You should have received a copy of the GNU General Public License
along with this program; if not, see <http://www.gnu.org/licenses/>.
*/
-#include "includes.h"
-#include "lib/events/events.h"
-#include "lib/tdb/include/tdb.h"
+#include "replace.h"
#include "system/time.h"
#include "system/network.h"
#include "system/filesys.h"
#include "system/wait.h"
-#include "../include/ctdb_private.h"
-#include "lib/util/dlinklist.h"
-#include "db_wrap.h"
-/*
- lock all databases - mark only
- */
-static int ctdb_lock_all_databases_mark(struct ctdb_context *ctdb)
-{
- struct ctdb_db_context *ctdb_db;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_ERR,("Attempt to mark all databases locked when not frozen\n"));
- return -1;
- }
- for (ctdb_db=ctdb->db_list;ctdb_db;ctdb_db=ctdb_db->next) {
- if (tdb_lockall_mark(ctdb_db->ltdb->tdb) != 0) {
- return -1;
- }
- }
- return 0;
-}
+#include <talloc.h>
+#include <tevent.h>
+#include <tdb.h>
-/*
- lock all databases - unmark only
- */
-static int ctdb_lock_all_databases_unmark(struct ctdb_context *ctdb)
-{
- struct ctdb_db_context *ctdb_db;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_ERR,("Attempt to unmark all databases locked when not frozen\n"));
- return -1;
- }
- for (ctdb_db=ctdb->db_list;ctdb_db;ctdb_db=ctdb_db->next) {
- if (tdb_lockall_unmark(ctdb_db->ltdb->tdb) != 0) {
- return -1;
- }
- }
- return 0;
-}
+#include "lib/tdb_wrap/tdb_wrap.h"
+#include "lib/util/dlinklist.h"
+#include "lib/util/debug.h"
+#include "lib/util/samba_util.h"
+#include "lib/util/util_process.h"
+
+#include "ctdb_private.h"
+#include "ctdb_client.h"
+#include "common/system.h"
+#include "common/common.h"
+#include "common/logging.h"
int
ctdb_control_getvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
{
- CHECK_CONTROL_DATA_SIZE(0);
struct ctdb_vnn_map_wire *map;
size_t len;
+ CHECK_CONTROL_DATA_SIZE(0);
+
len = offsetof(struct ctdb_vnn_map_wire, map) + sizeof(uint32_t)*ctdb->vnn_map->size;
map = talloc_size(outdata, len);
CTDB_NO_MEMORY(ctdb, map);
return 0;
}
-int
+int
ctdb_control_setvnnmap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
{
struct ctdb_vnn_map_wire *map = (struct ctdb_vnn_map_wire *)indata.dptr;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_ERR,("Attempt to set vnnmap when not frozen\n"));
+ if (ctdb->recovery_mode != CTDB_RECOVERY_ACTIVE) {
+ DEBUG(DEBUG_ERR, ("Attempt to set vnnmap when not in recovery\n"));
return -1;
}
{
uint32_t i, len;
struct ctdb_db_context *ctdb_db;
- struct ctdb_dbid_map *dbid_map;
+ struct ctdb_dbid_map_old *dbid_map;
CHECK_CONTROL_DATA_SIZE(0);
}
- outdata->dsize = offsetof(struct ctdb_dbid_map, dbs) + sizeof(dbid_map->dbs[0])*len;
+ outdata->dsize = offsetof(struct ctdb_dbid_map_old, dbs) + sizeof(dbid_map->dbs[0])*len;
outdata->dptr = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
if (!outdata->dptr) {
DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate dbmap array\n"));
exit(1);
}
- dbid_map = (struct ctdb_dbid_map *)outdata->dptr;
+ dbid_map = (struct ctdb_dbid_map_old *)outdata->dptr;
dbid_map->num = len;
for (i=0,ctdb_db=ctdb->db_list;ctdb_db;i++,ctdb_db=ctdb_db->next){
- dbid_map->dbs[i].dbid = ctdb_db->db_id;
- dbid_map->dbs[i].persistent = ctdb_db->persistent;
+ dbid_map->dbs[i].db_id = ctdb_db->db_id;
+ if (ctdb_db->persistent != 0) {
+ dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_PERSISTENT;
+ }
+ if (ctdb_db->readonly != 0) {
+ dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_READONLY;
+ }
+ if (ctdb_db->sticky != 0) {
+ dbid_map->dbs[i].flags |= CTDB_DB_FLAGS_STICKY;
+ }
}
return 0;
}
-int
+int
ctdb_control_getnodemap(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata, TDB_DATA *outdata)
{
- uint32_t i, num_nodes;
- struct ctdb_node_map *node_map;
-
CHECK_CONTROL_DATA_SIZE(0);
- num_nodes = ctdb->num_nodes;
-
- outdata->dsize = offsetof(struct ctdb_node_map, nodes) + num_nodes*sizeof(struct ctdb_node_and_flags);
- outdata->dptr = (unsigned char *)talloc_zero_size(outdata, outdata->dsize);
- if (!outdata->dptr) {
- DEBUG(DEBUG_ALERT, (__location__ " Failed to allocate nodemap array\n"));
- exit(1);
+ outdata->dptr = (unsigned char *)ctdb_node_list_to_map(ctdb->nodes,
+ ctdb->num_nodes,
+ outdata);
+ if (outdata->dptr == NULL) {
+ return -1;
}
- node_map = (struct ctdb_node_map *)outdata->dptr;
- node_map->num = num_nodes;
- for (i=0; i<num_nodes; i++) {
- inet_aton(ctdb->nodes[i]->address.address, &node_map->nodes[i].sin.sin_addr);
- node_map->nodes[i].pnn = ctdb->nodes[i]->pnn;
- node_map->nodes[i].flags = ctdb->nodes[i]->flags;
- }
+ outdata->dsize = talloc_get_size(outdata->dptr);
return 0;
}
-static void
-ctdb_reload_nodes_event(struct event_context *ev, struct timed_event *te,
- struct timeval t, void *private_data)
+/*
+ reload the nodes file
+*/
+int
+ctdb_control_reload_nodes_file(struct ctdb_context *ctdb, uint32_t opcode)
{
- int ret;
- struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
- int ctdb_tcp_init(struct ctdb_context *);
+ int i, num_nodes;
+ TALLOC_CTX *tmp_ctx;
+ struct ctdb_node **nodes;
- /* shut down the transport */
- if (ctdb->methods != NULL) {
- ctdb->methods->shutdown(ctdb);
- }
+ tmp_ctx = talloc_new(ctdb);
+
+ /* steal the old nodes file for a while */
+ talloc_steal(tmp_ctx, ctdb->nodes);
+ nodes = ctdb->nodes;
+ ctdb->nodes = NULL;
+ num_nodes = ctdb->num_nodes;
+ ctdb->num_nodes = 0;
- /* start the transport again */
+ /* load the new nodes file */
ctdb_load_nodes_file(ctdb);
- ret = ctdb_tcp_init(ctdb);
- if (ret != 0) {
- DEBUG(DEBUG_CRIT, (__location__ " Failed to init TCP\n"));
- exit(1);
- }
- if (ctdb->methods == NULL) {
- DEBUG(DEBUG_ALERT,(__location__ " Can not restart transport. ctdb->methods==NULL\n"));
- ctdb_fatal(ctdb, "can not reinitialize transport.");
+ for (i=0; i<ctdb->num_nodes; i++) {
+ /* keep any identical pre-existing nodes and connections */
+ if ((i < num_nodes) && ctdb_same_address(&ctdb->nodes[i]->address, &nodes[i]->address)) {
+ talloc_free(ctdb->nodes[i]);
+ ctdb->nodes[i] = talloc_steal(ctdb->nodes, nodes[i]);
+ continue;
+ }
+
+ if (ctdb->nodes[i]->flags & NODE_FLAGS_DELETED) {
+ continue;
+ }
+
+ /* any new or different nodes must be added */
+ if (ctdb->methods->add_node(ctdb->nodes[i]) != 0) {
+ DEBUG(DEBUG_CRIT, (__location__ " methods->add_node failed at %d\n", i));
+ ctdb_fatal(ctdb, "failed to add node. shutting down\n");
+ }
+ if (ctdb->methods->connect_node(ctdb->nodes[i]) != 0) {
+ DEBUG(DEBUG_CRIT, (__location__ " methods->add_connect failed at %d\n", i));
+ ctdb_fatal(ctdb, "failed to connect to node. shutting down\n");
+ }
}
- ctdb->methods->initialise(ctdb);
- ctdb->methods->start(ctdb);
- return;
-}
+ /* tell the recovery daemon to reaload the nodes file too */
+ ctdb_daemon_send_message(ctdb, ctdb->pnn, CTDB_SRVID_RELOAD_NODES, tdb_null);
-/*
- reload the nodes file after a short delay (so that we can send the response
- back first
-*/
-int
-ctdb_control_reload_nodes_file(struct ctdb_context *ctdb, uint32_t opcode)
-{
- event_add_timed(ctdb->ev, ctdb, timeval_current_ofs(1,0), ctdb_reload_nodes_event, ctdb);
+ talloc_free(tmp_ctx);
return 0;
}
*/
struct pulldb_data {
struct ctdb_context *ctdb;
- struct ctdb_control_pulldb_reply *pulldata;
+ struct ctdb_db_context *ctdb_db;
+ struct ctdb_marshall_buffer *pulldata;
uint32_t len;
+ uint32_t allocated_len;
bool failed;
};
static int traverse_pulldb(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data, void *p)
{
struct pulldb_data *params = (struct pulldb_data *)p;
- struct ctdb_rec_data *rec;
+ struct ctdb_rec_data_old *rec;
+ struct ctdb_context *ctdb = params->ctdb;
+ struct ctdb_db_context *ctdb_db = params->ctdb_db;
/* add the record to the blob */
rec = ctdb_marshall_record(params->pulldata, 0, key, NULL, data);
params->failed = true;
return -1;
}
- params->pulldata = talloc_realloc_size(NULL, params->pulldata, rec->length + params->len);
+ if (params->len + rec->length >= params->allocated_len) {
+ params->allocated_len = rec->length + params->len + ctdb->tunable.pulldb_preallocation_size;
+ params->pulldata = talloc_realloc_size(NULL, params->pulldata, params->allocated_len);
+ }
if (params->pulldata == NULL) {
- DEBUG(DEBUG_ERR,(__location__ " Failed to expand pulldb_data to %u (%u records)\n",
- rec->length + params->len, params->pulldata->count));
- params->failed = true;
- return -1;
+ DEBUG(DEBUG_CRIT,(__location__ " Failed to expand pulldb_data to %u\n", rec->length + params->len));
+ ctdb_fatal(params->ctdb, "failed to allocate memory for recovery. shutting down\n");
}
params->pulldata->count++;
memcpy(params->len+(uint8_t *)params->pulldata, rec, rec->length);
params->len += rec->length;
+
+ if (ctdb->tunable.db_record_size_warn != 0 && rec->length > ctdb->tunable.db_record_size_warn) {
+ DEBUG(DEBUG_ERR,("Data record in %s is big. Record size is %d bytes\n", ctdb_db->db_name, (int)rec->length));
+ }
+
talloc_free(rec);
return 0;
}
/*
- pul a bunch of records from a ltdb, filtering by lmaster
+ pull a bunch of records from a ltdb, filtering by lmaster
*/
int32_t ctdb_control_pull_db(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
{
- struct ctdb_control_pulldb *pull;
+ struct ctdb_pulldb *pull;
struct ctdb_db_context *ctdb_db;
struct pulldb_data params;
- struct ctdb_control_pulldb_reply *reply;
+ struct ctdb_marshall_buffer *reply;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_DEBUG,("rejecting ctdb_control_pull_db when not frozen\n"));
- return -1;
- }
+ pull = (struct ctdb_pulldb *)indata.dptr;
- pull = (struct ctdb_control_pulldb *)indata.dptr;
-
ctdb_db = find_ctdb_db(ctdb, pull->db_id);
if (!ctdb_db) {
DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", pull->db_id));
return -1;
}
- reply = talloc_zero(outdata, struct ctdb_control_pulldb_reply);
+ if (!ctdb_db_frozen(ctdb_db)) {
+ DEBUG(DEBUG_ERR,
+ ("rejecting ctdb_control_pull_db when not frozen\n"));
+ return -1;
+ }
+
+ reply = talloc_zero(outdata, struct ctdb_marshall_buffer);
CTDB_NO_MEMORY(ctdb, reply);
reply->db_id = pull->db_id;
params.ctdb = ctdb;
+ params.ctdb_db = ctdb_db;
params.pulldata = reply;
- params.len = offsetof(struct ctdb_control_pulldb_reply, data);
+ params.len = offsetof(struct ctdb_marshall_buffer, data);
+ params.allocated_len = params.len;
params.failed = false;
- if (ctdb_lock_all_databases_mark(ctdb) != 0) {
- DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entired db - failing\n"));
+ if (ctdb_db->unhealthy_reason) {
+ /* this is just a warning, as the tdb should be empty anyway */
+ DEBUG(DEBUG_WARNING,("db(%s) unhealty in ctdb_control_pull_db: %s\n",
+ ctdb_db->db_name, ctdb_db->unhealthy_reason));
+ }
+
+ if (ctdb_lockdb_mark(ctdb_db) != 0) {
+ DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entire db - failing\n"));
return -1;
}
if (tdb_traverse_read(ctdb_db->ltdb->tdb, traverse_pulldb, ¶ms) == -1) {
DEBUG(DEBUG_ERR,(__location__ " Failed to get traverse db '%s'\n", ctdb_db->db_name));
- ctdb_lock_all_databases_unmark(ctdb);
+ ctdb_lockdb_unmark(ctdb_db);
talloc_free(params.pulldata);
return -1;
}
- ctdb_lock_all_databases_unmark(ctdb);
+ ctdb_lockdb_unmark(ctdb_db);
outdata->dptr = (uint8_t *)params.pulldata;
outdata->dsize = params.len;
+ if (ctdb->tunable.db_record_count_warn != 0 && params.pulldata->count > ctdb->tunable.db_record_count_warn) {
+ DEBUG(DEBUG_ERR,("Database %s is big. Contains %d records\n", ctdb_db->db_name, params.pulldata->count));
+ }
+ if (ctdb->tunable.db_size_warn != 0 && outdata->dsize > ctdb->tunable.db_size_warn) {
+ DEBUG(DEBUG_ERR,("Database %s is big. Contains %d bytes\n", ctdb_db->db_name, (int)outdata->dsize));
+ }
+
+
return 0;
}
*/
int32_t ctdb_control_push_db(struct ctdb_context *ctdb, TDB_DATA indata)
{
- struct ctdb_control_pulldb_reply *reply = (struct ctdb_control_pulldb_reply *)indata.dptr;
+ struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
struct ctdb_db_context *ctdb_db;
int i, ret;
- struct ctdb_rec_data *rec;
+ struct ctdb_rec_data_old *rec;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_DEBUG,("rejecting ctdb_control_push_db when not frozen\n"));
- return -1;
- }
-
- if (indata.dsize < offsetof(struct ctdb_control_pulldb_reply, data)) {
+ if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
DEBUG(DEBUG_ERR,(__location__ " invalid data in pulldb reply\n"));
return -1;
}
return -1;
}
- if (ctdb_lock_all_databases_mark(ctdb) != 0) {
- DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entired db - failing\n"));
+ if (!ctdb_db_frozen(ctdb_db)) {
+ DEBUG(DEBUG_ERR,
+ ("rejecting ctdb_control_push_db when not frozen\n"));
+ return -1;
+ }
+
+ if (ctdb_lockdb_mark(ctdb_db) != 0) {
+ DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entire db - failing\n"));
return -1;
}
- rec = (struct ctdb_rec_data *)&reply->data[0];
+ rec = (struct ctdb_rec_data_old *)&reply->data[0];
DEBUG(DEBUG_INFO,("starting push of %u records for dbid 0x%x\n",
reply->count, reply->db_id));
goto failed;
}
hdr = (struct ctdb_ltdb_header *)data.dptr;
+ /* strip off any read only record flags. All readonly records
+ are revoked implicitely by a recovery
+ */
+ hdr->flags &= ~CTDB_REC_RO_FLAGS;
+
data.dptr += sizeof(*hdr);
data.dsize -= sizeof(*hdr);
goto failed;
}
- rec = (struct ctdb_rec_data *)(rec->length + (uint8_t *)rec);
+ rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
}
DEBUG(DEBUG_DEBUG,("finished push of %u records for dbid 0x%x\n",
reply->count, reply->db_id));
- ctdb_lock_all_databases_unmark(ctdb);
- return 0;
-
-failed:
- ctdb_lock_all_databases_unmark(ctdb);
- return -1;
-}
-
-
-static int traverse_setdmaster(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data, void *p)
-{
- uint32_t *dmaster = (uint32_t *)p;
- struct ctdb_ltdb_header *header = (struct ctdb_ltdb_header *)data.dptr;
- int ret;
-
- /* skip if already correct */
- if (header->dmaster == *dmaster) {
- return 0;
- }
-
- header->dmaster = *dmaster;
-
- ret = tdb_store(tdb, key, data, TDB_REPLACE);
- if (ret) {
- DEBUG(DEBUG_CRIT,(__location__ " failed to write tdb data back ret:%d\n",ret));
- return ret;
+ if (ctdb_db->readonly) {
+ DEBUG(DEBUG_CRIT,("Clearing the tracking database for dbid 0x%x\n",
+ ctdb_db->db_id));
+ if (tdb_wipe_all(ctdb_db->rottdb) != 0) {
+ DEBUG(DEBUG_ERR,("Failed to wipe tracking database for 0x%x. Dropping read-only delegation support\n", ctdb_db->db_id));
+ ctdb_db->readonly = false;
+ tdb_close(ctdb_db->rottdb);
+ ctdb_db->rottdb = NULL;
+ ctdb_db->readonly = false;
+ }
+ while (ctdb_db->revokechild_active != NULL) {
+ talloc_free(ctdb_db->revokechild_active);
+ }
}
- /* TODO: add error checking here */
-
+ ctdb_lockdb_unmark(ctdb_db);
return 0;
-}
-
-int32_t ctdb_control_set_dmaster(struct ctdb_context *ctdb, TDB_DATA indata)
-{
- struct ctdb_control_set_dmaster *p = (struct ctdb_control_set_dmaster *)indata.dptr;
- struct ctdb_db_context *ctdb_db;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_DEBUG,("rejecting ctdb_control_set_dmaster when not frozen\n"));
- return -1;
- }
-
- ctdb_db = find_ctdb_db(ctdb, p->db_id);
- if (!ctdb_db) {
- DEBUG(DEBUG_ERR,(__location__ " Unknown db 0x%08x\n", p->db_id));
- return -1;
- }
-
- if (ctdb_lock_all_databases_mark(ctdb) != 0) {
- DEBUG(DEBUG_ERR,(__location__ " Failed to get lock on entired db - failing\n"));
- return -1;
- }
-
- tdb_traverse(ctdb_db->ltdb->tdb, traverse_setdmaster, &p->dmaster);
-
- ctdb_lock_all_databases_unmark(ctdb);
-
- return 0;
+failed:
+ ctdb_lockdb_unmark(ctdb_db);
+ return -1;
}
struct ctdb_set_recmode_state {
struct ctdb_context *ctdb;
- struct ctdb_req_control *c;
+ struct ctdb_req_control_old *c;
uint32_t recmode;
int fd[2];
- struct timed_event *te;
- struct fd_event *fde;
+ struct tevent_timer *te;
+ struct tevent_fd *fde;
pid_t child;
+ struct timeval start_time;
};
/*
called if our set_recmode child times out. this would happen if
ctdb_recovery_lock() would block.
*/
-static void ctdb_set_recmode_timeout(struct event_context *ev, struct timed_event *te,
- struct timeval t, void *private_data)
+static void ctdb_set_recmode_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval t, void *private_data)
{
struct ctdb_set_recmode_state *state = talloc_get_type(private_data,
struct ctdb_set_recmode_state);
- ctdb_request_control_reply(state->ctdb, state->c, NULL, -1, "timeout in ctdb_set_recmode");
+ /* we consider this a success, not a failure, as we failed to
+ set the recovery lock which is what we wanted. This can be
+ caused by the cluster filesystem being very slow to
+ arbitrate locks immediately after a node failure.
+ */
+ DEBUG(DEBUG_ERR,(__location__ " set_recmode child process hung/timedout CFS slow to grant locks? (allowing recmode set anyway)\n"));
+ state->ctdb->recovery_mode = state->recmode;
+ ctdb_request_control_reply(state->ctdb, state->c, NULL, 0, NULL);
talloc_free(state);
}
*/
static int set_recmode_destructor(struct ctdb_set_recmode_state *state)
{
- kill(state->child, SIGKILL);
+ double l = timeval_elapsed(&state->start_time);
+
+ CTDB_UPDATE_RECLOCK_LATENCY(state->ctdb, "daemon reclock", reclock.ctdbd, l);
+
+ if (state->fd[0] != -1) {
+ state->fd[0] = -1;
+ }
+ if (state->fd[1] != -1) {
+ state->fd[1] = -1;
+ }
+ ctdb_kill(state->ctdb, state->child, SIGKILL);
return 0;
}
/* this is called when the client process has completed ctdb_recovery_lock()
and has written data back to us through the pipe.
*/
-static void set_recmode_handler(struct event_context *ev, struct fd_event *fde,
- uint16_t flags, void *private_data)
+static void set_recmode_handler(struct tevent_context *ev,
+ struct tevent_fd *fde,
+ uint16_t flags, void *private_data)
{
struct ctdb_set_recmode_state *state= talloc_get_type(private_data,
struct ctdb_set_recmode_state);
state->te = NULL;
- /* read the childs status when trying to lock the reclock file.
- child wrote 0 if everything is fine and 1 if it did manage
- to lock the file, which would be a problem since that means
- we got a request to exit from recovery but we could still lock
- the file which at this time SHOULD be locked by the recovery
- daemon on the recmaster
- */
- ret = read(state->fd[0], &c, 1);
+ /* If, as expected, the child was unable to take the recovery
+ * lock then it will have written 0 into the pipe, so
+ * continue. However, any other value (e.g. 1) indicates that
+ * it was able to take the recovery lock when it should have
+ * been held by the recovery daemon on the recovery master.
+ */
+ ret = sys_read(state->fd[0], &c, 1);
if (ret != 1 || c != 0) {
- ctdb_request_control_reply(state->ctdb, state->c, NULL, -1, "managed to lock reclock file from inside daemon");
+ ctdb_request_control_reply(
+ state->ctdb, state->c, NULL, -1,
+ "Took recovery lock from daemon during recovery - probably a cluster filesystem lock coherence problem");
talloc_free(state);
return;
}
state->ctdb->recovery_mode = state->recmode;
+ /* release any deferred attach calls from clients */
+ if (state->recmode == CTDB_RECOVERY_NORMAL) {
+ ctdb_process_deferred_attach(state->ctdb);
+ }
+
ctdb_request_control_reply(state->ctdb, state->c, NULL, 0, NULL);
talloc_free(state);
return;
}
+static void
+ctdb_drop_all_ips_event(struct tevent_context *ev, struct tevent_timer *te,
+ struct timeval t, void *private_data)
+{
+ struct ctdb_context *ctdb = talloc_get_type(private_data, struct ctdb_context);
+
+ DEBUG(DEBUG_ERR,(__location__ " Been in recovery mode for too long. Dropping all IPS\n"));
+ talloc_free(ctdb->release_ips_ctx);
+ ctdb->release_ips_ctx = NULL;
+
+ ctdb_release_all_ips(ctdb);
+}
+
+/*
+ * Set up an event to drop all public ips if we remain in recovery for too
+ * long
+ */
+int ctdb_deferred_drop_all_ips(struct ctdb_context *ctdb)
+{
+ if (ctdb->release_ips_ctx != NULL) {
+ talloc_free(ctdb->release_ips_ctx);
+ }
+ ctdb->release_ips_ctx = talloc_new(ctdb);
+ CTDB_NO_MEMORY(ctdb, ctdb->release_ips_ctx);
+
+ tevent_add_timer(ctdb->ev, ctdb->release_ips_ctx,
+ timeval_current_ofs(ctdb->tunable.recovery_drop_all_ips, 0),
+ ctdb_drop_all_ips_event, ctdb);
+ return 0;
+}
+
/*
set the recovery mode
*/
int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
- struct ctdb_req_control *c,
+ struct ctdb_req_control_old *c,
TDB_DATA indata, bool *async_reply,
const char **errormsg)
{
uint32_t recmode = *(uint32_t *)indata.dptr;
- int ret;
+ int i, ret;
struct ctdb_set_recmode_state *state;
pid_t parent = getpid();
+ struct ctdb_db_context *ctdb_db;
- if (ctdb->freeze_mode != CTDB_FREEZE_FROZEN) {
- DEBUG(DEBUG_ERR,("Attempt to change recovery mode to %u when not frozen\n",
- recmode));
- (*errormsg) = "Cannot change recovery mode while not frozen";
- return -1;
+ /* if we enter recovery but stay in recovery for too long
+ we will eventually drop all our ip addresses
+ */
+ if (recmode == CTDB_RECOVERY_NORMAL) {
+ talloc_free(ctdb->release_ips_ctx);
+ ctdb->release_ips_ctx = NULL;
+ } else {
+ if (ctdb_deferred_drop_all_ips(ctdb) != 0) {
+ DEBUG(DEBUG_ERR,("Failed to set up deferred drop all ips\n"));
+ }
}
if (recmode != ctdb->recovery_mode) {
/* some special handling when ending recovery mode */
- /* force the databased to thaw */
- if (ctdb->freeze_handle) {
- ctdb_control_thaw(ctdb);
+ for (ctdb_db = ctdb->db_list; ctdb_db != NULL; ctdb_db = ctdb_db->next) {
+ if (ctdb_db->generation != ctdb->vnn_map->generation) {
+ DEBUG(DEBUG_ERR,
+ ("Inconsistent DB generation %u for %s\n",
+ ctdb_db->generation, ctdb_db->db_name));
+ DEBUG(DEBUG_ERR, ("Recovery mode set to ACTIVE\n"));
+ return -1;
+ }
+ }
+
+ /* force the databases to thaw */
+ for (i=1; i<=NUM_DB_PRIORITIES; i++) {
+ if (ctdb_db_prio_frozen(ctdb, i)) {
+ ctdb_control_thaw(ctdb, i, false);
+ }
+ }
+
+ /* release any deferred attach calls from clients */
+ if (recmode == CTDB_RECOVERY_NORMAL) {
+ ctdb_process_deferred_attach(ctdb);
+ }
+
+ if (ctdb->recovery_lock_file == NULL) {
+ /* Not using recovery lock file */
+ ctdb->recovery_mode = recmode;
+ return 0;
}
state = talloc(ctdb, struct ctdb_set_recmode_state);
CTDB_NO_MEMORY(ctdb, state);
+ state->start_time = timeval_current();
+ state->fd[0] = -1;
+ state->fd[1] = -1;
+
/* For the rest of what needs to be done, we need to do this in
a child process since
1, the call to ctdb_recovery_lock() can block if the cluster
filesystem is in the process of recovery.
- 2, running of the script may take a while.
*/
ret = pipe(state->fd);
if (ret != 0) {
return -1;
}
- state->child = fork();
+ state->child = ctdb_fork(ctdb);
if (state->child == (pid_t)-1) {
close(state->fd[0]);
close(state->fd[1]);
char cc = 0;
close(state->fd[0]);
- /* we should not be able to get the lock on the nodes list,
- as it should be held by the recovery master
- */
- if (ctdb_recovery_lock(ctdb, false)) {
- DEBUG(DEBUG_CRIT,("ERROR: recovery lock file %s not locked when recovering!\n", ctdb->recovery_lock_file));
+ prctl_set_comment("ctdb_recmode");
+ debug_extra = talloc_asprintf(NULL, "set_recmode:");
+ /* Daemon should not be able to get the recover lock,
+ * as it should be held by the recovery master */
+ if (ctdb_recovery_lock(ctdb)) {
+ DEBUG(DEBUG_ERR,
+ ("ERROR: Daemon able to take recovery lock on \"%s\" during recovery\n",
+ ctdb->recovery_lock_file));
+ ctdb_recovery_unlock(ctdb);
cc = 1;
}
- write(state->fd[1], &cc, 1);
+ sys_write(state->fd[1], &cc, 1);
/* make sure we die when our parent dies */
- while (kill(parent, 0) == 0 || errno != ESRCH) {
+ while (ctdb_kill(ctdb, parent, 0) == 0 || errno != ESRCH) {
sleep(5);
+ sys_write(state->fd[1], &cc, 1);
}
_exit(0);
}
close(state->fd[1]);
+ set_close_on_exec(state->fd[0]);
+
+ state->fd[1] = -1;
talloc_set_destructor(state, set_recmode_destructor);
- state->te = event_add_timed(ctdb->ev, state, timeval_current_ofs(3, 0),
- ctdb_set_recmode_timeout, state);
+ DEBUG(DEBUG_DEBUG, (__location__ " Created PIPE FD:%d for setrecmode\n", state->fd[0]));
+
+ state->te = tevent_add_timer(ctdb->ev, state, timeval_current_ofs(5, 0),
+ ctdb_set_recmode_timeout, state);
+
+ state->fde = tevent_add_fd(ctdb->ev, state, state->fd[0], TEVENT_FD_READ,
+ set_recmode_handler, (void *)state);
- state->fde = event_add_fd(ctdb->ev, state, state->fd[0],
- EVENT_FD_READ|EVENT_FD_AUTOCLOSE,
- set_recmode_handler,
- (void *)state);
if (state->fde == NULL) {
talloc_free(state);
return -1;
}
+ tevent_fd_set_auto_close(state->fde);
state->ctdb = ctdb;
state->recmode = recmode;
}
+bool ctdb_recovery_have_lock(struct ctdb_context *ctdb)
+{
+ return ctdb->recovery_lock_fd != -1;
+}
+
/*
try and get the recovery lock in shared storage - should only work
on the recovery master recovery daemon. Anywhere else is a bug
*/
-bool ctdb_recovery_lock(struct ctdb_context *ctdb, bool keep)
+bool ctdb_recovery_lock(struct ctdb_context *ctdb)
{
struct flock lock;
- if (ctdb->recovery_lock_fd != -1) {
- close(ctdb->recovery_lock_fd);
- }
- ctdb->recovery_lock_fd = open(ctdb->recovery_lock_file, O_RDWR|O_CREAT, 0600);
+ ctdb->recovery_lock_fd = open(ctdb->recovery_lock_file,
+ O_RDWR|O_CREAT, 0600);
if (ctdb->recovery_lock_fd == -1) {
- DEBUG(DEBUG_ERR,("ctdb_recovery_lock: Unable to open %s - (%s)\n",
- ctdb->recovery_lock_file, strerror(errno)));
+ DEBUG(DEBUG_ERR,
+ ("ctdb_recovery_lock: Unable to open %s - (%s)\n",
+ ctdb->recovery_lock_file, strerror(errno)));
return false;
}
lock.l_pid = 0;
if (fcntl(ctdb->recovery_lock_fd, F_SETLK, &lock) != 0) {
+ int saved_errno = errno;
close(ctdb->recovery_lock_fd);
ctdb->recovery_lock_fd = -1;
- if (keep) {
- DEBUG(DEBUG_CRIT,("ctdb_recovery_lock: Failed to get recovery lock on '%s'\n", ctdb->recovery_lock_file));
+ /* Fail silently on these errors, since they indicate
+ * lock contention, but log an error for any other
+ * failure. */
+ if (saved_errno != EACCES &&
+ saved_errno != EAGAIN) {
+ DEBUG(DEBUG_ERR,("ctdb_recovery_lock: Failed to get "
+ "recovery lock on '%s' - (%s)\n",
+ ctdb->recovery_lock_file,
+ strerror(saved_errno)));
}
return false;
}
- if (!keep) {
+ return true;
+}
+
+void ctdb_recovery_unlock(struct ctdb_context *ctdb)
+{
+ if (ctdb->recovery_lock_fd != -1) {
+ DEBUG(DEBUG_NOTICE, ("Releasing recovery lock\n"));
close(ctdb->recovery_lock_fd);
ctdb->recovery_lock_fd = -1;
}
-
- DEBUG(DEBUG_NOTICE,("ctdb_recovery_lock: Got recovery lock on '%s'\n", ctdb->recovery_lock_file));
-
- return true;
}
/*
when the function returns)
or !0 is the record still exists in the tdb after returning.
*/
-static int delete_tdb_record(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, struct ctdb_rec_data *rec)
+static int delete_tdb_record(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, struct ctdb_rec_data_old *rec)
{
- TDB_DATA key, data;
+ TDB_DATA key, data, data2;
struct ctdb_ltdb_header *hdr, *hdr2;
/* these are really internal tdb functions - but we need them here for
return -1;
}
- data = tdb_fetch(ctdb_db->ltdb->tdb, key);
- if (data.dptr == NULL) {
+ data2 = tdb_fetch(ctdb_db->ltdb->tdb, key);
+ if (data2.dptr == NULL) {
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
return 0;
}
- if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
+ if (data2.dsize < sizeof(struct ctdb_ltdb_header)) {
if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) == 0) {
- tdb_delete(ctdb_db->ltdb->tdb, key);
+ if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
+ DEBUG(DEBUG_CRIT,(__location__ " Failed to delete corrupt record\n"));
+ }
tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
DEBUG(DEBUG_CRIT,(__location__ " Deleted corrupt record\n"));
}
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
- free(data.dptr);
+ free(data2.dptr);
return 0;
}
- hdr2 = (struct ctdb_ltdb_header *)data.dptr;
+ hdr2 = (struct ctdb_ltdb_header *)data2.dptr;
if (hdr2->rsn > hdr->rsn) {
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
DEBUG(DEBUG_INFO,(__location__ " Skipping record with rsn=%llu - called with rsn=%llu\n",
(unsigned long long)hdr2->rsn, (unsigned long long)hdr->rsn));
- free(data.dptr);
- return -1;
+ free(data2.dptr);
+ return -1;
+ }
+
+ /* do not allow deleting record that have readonly flags set. */
+ if (hdr->flags & CTDB_REC_RO_FLAGS) {
+ tdb_chainunlock(ctdb_db->ltdb->tdb, key);
+ DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly flags set\n"));
+ free(data2.dptr);
+ return -1;
+ }
+ if (hdr2->flags & CTDB_REC_RO_FLAGS) {
+ tdb_chainunlock(ctdb_db->ltdb->tdb, key);
+ DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly flags set\n"));
+ free(data2.dptr);
+ return -1;
}
if (hdr2->dmaster == ctdb->pnn) {
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
DEBUG(DEBUG_INFO,(__location__ " Attempted delete record where we are the dmaster\n"));
- free(data.dptr);
- return -1;
+ free(data2.dptr);
+ return -1;
}
if (tdb_lock_nonblock(ctdb_db->ltdb->tdb, -1, F_WRLCK) != 0) {
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
- free(data.dptr);
- return -1;
+ free(data2.dptr);
+ return -1;
}
if (tdb_delete(ctdb_db->ltdb->tdb, key) != 0) {
tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
DEBUG(DEBUG_INFO,(__location__ " Failed to delete record\n"));
- free(data.dptr);
- return -1;
+ free(data2.dptr);
+ return -1;
}
tdb_unlock(ctdb_db->ltdb->tdb, -1, F_WRLCK);
tdb_chainunlock(ctdb_db->ltdb->tdb, key);
- free(data.dptr);
- return 0;
+ free(data2.dptr);
+ return 0;
}
struct recovery_callback_state {
- struct ctdb_req_control *c;
+ struct ctdb_req_control_old *c;
};
struct recovery_callback_state *state = talloc_get_type(p, struct recovery_callback_state);
ctdb_enable_monitoring(ctdb);
+ CTDB_INCREMENT_STAT(ctdb, num_recoveries);
if (status != 0) {
DEBUG(DEBUG_ERR,(__location__ " recovered event script failed (status %d)\n", status));
+ if (status == -ETIME) {
+ ctdb_ban_self(ctdb);
+ }
}
ctdb_request_control_reply(ctdb, state->c, NULL, status, NULL);
talloc_free(state);
gettimeofday(&ctdb->last_recovery_finished, NULL);
+
+ if (ctdb->runstate == CTDB_RUNSTATE_FIRST_RECOVERY) {
+ ctdb_set_runstate(ctdb, CTDB_RUNSTATE_STARTUP);
+ }
}
/*
recovery has finished
*/
int32_t ctdb_control_end_recovery(struct ctdb_context *ctdb,
- struct ctdb_req_control *c,
+ struct ctdb_req_control_old *c,
bool *async_reply)
{
int ret;
DEBUG(DEBUG_NOTICE,("Recovery has finished\n"));
+ ctdb_persistent_finish_trans3_commits(ctdb);
+
state = talloc(ctdb, struct recovery_callback_state);
CTDB_NO_MEMORY(ctdb, state);
- state->c = talloc_steal(state, c);
+ state->c = c;
ctdb_disable_monitoring(ctdb);
- ret = ctdb_event_script_callback(ctdb,
- timeval_current_ofs(ctdb->tunable.script_timeout, 0),
- state,
+ ret = ctdb_event_script_callback(ctdb, state,
ctdb_end_recovery_callback,
- state, "recovered");
+ state,
+ CTDB_EVENT_RECOVERED, "%s", "");
if (ret != 0) {
ctdb_enable_monitoring(ctdb);
}
/* tell the control that we will be reply asynchronously */
+ state->c = talloc_steal(state, c);
*async_reply = true;
return 0;
}
run the startrecovery eventscript
*/
int32_t ctdb_control_start_recovery(struct ctdb_context *ctdb,
- struct ctdb_req_control *c,
+ struct ctdb_req_control_old *c,
bool *async_reply)
{
int ret;
ctdb_disable_monitoring(ctdb);
- ret = ctdb_event_script_callback(ctdb,
- timeval_current_ofs(ctdb->tunable.script_timeout, 0),
- state,
+ ret = ctdb_event_script_callback(ctdb, state,
ctdb_start_recovery_callback,
- state, "startrecovery");
+ state,
+ CTDB_EVENT_START_RECOVERY,
+ "%s", "");
if (ret != 0) {
DEBUG(DEBUG_ERR,(__location__ " Failed to start recovery\n"));
return 0;
}
-/*
- report the location for the reclock file
- */
-int32_t ctdb_control_get_reclock_file(struct ctdb_context *ctdb, TDB_DATA *outdata)
-{
- char *reclock = NULL;
-
- reclock = talloc_strdup(outdata, ctdb->recovery_lock_file);
- CTDB_NO_MEMORY(ctdb, reclock);
-
- outdata->dsize = strlen(reclock)+1;
- outdata->dptr = (uint8_t *)reclock;
-
- return 0;
-}
-
/*
try to delete all these records as part of the vacuuming process
and return the records we failed to delete
*/
int32_t ctdb_control_try_delete_records(struct ctdb_context *ctdb, TDB_DATA indata, TDB_DATA *outdata)
{
- struct ctdb_control_pulldb_reply *reply = (struct ctdb_control_pulldb_reply *)indata.dptr;
+ struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
struct ctdb_db_context *ctdb_db;
int i;
- struct ctdb_rec_data *rec;
- struct ctdb_control_pulldb_reply *records;
+ struct ctdb_rec_data_old *rec;
+ struct ctdb_marshall_buffer *records;
- if (indata.dsize < offsetof(struct ctdb_control_pulldb_reply, data)) {
+ if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
DEBUG(DEBUG_ERR,(__location__ " invalid data in try_delete_records\n"));
return -1;
}
/* create a blob to send back the records we couldnt delete */
- records = (struct ctdb_control_pulldb_reply *)
+ records = (struct ctdb_marshall_buffer *)
talloc_zero_size(outdata,
- offsetof(struct ctdb_control_pulldb_reply, data));
+ offsetof(struct ctdb_marshall_buffer, data));
if (records == NULL) {
DEBUG(DEBUG_ERR,(__location__ " Out of memory\n"));
return -1;
records->db_id = ctdb_db->db_id;
- rec = (struct ctdb_rec_data *)&reply->data[0];
+ rec = (struct ctdb_rec_data_old *)&reply->data[0];
for (i=0;i<reply->count;i++) {
TDB_DATA key, data;
memcpy(old_size+(uint8_t *)records, rec, rec->length);
}
- rec = (struct ctdb_rec_data *)(rec->length + (uint8_t *)rec);
+ rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
}
- outdata->dptr = (uint8_t *)records;
- outdata->dsize = talloc_get_size(records);
+ *outdata = ctdb_marshall_finish(records);
+
+ return 0;
+}
+
+/**
+ * Store a record as part of the vacuum process:
+ * This is called from the RECEIVE_RECORD control which
+ * the lmaster uses to send the current empty copy
+ * to all nodes for storing, before it lets the other
+ * nodes delete the records in the second phase with
+ * the TRY_DELETE_RECORDS control.
+ *
+ * Only store if we are not lmaster or dmaster, and our
+ * rsn is <= the provided rsn. Use non-blocking locks.
+ *
+ * return 0 if the record was successfully stored.
+ * return !0 if the record still exists in the tdb after returning.
+ */
+static int store_tdb_record(struct ctdb_context *ctdb,
+ struct ctdb_db_context *ctdb_db,
+ struct ctdb_rec_data_old *rec)
+{
+ TDB_DATA key, data, data2;
+ struct ctdb_ltdb_header *hdr, *hdr2;
+ int ret;
+
+ key.dsize = rec->keylen;
+ key.dptr = &rec->data[0];
+ data.dsize = rec->datalen;
+ data.dptr = &rec->data[rec->keylen];
+
+ if (ctdb_lmaster(ctdb, &key) == ctdb->pnn) {
+ DEBUG(DEBUG_INFO, (__location__ " Called store_tdb_record "
+ "where we are lmaster\n"));
+ return -1;
+ }
+
+ if (data.dsize != sizeof(struct ctdb_ltdb_header)) {
+ DEBUG(DEBUG_ERR, (__location__ " Bad record size\n"));
+ return -1;
+ }
+
+ hdr = (struct ctdb_ltdb_header *)data.dptr;
+
+ /* use a non-blocking lock */
+ if (tdb_chainlock_nonblock(ctdb_db->ltdb->tdb, key) != 0) {
+ DEBUG(DEBUG_INFO, (__location__ " Failed to lock chain in non-blocking mode\n"));
+ return -1;
+ }
+
+ data2 = tdb_fetch(ctdb_db->ltdb->tdb, key);
+ if (data2.dptr == NULL || data2.dsize < sizeof(struct ctdb_ltdb_header)) {
+ if (tdb_store(ctdb_db->ltdb->tdb, key, data, 0) == -1) {
+ DEBUG(DEBUG_ERR, (__location__ "Failed to store record\n"));
+ ret = -1;
+ goto done;
+ }
+ DEBUG(DEBUG_INFO, (__location__ " Stored record\n"));
+ ret = 0;
+ goto done;
+ }
+
+ hdr2 = (struct ctdb_ltdb_header *)data2.dptr;
+
+ if (hdr2->rsn > hdr->rsn) {
+ DEBUG(DEBUG_INFO, (__location__ " Skipping record with "
+ "rsn=%llu - called with rsn=%llu\n",
+ (unsigned long long)hdr2->rsn,
+ (unsigned long long)hdr->rsn));
+ ret = -1;
+ goto done;
+ }
+
+ /* do not allow vacuuming of records that have readonly flags set. */
+ if (hdr->flags & CTDB_REC_RO_FLAGS) {
+ DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly "
+ "flags set\n"));
+ ret = -1;
+ goto done;
+ }
+ if (hdr2->flags & CTDB_REC_RO_FLAGS) {
+ DEBUG(DEBUG_INFO,(__location__ " Skipping record with readonly "
+ "flags set\n"));
+ ret = -1;
+ goto done;
+ }
+
+ if (hdr2->dmaster == ctdb->pnn) {
+ DEBUG(DEBUG_INFO, (__location__ " Attempted to store record "
+ "where we are the dmaster\n"));
+ ret = -1;
+ goto done;
+ }
+
+ if (tdb_store(ctdb_db->ltdb->tdb, key, data, 0) != 0) {
+ DEBUG(DEBUG_INFO,(__location__ " Failed to store record\n"));
+ ret = -1;
+ goto done;
+ }
+
+ ret = 0;
+
+done:
+ tdb_chainunlock(ctdb_db->ltdb->tdb, key);
+ free(data2.dptr);
+ return ret;
+}
+
+
+
+/**
+ * Try to store all these records as part of the vacuuming process
+ * and return the records we failed to store.
+ */
+int32_t ctdb_control_receive_records(struct ctdb_context *ctdb,
+ TDB_DATA indata, TDB_DATA *outdata)
+{
+ struct ctdb_marshall_buffer *reply = (struct ctdb_marshall_buffer *)indata.dptr;
+ struct ctdb_db_context *ctdb_db;
+ int i;
+ struct ctdb_rec_data_old *rec;
+ struct ctdb_marshall_buffer *records;
+
+ if (indata.dsize < offsetof(struct ctdb_marshall_buffer, data)) {
+ DEBUG(DEBUG_ERR,
+ (__location__ " invalid data in receive_records\n"));
+ return -1;
+ }
+
+ ctdb_db = find_ctdb_db(ctdb, reply->db_id);
+ if (!ctdb_db) {
+ DEBUG(DEBUG_ERR, (__location__ " Unknown db 0x%08x\n",
+ reply->db_id));
+ return -1;
+ }
+
+ DEBUG(DEBUG_DEBUG, ("starting receive_records of %u records for "
+ "dbid 0x%x\n", reply->count, reply->db_id));
+
+ /* create a blob to send back the records we could not store */
+ records = (struct ctdb_marshall_buffer *)
+ talloc_zero_size(outdata,
+ offsetof(struct ctdb_marshall_buffer, data));
+ if (records == NULL) {
+ DEBUG(DEBUG_ERR, (__location__ " Out of memory\n"));
+ return -1;
+ }
+ records->db_id = ctdb_db->db_id;
+
+ rec = (struct ctdb_rec_data_old *)&reply->data[0];
+ for (i=0; i<reply->count; i++) {
+ TDB_DATA key, data;
+
+ key.dptr = &rec->data[0];
+ key.dsize = rec->keylen;
+ data.dptr = &rec->data[key.dsize];
+ data.dsize = rec->datalen;
+
+ if (data.dsize < sizeof(struct ctdb_ltdb_header)) {
+ DEBUG(DEBUG_CRIT, (__location__ " bad ltdb record "
+ "in indata\n"));
+ return -1;
+ }
+
+ /*
+ * If we can not store the record we must add it to the reply
+ * so the lmaster knows it may not purge this record.
+ */
+ if (store_tdb_record(ctdb, ctdb_db, rec) != 0) {
+ size_t old_size;
+ struct ctdb_ltdb_header *hdr;
+
+ hdr = (struct ctdb_ltdb_header *)data.dptr;
+ data.dptr += sizeof(*hdr);
+ data.dsize -= sizeof(*hdr);
+
+ DEBUG(DEBUG_INFO, (__location__ " Failed to store "
+ "record with hash 0x%08x in vacuum "
+ "via RECEIVE_RECORDS\n",
+ ctdb_hash(&key)));
+
+ old_size = talloc_get_size(records);
+ records = talloc_realloc_size(outdata, records,
+ old_size + rec->length);
+ if (records == NULL) {
+ DEBUG(DEBUG_ERR, (__location__ " Failed to "
+ "expand\n"));
+ return -1;
+ }
+ records->count++;
+ memcpy(old_size+(uint8_t *)records, rec, rec->length);
+ }
+
+ rec = (struct ctdb_rec_data_old *)(rec->length + (uint8_t *)rec);
+ }
+
+ *outdata = ctdb_marshall_finish(records);
return 0;
}
+
/*
report capabilities
*/
return 0;
}
+/* The recovery daemon will ping us at regular intervals.
+ If we havent been pinged for a while we assume the recovery
+ daemon is inoperable and we restart.
+*/
+static void ctdb_recd_ping_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval t, void *p)
+{
+ struct ctdb_context *ctdb = talloc_get_type(p, struct ctdb_context);
+ uint32_t *count = talloc_get_type(ctdb->recd_ping_count, uint32_t);
+
+ DEBUG(DEBUG_ERR, ("Recovery daemon ping timeout. Count : %u\n", *count));
+
+ if (*count < ctdb->tunable.recd_ping_failcount) {
+ (*count)++;
+ tevent_add_timer(ctdb->ev, ctdb->recd_ping_count,
+ timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
+ ctdb_recd_ping_timeout, ctdb);
+ return;
+ }
+
+ DEBUG(DEBUG_ERR, ("Final timeout for recovery daemon ping. Restarting recovery daemon. (This can be caused if the cluster filesystem has hung)\n"));
+
+ ctdb_stop_recoverd(ctdb);
+ ctdb_start_recoverd(ctdb);
+}
+
+int32_t ctdb_control_recd_ping(struct ctdb_context *ctdb)
+{
+ talloc_free(ctdb->recd_ping_count);
+
+ ctdb->recd_ping_count = talloc_zero(ctdb, uint32_t);
+ CTDB_NO_MEMORY(ctdb, ctdb->recd_ping_count);
+
+ if (ctdb->tunable.recd_ping_timeout != 0) {
+ tevent_add_timer(ctdb->ev, ctdb->recd_ping_count,
+ timeval_current_ofs(ctdb->tunable.recd_ping_timeout, 0),
+ ctdb_recd_ping_timeout, ctdb);
+ }
+
+ return 0;
+}
+
+
+
+int32_t ctdb_control_set_recmaster(struct ctdb_context *ctdb, uint32_t opcode, TDB_DATA indata)
+{
+ uint32_t new_recmaster;
+
+ CHECK_CONTROL_DATA_SIZE(sizeof(uint32_t));
+ new_recmaster = ((uint32_t *)(&indata.dptr[0]))[0];
+
+ if (ctdb->pnn != new_recmaster && ctdb->recovery_master == ctdb->pnn) {
+ DEBUG(DEBUG_NOTICE,
+ ("This node (%u) is no longer the recovery master\n", ctdb->pnn));
+ }
+
+ if (ctdb->pnn == new_recmaster && ctdb->recovery_master != new_recmaster) {
+ DEBUG(DEBUG_NOTICE,
+ ("This node (%u) is now the recovery master\n", ctdb->pnn));
+ }
+
+ ctdb->recovery_master = new_recmaster;
+ return 0;
+}
+
+
+int32_t ctdb_control_stop_node(struct ctdb_context *ctdb)
+{
+ DEBUG(DEBUG_NOTICE, ("Stopping node\n"));
+ ctdb_disable_monitoring(ctdb);
+ ctdb->nodes[ctdb->pnn]->flags |= NODE_FLAGS_STOPPED;
+
+ return 0;
+}
+
+int32_t ctdb_control_continue_node(struct ctdb_context *ctdb)
+{
+ DEBUG(DEBUG_NOTICE, ("Continue node\n"));
+ ctdb->nodes[ctdb->pnn]->flags &= ~NODE_FLAGS_STOPPED;
+
+ return 0;
+}
+