ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/BDB/BDB.pm
(Generate patch)

Comparing BDB/BDB.pm (file contents):
Revision 1.7 by root, Mon Mar 5 19:47:01 2007 UTC vs.
Revision 1.20 by root, Fri Dec 7 13:14:41 2007 UTC

4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use BDB; 7 use BDB;
8 8
9 my $env = db_env_create;
10
11 mkdir "bdtest", 0700;
12 db_env_open
13 $env,
14 "bdtest",
15 BDB::INIT_LOCK | BDB::INIT_LOG | BDB::INIT_MPOOL
16 | BDB::INIT_TXN | BDB::RECOVER | BDB::USE_ENVIRON | BDB::CREATE,
17 0600;
18
19 $env->set_flags (BDB::AUTO_COMMIT | BDB::TXN_NOSYNC, 1);
20
21 my $db = db_create $env;
22 db_open $db, undef, "table", undef, BDB::BTREE, BDB::AUTO_COMMIT | BDB::CREATE
23 | BDB::READ_UNCOMMITTED, 0600;
24 db_put $db, undef, "key", "data", 0, sub {
25 db_del $db, undef, "key";
26 };
27 db_sync $db;
28
29 # automatic result processing with AnyEvent:
30 our $FH; open $FH, "<&=" . BDB::poll_fileno;
31 our $WATCHER = AnyEvent->io (fh => $FH, poll => 'r', cb => \&BDB::poll_cb);
32
33 # automatic result processing with EV:
34 my $WATCHER = EV::io BDB::poll_fileno, EV::READ, \&BDB::poll_cb;
35
36 # with Glib:
37 add_watch Glib::IO BDB::poll_fileno,
38 in => sub { BDB::poll_cb; 1 };
39
40 # or simply flush manually
41 BDB::flush;
42
43
9=head1 DESCRIPTION 44=head1 DESCRIPTION
10 45
11See the eg/ directory in the distribution and the berkeleydb C 46See the BerkeleyDB documentation (L<http://www.oracle.com/technology/documentation/berkeley-db/db/index.html>).
12documentation. This is inadequate, but the only sources of documentation 47The BDB API is very similar to the C API (the translation has been very faithful).
13known for this module so far.
14 48
15=head2 EXAMPLE 49See also the example sections in the document below and possibly the eg/
50subdirectory of the BDB distribution. Last not least see the IO::AIO
51documentation, as that module uses almost the same asynchronous request
52model as this module.
53
54I know this is woefully inadequate documentation. Send a patch!
55
16 56
17=head1 REQUEST ANATOMY AND LIFETIME 57=head1 REQUEST ANATOMY AND LIFETIME
18 58
19Every request method creates a request. which is a C data structure not 59Every request method creates a request. which is a C data structure not
20directly visible to Perl. 60directly visible to Perl.
67use strict 'vars'; 107use strict 'vars';
68 108
69use base 'Exporter'; 109use base 'Exporter';
70 110
71BEGIN { 111BEGIN {
72 our $VERSION = '0.1'; 112 our $VERSION = '1.2';
73 113
74 our @BDB_REQ = qw( 114 our @BDB_REQ = qw(
75 db_env_open db_env_close db_env_txn_checkpoint db_env_lock_detect 115 db_env_open db_env_close db_env_txn_checkpoint db_env_lock_detect
76 db_env_memp_sync db_env_memp_trickle 116 db_env_memp_sync db_env_memp_trickle
77 db_open db_close db_compact db_sync db_put db_get db_pget db_del db_key_range 117 db_open db_close db_compact db_sync db_put db_get db_pget db_del db_key_range
78 db_txn_commit db_txn_abort 118 db_txn_commit db_txn_abort db_txn_finish
79 db_c_close db_c_count db_c_put db_c_get db_c_pget db_c_del 119 db_c_close db_c_count db_c_put db_c_get db_c_pget db_c_del
80 db_sequence_open db_sequence_close 120 db_sequence_open db_sequence_close
81 db_sequence_get db_sequence_remove 121 db_sequence_get db_sequence_remove
82 ); 122 );
83 our @EXPORT = (@BDB_REQ, qw(dbreq_pri dbreq_nice db_env_create db_create)); 123 our @EXPORT = (@BDB_REQ, qw(dbreq_pri dbreq_nice db_env_create db_create));
90 130
91 require XSLoader; 131 require XSLoader;
92 XSLoader::load ("BDB", $VERSION); 132 XSLoader::load ("BDB", $VERSION);
93} 133}
94 134
135=head2 BERKELEYDB FUNCTIONS
136
137All of these are functions. The create functions simply return a new
138object and never block. All the remaining functions all take an optional
139callback as last argument. If it is missing, then the fucntion will be
140executed synchronously.
141
142BDB functions that cannot block (mostly functions that manipulate
143settings) are method calls on the relevant objects, so the rule of thumb
144is: if its a method, its not blocking, if its a function, it takes a
145callback as last argument.
146
147In the following, C<$int> signifies an integer return value,
148C<octetstring> is a "binary string" (i.e. a perl string with no character
149indices >255), C<U32> is an unsigned 32 bit integer, C<int> is some
150integer, C<NV> is a floating point value.
151
152The C<SV *> types are generic perl scalars (for input and output of data
153values), and the C<SV *callback> is the optional callback function to call
154when the request is completed.
155
156The various C<DB_ENV> etc. arguments are handles return by
157C<db_env_create>, C<db_create>, C<txn_begin> and so on. If they have an
158appended C<_ornull> this means they are optional and you can pass C<undef>
159for them, resulting a NULL pointer on the C level.
160
161=head3 BDB functions
162
163Functions in the BDB namespace, exported by default:
164
165 $env = db_env_create (U32 env_flags = 0)
166 flags: RPCCLIENT
167
168 db_env_open (DB_ENV *env, octetstring db_home, U32 open_flags, int mode, SV *callback = &PL_sv_undef)
169 open_flags: INIT_CDB INIT_LOCK INIT_LOG INIT_MPOOL INIT_REP INIT_TXN RECOVER RECOVER_FATAL USE_ENVIRON USE_ENVIRON_ROOT CREATE LOCKDOWN PRIVATE REGISTER SYSTEM_MEM
170 db_env_close (DB_ENV *env, U32 flags = 0, SV *callback = &PL_sv_undef)
171 db_env_txn_checkpoint (DB_ENV *env, U32 kbyte = 0, U32 min = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
172 flags: FORCE
173 db_env_lock_detect (DB_ENV *env, U32 flags = 0, U32 atype = DB_LOCK_DEFAULT, SV *dummy = 0, SV *callback = &PL_sv_undef)
174 atype: LOCK_DEFAULT LOCK_EXPIRE LOCK_MAXLOCKS LOCK_MAXWRITE LOCK_MINLOCKS LOCK_MINWRITE LOCK_OLDEST LOCK_RANDOM LOCK_YOUNGEST
175 db_env_memp_sync (DB_ENV *env, SV *dummy = 0, SV *callback = &PL_sv_undef)
176 db_env_memp_trickle (DB_ENV *env, int percent, SV *dummy = 0, SV *callback = &PL_sv_undef)
177
178 $db = db_create (DB_ENV *env = 0, U32 flags = 0)
179 flags: XA_CREATE
180
181 db_open (DB *db, DB_TXN_ornull *txnid, octetstring file, octetstring database, int type, U32 flags, int mode, SV *callback = &PL_sv_undef)
182 flags: AUTO_COMMIT CREATE EXCL MULTIVERSION NOMMAP RDONLY READ_UNCOMMITTED THREAD TRUNCATE
183 db_close (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
184 flags: DB_NOSYNC
185 db_compact (DB *db, DB_TXN_ornull *txn = 0, SV *start = 0, SV *stop = 0, SV *unused1 = 0, U32 flags = DB_FREE_SPACE, SV *unused2 = 0, SV *callback = &PL_sv_undef)
186 flags: FREELIST_ONLY FREE_SPACE
187 db_sync (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
188 db_key_range (DB *db, DB_TXN_ornull *txn, SV *key, SV *key_range, U32 flags = 0, SV *callback = &PL_sv_undef)
189 db_put (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
190 flags: APPEND NODUPDATA NOOVERWRITE
191 db_get (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
192 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
193 db_pget (DB *db, DB_TXN_ornull *txn, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
194 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
195 db_del (DB *db, DB_TXN_ornull *txn, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
196 db_txn_commit (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
197 flags: TXN_NOSYNC TXN_SYNC
198 db_txn_abort (DB_TXN *txn, SV *callback = &PL_sv_undef)
199
200 db_c_close (DBC *dbc, SV *callback = &PL_sv_undef)
201 db_c_count (DBC *dbc, SV *count, U32 flags = 0, SV *callback = &PL_sv_undef)
202 db_c_put (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
203 flags: AFTER BEFORE CURRENT KEYFIRST KEYLAST NODUPDATA
204 db_c_get (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
205 flags: CURRENT FIRST GET_BOTH GET_BOTH_RANGE GET_RECNO JOIN_ITEM LAST NEXT NEXT_DUP NEXT_NODUP PREV PREV_DUP PREV_NODUP SET SET_RANGE SET_RECNO READ_UNCOMMITTED MULTIPLE MULTIPLE_KEY RMW
206 db_c_pget (DBC *dbc, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
207 db_c_del (DBC *dbc, U32 flags = 0, SV *callback = &PL_sv_undef)
208
209 db_sequence_open (DB_SEQUENCE *seq, DB_TXN_ornull *txnid, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
210 flags: CREATE EXCL
211 db_sequence_close (DB_SEQUENCE *seq, U32 flags = 0, SV *callback = &PL_sv_undef)
212 db_sequence_get (DB_SEQUENCE *seq, DB_TXN_ornull *txnid, int delta, SV *seq_value, U32 flags = DB_TXN_NOSYNC, SV *callback = &PL_sv_undef)
213 flags: TXN_NOSYNC
214 db_sequence_remove (DB_SEQUENCE *seq, DB_TXN_ornull *txnid = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
215 flags: TXN_NOSYNC
216
217=head4 db_txn_finish (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
218
219This is not actually a Berkeley DB function but a BDB module
220extension. The background for this exytension is: It is very annoying to
221have to check every single BDB function for error returns and provide a
222codepath out of your transaction. While the BDB module still makes this
223possible, it contains the following extensions:
224
225When a transaction-protected function returns any operating system
226error (errno > 0), BDB will set the C<TXN_DEADLOCK> flag on the
227transaction. This flag is also set by Berkeley DB functions themselves
228when an operation fails with LOCK_DEADLOCK, and it causes all further
229operations on that transaction (including C<db_txn_commit>) to fail.
230
231The C<db_txn_finish> request will look at this flag, and, if it is set,
232will automatically call C<db_txn_abort> (setting errno to C<LOCK_DEADLOCK>
233if it isn't set to something else yet). If it isn't set, it will call
234C<db_txn_commit> and return the error normally.
235
236How to use this? Easy: just write your transaction normally:
237
238 my $txn = $db_env->txn_begin;
239 db_get $db, $txn, "key", my $data;
240 db_put $db, $txn, "key", $data + 1 unless $! == BDB::NOTFOUND;
241 db_txn_finish $txn;
242 die "transaction failed" if $!;
243
244That is, handle only the expected errors. If something unexpected happens
245(EIO, LOCK_NOTGRANTED or a deadlock in either db_get or db_put), then the remaining
246requests (db_put in this case) will simply be skipped (they will fail with
247LOCK_DEADLOCK) and the transaction will be aborted.
248
249You can use the C<< $txn->failed >> method to check wether a transaction
250has failed in this way and abort further processing (excluding
251C<db_txn_finish>).
252
253=head3 DB_ENV/database environment methods
254
255Methods available on DB_ENV/$env handles:
256
257 DESTROY (DB_ENV_ornull *env)
258 CODE:
259 if (env)
260 env->close (env, 0);
261
262 $int = $env->set_data_dir (const char *dir)
263 $int = $env->set_tmp_dir (const char *dir)
264 $int = $env->set_lg_dir (const char *dir)
265 $int = $env->set_shm_key (long shm_key)
266 $int = $env->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
267 $int = $env->set_flags (U32 flags, int onoff)
268 $env->set_errfile (FILE *errfile = 0)
269 $env->set_msgfile (FILE *msgfile = 0)
270 $int = $env->set_verbose (U32 which, int onoff = 1)
271 $int = $env->set_encrypt (const char *password, U32 flags = 0)
272 $int = $env->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
273 $int = $env->set_mp_max_openfd (int maxopenfd);
274 $int = $env->set_mp_max_write (int maxwrite, int maxwrite_sleep);
275 $int = $env->set_mp_mmapsize (int mmapsize_mb)
276 $int = $env->set_lk_detect (U32 detect = DB_LOCK_DEFAULT)
277 $int = $env->set_lk_max_lockers (U32 max)
278 $int = $env->set_lk_max_locks (U32 max)
279 $int = $env->set_lk_max_objects (U32 max)
280 $int = $env->set_lg_bsize (U32 max)
281 $int = $env->set_lg_max (U32 max)
282 $int = $env->mutex_set_increment (U32 increment)
283 $int = $env->mutex_set_tas_spins (U32 tas_spins)
284 $int = $env->mutex_set_max (U32 max)
285 $int = $env->mutex_set_align (U32 align)
286
287 $txn = $env->txn_begin (DB_TXN_ornull *parent = 0, U32 flags = 0)
288 flags: READ_COMMITTED READ_UNCOMMITTED TXN_NOSYNC TXN_NOWAIT TXN_SNAPSHOT TXN_SYNC TXN_WAIT TXN_WRITE_NOSYNC
289
290=head4 Example:
291
292 use AnyEvent;
293 use BDB;
294
295 our $FH; open $FH, "<&=" . BDB::poll_fileno;
296 our $WATCHER = AnyEvent->io (fh => $FH, poll => 'r', cb => \&BDB::poll_cb);
297
298 BDB::min_parallel 8;
299
300 my $env = db_env_create;
301
302 mkdir "bdtest", 0700;
303 db_env_open
304 $env,
305 "bdtest",
306 BDB::INIT_LOCK | BDB::INIT_LOG | BDB::INIT_MPOOL | BDB::INIT_TXN | BDB::RECOVER | BDB::USE_ENVIRON | BDB::CREATE,
307 0600;
308
309 $env->set_flags (BDB::AUTO_COMMIT | BDB::TXN_NOSYNC, 1);
310
311
312=head3 DB/database methods
313
314Methods available on DB/$db handles:
315
316 DESTROY (DB_ornull *db)
317 CODE:
318 if (db)
319 {
320 SV *env = (SV *)db->app_private;
321 db->close (db, 0);
322 SvREFCNT_dec (env);
323 }
324
325 $int = $db->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
326 $int = $db->set_flags (U32 flags)
327 flags: CHKSUM ENCRYPT TXN_NOT_DURABLE
328 Btree: DUP DUPSORT RECNUM REVSPLITOFF
329 Hash: DUP DUPSORT
330 Queue: INORDER
331 Recno: RENUMBER SNAPSHOT
332
333 $int = $db->set_encrypt (const char *password, U32 flags)
334 $int = $db->set_lorder (int lorder)
335 $int = $db->set_bt_minkey (U32 minkey)
336 $int = $db->set_re_delim (int delim)
337 $int = $db->set_re_pad (int re_pad)
338 $int = $db->set_re_source (char *source)
339 $int = $db->set_re_len (U32 re_len)
340 $int = $db->set_h_ffactor (U32 h_ffactor)
341 $int = $db->set_h_nelem (U32 h_nelem)
342 $int = $db->set_q_extentsize (U32 extentsize)
343
344 $dbc = $db->cursor (DB_TXN_ornull *txn = 0, U32 flags = 0)
345 flags: READ_COMMITTED READ_UNCOMMITTED WRITECURSOR TXN_SNAPSHOT
346 $seq = $db->sequence (U32 flags = 0)
347
348=head4 Example:
349
350 my $db = db_create $env;
351 db_open $db, undef, "table", undef, BDB::BTREE, BDB::AUTO_COMMIT | BDB::CREATE | BDB::READ_UNCOMMITTED, 0600;
352
353 for (1..1000) {
354 db_put $db, undef, "key $_", "data $_";
355
356 db_key_range $db, undef, "key $_", my $keyrange;
357 my ($lt, $eq, $gt) = @$keyrange;
358 }
359
360 db_del $db, undef, "key $_" for 1..1000;
361
362 db_sync $db;
363
364
365=head3 DB_TXN/transaction methods
366
367Methods available on DB_TXN/$txn handles:
368
369 DESTROY (DB_TXN_ornull *txn)
370 CODE:
371 if (txn)
372 txn->abort (txn);
373
374 $int = $txn->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
375 flags: SET_LOCK_TIMEOUT SET_TXN_TIMEOUT
376
377 $bool = $txn->failed
378 # see db_txn_finish documentation, above
379
380
381=head3 DBC/cursor methods
382
383Methods available on DBC/$dbc handles:
384
385 DESTROY (DBC_ornull *dbc)
386 CODE:
387 if (dbc)
388 dbc->c_close (dbc);
389
390=head4 Example:
391
392 my $c = $db->cursor;
393
394 for (;;) {
395 db_c_get $c, my $key, my $data, BDB::NEXT;
396 warn "<$!,$key,$data>";
397 last if $!;
398 }
399
400 db_c_close $c;
401
402
403=head3 DB_SEQUENCE/sequence methods
404
405Methods available on DB_SEQUENCE/$seq handles:
406
407 DESTROY (DB_SEQUENCE_ornull *seq)
408 CODE:
409 if (seq)
410 seq->close (seq, 0);
411
412 $int = $seq->initial_value (db_seq_t value)
413 $int = $seq->set_cachesize (U32 size)
414 $int = $seq->set_flags (U32 flags)
415 flags: SEQ_DEC SEQ_INC SEQ_WRAP
416 $int = $seq->set_range (db_seq_t min, db_seq_t max)
417
418=head4 Example:
419
420 my $seq = $db->sequence;
421
422 db_sequence_open $seq, undef, "seq", BDB::CREATE;
423 db_sequence_get $seq, undef, 1, my $value;
424
425
95=head2 SUPPORT FUNCTIONS 426=head2 SUPPORT FUNCTIONS
96 427
97=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION 428=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
98 429
99=over 4 430=over 4
143interactiveness when perl is not fast enough to process all requests in 474interactiveness when perl is not fast enough to process all requests in
144time. 475time.
145 476
146For interactive programs, values such as C<0.01> to C<0.1> should be fine. 477For interactive programs, values such as C<0.01> to C<0.1> should be fine.
147 478
148Example: Install an Event watcher that automatically calls 479Example: Install an EV watcher that automatically calls
149BDB::poll_cb with low priority, to ensure that other parts of the 480BDB::poll_cb with low priority, to ensure that other parts of the
150program get the CPU sometimes even under high AIO load. 481program get the CPU sometimes even under high load.
151 482
152 # try not to spend much more than 0.1s in poll_cb 483 # try not to spend much more than 0.1s in poll_cb
153 BDB::max_poll_time 0.1; 484 BDB::max_poll_time 0.1;
154 485
155 # use a low priority so other tasks have priority 486 my $bdb_poll = EV::io BDB::poll_fileno, EV::READ, \&BDB::poll_cb);
156 Event->io (fd => BDB::poll_fileno,
157 poll => 'r', nice => 1,
158 cb => &BDB::poll_cb);
159 487
160=item BDB::poll_wait 488=item BDB::poll_wait
161 489
162If there are any outstanding requests and none of them in the result 490If there are any outstanding requests and none of them in the result
163phase, wait till the result filehandle becomes ready for reading (simply 491phase, wait till the result filehandle becomes ready for reading (simply
175 503
176 BDB::poll_wait, BDB::poll_cb 504 BDB::poll_wait, BDB::poll_cb
177 505
178=item BDB::flush 506=item BDB::flush
179 507
180Wait till all outstanding AIO requests have been handled. 508Wait till all outstanding BDB requests have been handled.
181 509
182Strictly equivalent to: 510Strictly equivalent to:
183 511
184 BDB::poll_wait, BDB::poll_cb 512 BDB::poll_wait, BDB::poll_cb
185 while BDB::nreqs; 513 while BDB::nreqs;
186 514
515=back
516
187=head3 CONTROLLING THE NUMBER OF THREADS 517=head3 CONTROLLING THE NUMBER OF THREADS
188 518
519=over 4
520
189=item BDB::min_parallel $nthreads 521=item BDB::min_parallel $nthreads
190 522
191Set the minimum number of AIO threads to C<$nthreads>. The current 523Set the minimum number of BDB threads to C<$nthreads>. The current
192default is C<8>, which means eight asynchronous operations can execute 524default is C<8>, which means eight asynchronous operations can execute
193concurrently at any one time (the number of outstanding requests, 525concurrently at any one time (the number of outstanding requests,
194however, is unlimited). 526however, is unlimited).
195 527
196BDB starts threads only on demand, when an AIO request is queued and 528BDB starts threads only on demand, when an BDB request is queued and
197no free thread exists. Please note that queueing up a hundred requests can 529no free thread exists. Please note that queueing up a hundred requests can
198create demand for a hundred threads, even if it turns out that everything 530create demand for a hundred threads, even if it turns out that everything
199is in the cache and could have been processed faster by a single thread. 531is in the cache and could have been processed faster by a single thread.
200 532
201It is recommended to keep the number of threads relatively low, as some 533It is recommended to keep the number of threads relatively low, as some
206Under most circumstances you don't need to call this function, as the 538Under most circumstances you don't need to call this function, as the
207module selects a default that is suitable for low to moderate load. 539module selects a default that is suitable for low to moderate load.
208 540
209=item BDB::max_parallel $nthreads 541=item BDB::max_parallel $nthreads
210 542
211Sets the maximum number of AIO threads to C<$nthreads>. If more than the 543Sets the maximum number of BDB threads to C<$nthreads>. If more than the
212specified number of threads are currently running, this function kills 544specified number of threads are currently running, this function kills
213them. This function blocks until the limit is reached. 545them. This function blocks until the limit is reached.
214 546
215While C<$nthreads> are zero, aio requests get queued but not executed 547While C<$nthreads> are zero, aio requests get queued but not executed
216until the number of threads has been increased again. 548until the number of threads has been increased again.
319 651
320=head2 FORK BEHAVIOUR 652=head2 FORK BEHAVIOUR
321 653
322This module should do "the right thing" when the process using it forks: 654This module should do "the right thing" when the process using it forks:
323 655
324Before the fork, IO::AIO enters a quiescent state where no requests 656Before the fork, BDB enters a quiescent state where no requests
325can be added in other threads and no results will be processed. After 657can be added in other threads and no results will be processed. After
326the fork the parent simply leaves the quiescent state and continues 658the fork the parent simply leaves the quiescent state and continues
327request/result processing, while the child frees the request/result queue 659request/result processing, while the child frees the request/result queue
328(so that the requests started before the fork will only be handled in the 660(so that the requests started before the fork will only be handled in the
329parent). Threads will be started on demand until the limit set in the 661parent). Threads will be started on demand until the limit set in the
330parent process has been reached again. 662parent process has been reached again.
331 663
332In short: the parent will, after a short pause, continue as if fork had 664In short: the parent will, after a short pause, continue as if fork had
333not been called, while the child will act as if IO::AIO has not been used 665not been called, while the child will act as if BDB has not been used
334yet. 666yet.
335 667
336=head2 MEMORY USAGE 668=head2 MEMORY USAGE
337 669
338Per-request usage: 670Per-request usage:
341bytes of memory. In addition, stat requests need a stat buffer (possibly 673bytes of memory. In addition, stat requests need a stat buffer (possibly
342a few hundred bytes), readdir requires a result buffer and so on. Perl 674a few hundred bytes), readdir requires a result buffer and so on. Perl
343scalars and other data passed into aio requests will also be locked and 675scalars and other data passed into aio requests will also be locked and
344will consume memory till the request has entered the done state. 676will consume memory till the request has entered the done state.
345 677
346This is now awfully much, so queuing lots of requests is not usually a 678This is not awfully much, so queuing lots of requests is not usually a
347problem. 679problem.
348 680
349Per-thread usage: 681Per-thread usage:
350 682
351In the execution phase, some aio requests require more memory for 683In the execution phase, some aio requests require more memory for
352temporary buffers, and each thread requires a stack and other data 684temporary buffers, and each thread requires a stack and other data
353structures (usually around 16k-128k, depending on the OS). 685structures (usually around 16k-128k, depending on the OS).
354 686
355=head1 KNOWN BUGS 687=head1 KNOWN BUGS
356 688
357Known bugs will be fixed in the next release. 689Known bugs will be fixed in the next release, except:
690
691 If you use a transaction in any request, and the request returns
692 with an operating system error or DB_LOCK_NOTGRANTED, the internal
693 TXN_DEADLOCK flag will be set on the transaction. See C<db_txn_finish>,
694 above.
358 695
359=head1 SEE ALSO 696=head1 SEE ALSO
360 697
361L<Coro::AIO>. 698L<Coro::BDB>, L<IO::AIO>.
362 699
363=head1 AUTHOR 700=head1 AUTHOR
364 701
365 Marc Lehmann <schmorp@schmorp.de> 702 Marc Lehmann <schmorp@schmorp.de>
366 http://home.schmorp.de/ 703 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines