DEBUG_ERR,(__location__ " recovered event script failed (status %d)\n", status));
 909                 if (status == -ETIME) {
 910                         ctdb_ban_self(ctdb);
 911                 }
 912         }
 913
 914         ctdb_request_control_reply(ctdb, state->c, NULL, status, NULL);
 915         talloc_free(state);
 916
 917         gettimeofday(&ctdb->last_recovery_finished, NULL);
 918
 919         if (ctdb->runstate == CTDB_RUNSTATE_FIRST_RECOVERY) {
 920                 ctdb_set_runstate(ctdb, CTDB_RUNSTATE_STARTUP);
 921         }
 922 }
 923
 924 /*
 925   recovery has finished
 926  */
 927 int32_t ctdb_control_end_recovery(struct ctdb_context *ctdb,
 928                                 struct ctdb_req_control *c,
 929                                 bool *async_reply)
 930 {
 931         int ret;
 932         struct recovery_callback_state *state;
 933
 934         DEBUG(DEBUG_NOTICE,("Recovery has finished\n"));
 935
 936         ctdb_persistent_finish_trans3_commits(ctdb);
 937
 938         state = talloc(ctdb, struct recovery_callback_state);
 939         CTDB_NO_MEMORY(ctdb, state);
 940
 941         state->c    = c;
 942
 943         ctdb_disable_monitoring(ctdb);
 944
 945         ret = ctdb_event_script_callback(ctdb, state,
 946                                          ctdb_end_recovery_callback,
 947                                          state,
 948                                          CTDB_EVENT_RECOVERED, "%s", "");
 949
 950         if (ret != 0) {
 951                 ctdb_enable_monitoring(ctdb);
 952
 953                 DEBUG(DEBUG_ERR,(__location__ " Failed to end recovery\n"));
 954                 talloc_free(state);
 955                 return -1;
 956         }
 957
 958         /* tell the control that we will be reply asynchronously */
 959         state->c    = talloc_steal(state, c);
 960         *async_reply = true;
 961         return 0;
 962 }
 963
 964 /*
 965   called when the 'startrecovery' event script has finished
 966  */
 967 static void ctdb_start_recovery_callback(struct ctdb_context *ctdb, int status, void *p)
 968 {
 969         struct recovery_callback_state *state = talloc_get_type(p, struct recovery_callback_state);
 970
 971         if (status != 0) {
 972                 DEBUG(DEBUG_ERR,(__location__ " startrecovery event script failed (status %d)\n", status));
 973         }
 974
 975         ctdb_request_control_reply(ctdb, state->c, NULL, status, NULL);
 976         talloc_free(state);
 977 }
 978
 979 /*
 980   run the startrecovery eventscript
 981  */
 982 int32_t ctdb_control_start_recovery(struct ctdb_context *ctdb,
 983                                 struct ctdb_req_control *c,
 984                                 bool *async_reply)
 985 {
 986         int ret;
 987         struct recovery_callback_state *state;
 988
 989         DEBUG(DEBUG_NOTICE,(__location__ " startrecovery eventscript has been invoked\n"));
 990         gettimeofday(&ctdb->last_recovery_started, NULL);
 991
 992         state = talloc(ctdb, struct recovery_callback_state);
 993         CTDB_NO_MEMORY(ctdb, state);
 994
 995         state->c    = talloc_steal(state, c);
 996
 997         ctdb_disable_monitoring(ctdb);
 998
 999         ret = ctdb_event_script_callback(ctdb, state,
1000                                          ctdb_start_recovery_callback,
1001                                          state,
1002                                          CTDB_EVENT_START_RECOVERY,
1003                                          "%s", "");
1004
1005         if (ret != 0) {
1006                 DEBUG(DEBUG_ERR,(__location__ " Failed to start recovery\n"));
1007                 talloc_free(state);
1008                 return -1;
1009         }
1010
1011         /* tell the control that we will be reply asynchronously */
1012         *async_reply = true;
1013         return 0;
1014 }
1015
1016 /*
1017  try to delete all these records as part of the vacuuming process
1018  and return the records we failed to delete
1019 */
1020 int32_t ctdb_control_try_delete_records(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
1021 {
1022         struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
1023         struct ctdb_db_context *ctdb_db;
1024         int i;
1025         struct ctdb_rec_data *rec;
1026         struct ctdb_marshall_buffer *records;
1027
1028         if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
1029                 DEBUG(DEBUG_ERR,(__location__ " invalid data in try_delete_records\n"));
1030                 return -1;
1031         }
1032
1033         ctdb_db = find_ctdb_db(ctdb, reply->db_id);
1034         if (!ctdb_db) {
1035                 DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", reply->db_id));
1036                 return -1;
1037         }
1038
1039
1040         DEBUG(DEBUG_DEBUG,("starting try_delete_records of %u records for dbid 0x%x\n",
1041                  reply->count, reply->db_id));
1042
1043
1044         /* create a blob to send back the records we couldnt delete */
1045         records = (struct ctdb_marshall_buffer *)
1046                         talloc_zero_size(outdata,
1047                                     offsetof(struct ctdb_marshall_buffer, data));
1048         if (records == NULL) {
1049                 DEBUG(DEBUG_ERR,(__location__ " Out of memory\n"));
1050                 return -1;
1051         }
1052         records->db_id = ctdb_db->db_id;
1053
1054
1055         rec = (struct ctdb_rec_data *)&reply->data[0];
1056         for (i=0;i<reply->count;i++) {
1057                 TDB_DATA key, data;
1058
1059                 key.dptr = &rec->data[0];
1060                 key.dsize = rec->keylen;
1061                 data.dptr = &rec->data[key.dsize];
1062                 data.dsize = rec->datalen;
1063
1064                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
1065                         DEBUG(DEBUG_CRIT,(__location__ " bad ltdb record in indata\n"));
1066                         return -1;
1067                 }
1068
1069                 /* If we cant delete the record we must add it to the reply
1070                    so the lmaster knows it may not purge this record
1071                 */
1072                 if (delete_tdb_record(ctdb, ctdb_db, rec) != 0) {
1073                         size_t old_size;
1074                         struct ctdb_ltdb_header *hdr;
1075
1076                         hdr = (struct ctdb_ltdb_header *)data.dptr;
1077                         data.dptr += sizeof(*hdr);
1078                         data.dsize -= sizeof(*hdr);
1079
1080                         DEBUG(DEBUG_INFO, (__location__ " Failed to vacuum delete record with hash 0x%08x\n", ctdb_hash(&key)));
1081
1082                         old_size = talloc_get_size(records);
1083                         records = talloc_realloc_size(outdata, records, old_size + rec->length);
1084                         if (records == NULL) {
1085                                 DEBUG(DEBUG_ERR,(__location__ " Failed to expand\n"));
1086                                 return -1;
1087                         }
1088                         records->count++;
1089                         memcpy(old_size+(uint8_t *)records, rec, rec->length);
1090                 }
1091
1092                 rec = (struct ctdb_rec_data *)(rec->length + (uint8_t *)rec);
1093         }
1094
1095
1096         *outdata = ctdb_marshall_finish(records);
1097
1098         return 0;
1099 }
1100
1101 /**
1102  * Store a record as part of the vacuum process:
1103  * This is called from the RECEIVE_RECORD control which
1104  * the lmaster uses to send the current empty copy
1105  * to all nodes for storing, before it lets the other
1106  * nodes delete the records in the second phase with
1107  * the TRY_DELETE_RECORDS control.
1108  *
1109  * Only store if we are not lmaster or dmaster, and our
1110  * rsn is <= the provided rsn. Use non-blocking locks.
1111  *
1112  * return 0 if the record was successfully stored.
1113  * return !0 if the record still exists in the tdb after returning.
1114  */
1115 static int store_tdb_record(struct ctdb_context *ctdb,
1116                             struct ctdb_db_context *ctdb_db,
1117                             struct ctdb_rec_data *rec)
1118 {
1119         TDB_DATA key, data, data2;
1120         struct ctdb_ltdb_header *hdr, *hdr2;
1121         int ret;
1122
1123         key.dsize = rec->keylen;
1124         key.dptr = &rec->data[0];
1125         data.dsize = rec->datalen;
1126         data.dptr = &rec->data[rec->keylen];
1127
1128         if (ctdb_lmaster(ctdb, &key) == ctdb->pnn) {
1129                 DEBUG(DEBUG_INFO, (__location__ " Called store_tdb_record "
1130                                    "where we are lmaster\n"));
1131                 return -1;
1132         }
1133
1134         if (data.dsize != sizeof(struct ctdb_ltdb_header)) {
1135                 DEBUG(DEBUG_ERR, (__location__ " Bad record size\n"));
1136                 return -1;
1137         }
1138
1139         hdr = (struct ctdb_ltdb_header *)data.dptr;
1140
1141         /* use a non-blocking lock */
1142         if (tdb_chainlock_nonblock(ctdb_db->ltdb->tdb, key) != 0) {
1143                 DEBUG(DEBUG_INFO, (__location__ " Failed to lock chain in non-blocking mode\n"));
1144                 return -1;
1145         }
1146
1147         data2 = tdb_fetch(ctdb_db->ltdb->tdb, key);
1148         if (data2.dptr == NULL || data2.dsize < sizeof(struct ctdb_ltdb_header)) {
1149                 if (tdb_store(ctdb_db->ltdb->tdb, key, data, 0) == -1) {
1150                         DEBUG(DEBUG_ERR, (__location__ "Failed to store record\n"));
1151                         ret = -1;
1152                         goto done;
1153                 }
1154                 DEBUG(DEBUG_INFO, (__location__ " Stored record\n"));
1155                 ret = 0;
1156                 goto done;
1157         }
1158
1159         hdr2 = (struct ctdb_ltdb_header *)data2.dptr;
1160
1161         if (hdr2->rsn > hdr->rsn) {
1162                 DEBUG(DEBUG_INFO, (__location__ " Skipping record with "
1163                                    "rsn=%llu - called with rsn=%llu\n",
1164                                    (unsigned long long)hdr2->rsn,
1165                                    (unsigned long long)hdr->rsn));
1166                 ret = -1;
1167                 goto done;
1168         }
1169
1170         /* do not allow vacuuming of records that have readonly flags set. */
1171         if (hdr->flags & CTDB_REC_RO_FLAGS) {
1172                 DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly "
1173                                   "flags set\n"));
1174                 ret = -1;
1175                 goto done;
1176         }
1177         if (hdr2->flags & CTDB_REC_RO_FLAGS) {
1178                 DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly "
1179                                   "flags set\n"));
1180                 ret = -1;
1181                 goto done;
1182         }
1183
1184         if (hdr2->dmaster == ctdb->pnn) {
1185                 DEBUG(DEBUG_INFO, (__location__ " Attempted to store record "
1186                                    "where we are the dmaster\n"));
1187                 ret = -1;
1188                 goto done;
1189         }
1190
1191         if (tdb_store(ctdb_db->ltdb->tdb, key, data, 0) != 0) {
1192                 DEBUG(DEBUG_INFO,(__location__ " Failed to store record\n"));
1193                 ret = -1;
1194                 goto done;
1195         }
1196
1197         ret = 0;
1198
1199 done:
1200         tdb_chainunlock(ctdb_db->ltdb->tdb, key);
1201         free(data2.dptr);
1202         return  ret;
1203 }
1204
1205
1206
1207 /**
1208  * Try to store all these records as part of the vacuuming process
1209  * and return the records we failed to store.
1210  */
1211 int32_t ctdb_control_receive_records(struct ctdb_context *ctdb,
1212                                      TDB_DATA indata, TDB_DATA *outdata)
1213 {
1214         struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
1215         struct ctdb_db_context *ctdb_db;
1216         int i;
1217         struct ctdb_rec_data *rec;
1218         struct ctdb_marshall_buffer *records;
1219
1220         if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
1221                 DEBUG(DEBUG_ERR,
1222                       (__location__ " invalid data in receive_records\n"));
1223                 return -1;
1224         }
1225
1226         ctdb_db = find_ctdb_db(ctdb, reply->db_id);
1227         if (!ctdb_db) {
1228                 DEBUG(DEBUG_ERR, (__location__ " Unknown db 0x%08x\n",
1229                                   reply->db_id));
1230                 return -1;
1231         }
1232
1233         DEBUG(DEBUG_DEBUG, ("starting receive_records of %u records for "
1234                             "dbid 0x%x\n", reply->count, reply->db_id));
1235
1236         /* create a blob to send back the records we could not store */
1237         records = (struct ctdb_marshall_buffer *)
1238                         talloc_zero_size(outdata,
1239                                 offsetof(struct ctdb_marshall_buffer, data));
1240         if (records == NULL) {
1241                 DEBUG(DEBUG_ERR, (__location__ " Out of memory\n"));
1242                 return -1;
1243         }
1244         records->db_id = ctdb_db->db_id;
1245
1246         rec = (struct ctdb_rec_data *)&reply->data[0];
1247         for (i=0; i<reply->count; i++) {
1248                 TDB_DATA key, data;
1249
1250                 key.dptr = &rec->data[0];
1251                 key.dsize = rec->keylen;
1252                 data.dptr = &rec->data[key.dsize];
1253                 data.dsize = rec->datalen;
1254
1255                 if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
1256                         DEBUG(DEBUG_CRIT, (__location__ " bad ltdb record "
1257                                            "in indata\n"));
1258                         return -1;
1259                 }
1260
1261                 /*
1262                  * If we can not store the record we must add it to the reply
1263                  * so the lmaster knows it may not purge this record.
1264                  */
1265                 if (store_tdb_record(ctdb, ctdb_db, rec) != 0) {
1266                         size_t old_size;
1267                         struct ctdb_ltdb_header *hdr;
1268
1269                         hdr = (struct ctdb_ltdb_header *)data.dptr;
1270                         data.dptr += sizeof(*hdr);
1271                         data.dsize -= sizeof(*hdr);
1272
1273                         DEBUG(DEBUG_INFO, (__location__ " Failed to store "
1274                                            "record with hash 0x%08x in vacuum "
1275                                            "via RECEIVE_RECORDS\n",
1276                                            ctdb_hash(&key)));
1277
1278                         old_size = talloc_get_size(records);
1279                         records = talloc_realloc_size(outdata, records,
1280                                                       old_size + rec->length);
1281                         if (records == NULL) {
1282                                 DEBUG(DEBUG_ERR, (__location__ " Failed to "
1283                                                   "expand\n"));
1284                                 return -1;
1285                         }
1286                         records->count++;
1287                         memcpy(old_size+(uint8_t *)records, rec, rec->length);
1288                 }
1289
1290                 rec = (struct ctdb_rec_data *)(rec->length + (uint8_t *)rec);
1291         }
1292
1293         *outdata = ctdb_marshall_finish(records);
1294
1295         return 0;
1296 }
1297
1298
1299 /*
1300   report capabilities
1301  */
1302 int32_t ctdb_control_get_capabilities(struct ctdb_context *ctdb, TDB_DATA *outdata)
1303 {
1304         uint32_t *capabilities = NULL;
1305
1306         capabilities = talloc(outdata, uint32_t);
1307         CTDB_NO_MEMORY(ctdb, capabilities);
1308         *capabilities = ctdb->capabilities;
1309
1310         outdata->dsize = sizeof(uint32_t);
1311         outdata->dptr = (uint8_t *)capabilities;
1312
1313         return 0;
1314 }
1315
1316 /* The recovery daemon will ping us at regular intervals.
1317    If we havent been pinged for a while we assume the recovery
1318    daemon is inoperable and we restart.
1319 */
1320 static void ctdb_recd_ping_timeout(struct event_context *ev, struct timed_event *te, struct timeval t, void *p)
1321 {
1322         struct ctdb_context *ctdb = talloc_get_type(p, struct ctdb_context);
1323         uint32_t *count = talloc_get_type(ctdb->recd_ping_count, uint32_t);
1324
1325         DEBUG(DEBUG_ERR, ("Recovery daemon ping timeout. Count : %u\n", *count));
1326
1327         if (*count < ctdb->tunable.recd_ping_failcount) {
1328                 (*count)++;
1329                 event_add_timed(ctdb->ev, ctdb->recd_ping_count,
1330                         timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
1331                         ctdb_recd_ping_timeout, ctdb);
1332                 return;
1333         }
1334
1335         DEBUG(DEBUG_ERR, ("Final timeout for recovery daemon ping. Restarting recovery daemon. (This can be caused if the cluster filesystem has hung)\n"));
1336
1337         ctdb_stop_recoverd(ctdb);
1338         ctdb_start_recoverd(ctdb);
1339 }
1340
1341 int32_t ctdb_control_recd_ping(struct ctdb_context *ctdb)
1342 {
1343         talloc_free(ctdb->recd_ping_count);
1344
1345         ctdb->recd_ping_count = talloc_zero(ctdb, uint32_t);
1346         CTDB_NO_MEMORY(ctdb, ctdb->recd_ping_count);
1347
1348         if (ctdb->tunable.recd_ping_timeout != 0) {
1349                 event_add_timed(ctdb->ev, ctdb->recd_ping_count,
1350                         timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
1351                         ctdb_recd_ping_timeout, ctdb);
1352         }
1353
1354         return 0;
1355 }
1356
1357
1358
1359 int32_t ctdb_control_set_recmaster(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata)
1360 {
1361         uint32_t new_recmaster;
1362
1363         CHECK_CONTROL_DATA_SIZE(sizeof(uint32_t));
1364         new_recmaster = ((uint32_t *)(&indata.dptr[0]))[0];
1365
1366         if (ctdb->pnn != new_recmaster && ctdb->recovery_master == ctdb->pnn) {
1367                 DEBUG(DEBUG_NOTICE,
1368                       ("This node (%u) is no longer the recovery master\n", ctdb->pnn));
1369         }
1370
1371         if (ctdb->pnn == new_recmaster && ctdb->recovery_master != new_recmaster) {
1372                 DEBUG(DEBUG_NOTICE,
1373                       ("This node (%u) is now the recovery master\n", ctdb->pnn));
1374         }
1375
1376         ctdb->recovery_master = new_recmaster;
1377         return 0;
1378 }
1379
1380
1381 int32_t ctdb_control_stop_node(struct ctdb_context *ctdb)
1382 {
1383         DEBUG(DEBUG_NOTICE, ("Stopping node\n"));
1384         ctdb_disable_monitoring(ctdb);
1385         ctdb->nodes[ctdb->pnn]->flags |= NODE_FLAGS_STOPPED;
1386
1387         return 0;
1388 }
1389
1390 int32_t ctdb_control_continue_node(struct ctdb_context *ctdb)
1391 {
1392         DEBUG(DEBUG_NOTICE, ("Continue node\n"));
1393         ctdb->nodes[ctdb->pnn]->flags &= ~NODE_FLAGS_STOPPED;
1394
1395         return 0;
1396 }
1397