-
Notifications
You must be signed in to change notification settings - Fork 1.9k
/
table_interface.cc
360 lines (317 loc) · 13.4 KB
/
table_interface.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
// Copyright 2010-2015 RethinkDB, all rights reserved.
#include "clustering/administration/persist/table_interface.hpp"
#include <algorithm>
#include <array>
#include "clustering/administration/persist/branch_history_manager.hpp"
#include "clustering/administration/persist/file_keys.hpp"
#include "clustering/administration/persist/raft_storage_interface.hpp"
#include "clustering/administration/perfmon_collection_repo.hpp"
#include "logger.hpp"
#include "rdb_protocol/store.hpp"
#include "serializer/log/log_serializer.hpp"
#include "serializer/merger.hpp"
#include "serializer/translator.hpp"
class real_multistore_ptr_t :
public multistore_ptr_t {
public:
real_multistore_ptr_t(
const namespace_id_t &table_id,
const serializer_filepath_t &path,
scoped_ptr_t<real_branch_history_manager_t> &&bhm,
const base_path_t &base_path,
io_backender_t *io_backender,
cache_balancer_t *cache_balancer,
rdb_context_t *rdb_context,
perfmon_collection_t *perfmon_collection_serializers,
scoped_ptr_t<thread_allocation_t> &&serializer_thread,
std::vector<scoped_ptr_t<thread_allocation_t> > &&store_threads,
std::map<
namespace_id_t, std::pair<real_multistore_ptr_t *, auto_drainer_t::lock_t>
> *real_multistores) :
branch_history_manager(std::move(bhm)),
serializer_thread_allocation(std::move(serializer_thread)),
store_thread_allocations(std::move(store_threads)),
map_insertion_sentry(
real_multistores, table_id, std::make_pair(this, drainer.lock()))
{
// TODO: If the server gets killed when starting up, we can
// get a database in an invalid startup state.
// TODO: This is quite suspicious in that we check if the file
// exists and then assume it exists or does not exist when
// loading or creating it.
// TODO: We should use N slices on M serializers, not N slices
// on 1 serializer.
int res = access(path.permanent_path().c_str(), R_OK | W_OK);
bool create = (res != 0);
on_thread_t thread_switcher(serializer_thread_allocation->get_thread());
filepath_file_opener_t file_opener(path, io_backender);
if (create) {
log_serializer_t::create(
&file_opener,
log_serializer_t::static_config_t());
}
// TODO: Could we handle failure when loading the serializer? Right
// now, we don't.
scoped_ptr_t<serializer_t> inner_serializer(new log_serializer_t(
log_serializer_t::dynamic_config_t(),
&file_opener,
perfmon_collection_serializers));
serializer.init(new merger_serializer_t(
std::move(inner_serializer),
MERGER_SERIALIZER_MAX_ACTIVE_WRITES));
std::vector<serializer_t *> ptrs;
ptrs.push_back(serializer.get());
if (create) {
serializer_multiplexer_t::create(ptrs, CPU_SHARDING_FACTOR);
}
multiplexer.init(new serializer_multiplexer_t(ptrs));
pmap(CPU_SHARDING_FACTOR, [&](int ix) {
// TODO: Exceptions? If exceptions are being thrown in here, nothing is
// handling them.
on_thread_t thread_switcher_2(store_thread_allocations[ix]->get_thread());
stores[ix].init(new store_t(
cpu_sharding_subspace(ix),
multiplexer->proxies[ix],
cache_balancer,
strprintf("shard_%d", ix),
create,
perfmon_collection_serializers,
rdb_context,
io_backender,
base_path,
table_id,
update_sindexes_t::UPDATE));
/* Initialize the metainfo if necessary */
if (create) {
order_source_t order_source;
cond_t non_interruptor;
write_token_t write_token;
stores[ix]->new_write_token(&write_token);
stores[ix]->set_metainfo(
region_map_t<binary_blob_t>(
region_t::universe(),
binary_blob_t(version_t::zero())),
order_source.check_in("real_multistore_ptr_t"),
&write_token,
write_durability_t::HARD,
&non_interruptor);
}
});
if (create) {
file_opener.move_serializer_file_to_permanent_location();
}
}
~real_multistore_ptr_t() {
serializer_thread_allocation.reset();
store_thread_allocations.clear();
map_insertion_sentry.reset();
drainer.drain();
pmap(CPU_SHARDING_FACTOR, [this](int ix) {
if (stores[ix].has()) {
on_thread_t thread_switcher(stores[ix]->home_thread());
stores[ix].reset();
}
});
if (serializer.has()) {
on_thread_t thread_switcher(serializer->home_thread());
if (multiplexer.has()) {
multiplexer.reset();
}
serializer.reset();
}
}
branch_history_manager_t *get_branch_history_manager() {
return branch_history_manager.get();
}
serializer_t *get_serializer() {
return serializer.get_or_null();
}
store_view_t *get_cpu_sharded_store(size_t i) {
return stores[i].get();
}
store_t *get_underlying_store(size_t i) {
return stores[i].get();
}
bool is_gc_active() {
rassert(!drainer.is_draining());
if (serializer.has()) {
return serializer->is_gc_active();
} else {
return false;
}
}
private:
scoped_ptr_t<real_branch_history_manager_t> branch_history_manager;
scoped_ptr_t<serializer_t> serializer;
scoped_ptr_t<serializer_multiplexer_t> multiplexer;
scoped_ptr_t<store_t> stores[CPU_SHARDING_FACTOR];
scoped_ptr_t<thread_allocation_t> serializer_thread_allocation;
std::vector<scoped_ptr_t<thread_allocation_t> > store_thread_allocations;
auto_drainer_t drainer;
map_insertion_sentry_t<
namespace_id_t, std::pair<real_multistore_ptr_t *, auto_drainer_t::lock_t>
> map_insertion_sentry;
DISABLE_COPYING(real_multistore_ptr_t);
};
void real_table_persistence_interface_t::read_all_metadata(
const std::function<void(
const namespace_id_t &table_id,
const table_active_persistent_state_t &state,
raft_storage_interface_t<table_raft_state_t> *raft_storage,
metadata_file_t::read_txn_t *metadata_read_txn)> &active_cb,
const std::function<void(
const namespace_id_t &table_id,
const table_inactive_persistent_state_t &state,
metadata_file_t::read_txn_t *metadata_read_txn)> &inactive_cb,
signal_t *interruptor) {
metadata_file_t::read_txn_t read_txn(metadata_file, interruptor);
std::map<namespace_id_t, table_active_persistent_state_t> active_tables;
read_txn.read_many<table_active_persistent_state_t>(
mdprefix_table_active(),
[&](const std::string &uuid_str, const table_active_persistent_state_t &state) {
active_tables[str_to_uuid(uuid_str)] = state;
},
interruptor);
storage_interfaces.clear();
for (const auto &pair : active_tables) {
storage_interfaces[pair.first].init(new table_raft_storage_interface_t(
metadata_file, &read_txn, pair.first, interruptor));
active_cb(
pair.first, pair.second, storage_interfaces[pair.first].get(), &read_txn);
}
read_txn.read_many<table_inactive_persistent_state_t>(
mdprefix_table_inactive(),
[&](const std::string &uuid_str, const table_inactive_persistent_state_t &s) {
inactive_cb(str_to_uuid(uuid_str), s, &read_txn);
},
interruptor);
}
void real_table_persistence_interface_t::write_metadata_active(
const namespace_id_t &table_id,
const table_active_persistent_state_t &state,
const raft_persistent_state_t<table_raft_state_t> &raft_state,
raft_storage_interface_t<table_raft_state_t> **raft_storage_out) {
cond_t non_interruptor;
storage_interfaces.erase(table_id);
metadata_file_t::write_txn_t write_txn(metadata_file, &non_interruptor);
write_txn.erase(
mdprefix_table_inactive().suffix(uuid_to_str(table_id)),
&non_interruptor);
table_raft_storage_interface_t::erase(&write_txn, table_id);
write_txn.write(
mdprefix_table_active().suffix(uuid_to_str(table_id)),
state,
&non_interruptor);
storage_interfaces[table_id].init(new table_raft_storage_interface_t(
metadata_file, &write_txn, table_id, raft_state));
*raft_storage_out = storage_interfaces[table_id].get();
}
void real_table_persistence_interface_t::write_metadata_inactive(
const namespace_id_t &table_id,
const table_inactive_persistent_state_t &state) {
cond_t non_interruptor;
storage_interfaces.erase(table_id);
metadata_file_t::write_txn_t write_txn(metadata_file, &non_interruptor);
write_txn.erase(
mdprefix_table_active().suffix(uuid_to_str(table_id)),
&non_interruptor);
write_txn.write(
mdprefix_table_inactive().suffix(uuid_to_str(table_id)),
state,
&non_interruptor);
table_raft_storage_interface_t::erase(&write_txn, table_id);
real_branch_history_manager_t::erase(&write_txn, table_id);
}
void real_table_persistence_interface_t::delete_metadata(
const namespace_id_t &table_id) {
cond_t non_interruptor;
storage_interfaces.erase(table_id);
metadata_file_t::write_txn_t write_txn(metadata_file, &non_interruptor);
write_txn.erase(
mdprefix_table_active().suffix(uuid_to_str(table_id)),
&non_interruptor);
write_txn.erase(
mdprefix_table_inactive().suffix(uuid_to_str(table_id)),
&non_interruptor);
table_raft_storage_interface_t::erase(&write_txn, table_id);
real_branch_history_manager_t::erase(&write_txn, table_id);
}
void real_table_persistence_interface_t::load_multistore(
const namespace_id_t &table_id,
metadata_file_t::read_txn_t *metadata_read_txn,
scoped_ptr_t<multistore_ptr_t> *multistore_ptr_out,
signal_t *interruptor,
perfmon_collection_t *perfmon_collection_serializers) {
scoped_ptr_t<real_branch_history_manager_t> bhm(
new real_branch_history_manager_t(
table_id, metadata_file, metadata_read_txn, interruptor));
scoped_ptr_t<thread_allocation_t> serializer_thread(
new thread_allocation_t(&thread_allocator));
std::vector<scoped_ptr_t<thread_allocation_t> > store_threads;
for (size_t i = 0; i < CPU_SHARDING_FACTOR; ++i) {
store_threads.emplace_back(new thread_allocation_t(&thread_allocator));
}
multistore_ptr_out->init(new real_multistore_ptr_t(
table_id,
file_name_for(table_id),
std::move(bhm),
base_path,
io_backender,
cache_balancer,
rdb_context,
perfmon_collection_serializers,
std::move(serializer_thread),
std::move(store_threads),
&real_multistores));
}
void real_table_persistence_interface_t::create_multistore(
const namespace_id_t &table_id,
scoped_ptr_t<multistore_ptr_t> *multistore_ptr_out,
signal_t *interruptor,
perfmon_collection_t *perfmon_collection_serializers) {
metadata_file_t::read_txn_t read_txn(metadata_file, interruptor);
load_multistore(
table_id, &read_txn, multistore_ptr_out, interruptor,
perfmon_collection_serializers);
}
void real_table_persistence_interface_t::destroy_multistore(
const namespace_id_t &table_id,
scoped_ptr_t<multistore_ptr_t> *multistore_ptr_in) {
guarantee(multistore_ptr_in->has());
multistore_ptr_in->reset();
std::string filepath = file_name_for(table_id).permanent_path();
logNTC("Removing file %s\n", filepath.c_str());
const int res = ::unlink(filepath.c_str());
guarantee_err(res == 0 || get_errno() == ENOENT,
"unlink failed for file %s", filepath.c_str());
}
serializer_filepath_t real_table_persistence_interface_t::file_name_for(
const namespace_id_t &table_id) {
return serializer_filepath_t(base_path, uuid_to_str(table_id));
}
bool real_table_persistence_interface_t::is_gc_active() const {
for (int thread = 0; thread < get_num_db_threads(); ++thread) {
std::map<serializer_t *, auto_drainer_t::lock_t> serializers_copy;
// Note the copy in the loop below is intentional, one of the members is a
// `auto_drainer_t::lock_t` that we want to hold.
for (auto real_multistore : real_multistores) {
serializer_t *serializer =
real_multistore.second.first->get_serializer();
if (serializer == nullptr ||
serializer->home_thread() != threadnum_t(thread)) {
continue;
}
serializers_copy.insert(
std::make_pair(serializer, real_multistore.second.second));
}
{
on_thread_t on_thread((threadnum_t(thread)));
for (auto const &serializer : serializers_copy) {
if (serializer.first->is_gc_active()) {
return true;
}
}
}
}
return false;
}