… | |
… | |
3 | BDB - Asynchronous Berkeley DB access |
3 | BDB - Asynchronous Berkeley DB access |
4 | |
4 | |
5 | =head1 SYNOPSIS |
5 | =head1 SYNOPSIS |
6 | |
6 | |
7 | use BDB; |
7 | use BDB; |
|
|
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 | |
8 | |
43 | |
9 | =head1 DESCRIPTION |
44 | =head1 DESCRIPTION |
10 | |
45 | |
11 | See the BerkeleyDB documentation (L<http://www.oracle.com/technology/documentation/berkeley-db/db/index.html>). |
46 | See the BerkeleyDB documentation (L<http://www.oracle.com/technology/documentation/berkeley-db/db/index.html>). |
12 | The BDB API is very similar to the C API (the translation has been very faithful). |
47 | The BDB API is very similar to the C API (the translation has been very faithful). |
… | |
… | |
72 | use strict 'vars'; |
107 | use strict 'vars'; |
73 | |
108 | |
74 | use base 'Exporter'; |
109 | use base 'Exporter'; |
75 | |
110 | |
76 | BEGIN { |
111 | BEGIN { |
77 | our $VERSION = '1.0'; |
112 | our $VERSION = '1.3'; |
78 | |
113 | |
79 | our @BDB_REQ = qw( |
114 | our @BDB_REQ = qw( |
80 | 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 |
81 | db_env_memp_sync db_env_memp_trickle |
116 | db_env_memp_sync db_env_memp_trickle |
82 | 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 |
83 | db_txn_commit db_txn_abort |
118 | db_txn_commit db_txn_abort db_txn_finish |
84 | 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 |
85 | db_sequence_open db_sequence_close |
120 | db_sequence_open db_sequence_close |
86 | db_sequence_get db_sequence_remove |
121 | db_sequence_get db_sequence_remove |
87 | ); |
122 | ); |
88 | 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)); |
… | |
… | |
99 | |
134 | |
100 | =head2 BERKELEYDB FUNCTIONS |
135 | =head2 BERKELEYDB FUNCTIONS |
101 | |
136 | |
102 | All of these are functions. The create functions simply return a new |
137 | All of these are functions. The create functions simply return a new |
103 | object and never block. All the remaining functions all take an optional |
138 | object and never block. All the remaining functions all take an optional |
104 | callback as last argument. If it is missing, then the fucntion will be |
139 | callback as last argument that gets passed the return value. If it is |
105 | executed synchronously. |
140 | missing, then the function will be executed synchronously, and the return |
|
|
141 | value is returned as normally. |
106 | |
142 | |
107 | BDB functions that cannot block (mostly functions that manipulate |
143 | BDB functions that cannot block (mostly functions that manipulate |
108 | settings) are method calls on the relevant objects, so the rule of thumb |
144 | settings) are method calls on the relevant objects, so the rule of thumb |
109 | is: if its a method, its not blocking, if its a function, it takes a |
145 | is: if its a method, its not blocking, if its a function, it takes a |
110 | callback as last argument. |
146 | callback as last argument. |
… | |
… | |
177 | 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 | 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) |
178 | flags: TXN_NOSYNC |
214 | flags: TXN_NOSYNC |
179 | db_sequence_remove (DB_SEQUENCE *seq, DB_TXN_ornull *txnid = 0, U32 flags = 0, SV *callback = &PL_sv_undef) |
215 | db_sequence_remove (DB_SEQUENCE *seq, DB_TXN_ornull *txnid = 0, U32 flags = 0, SV *callback = &PL_sv_undef) |
180 | flags: TXN_NOSYNC |
216 | flags: TXN_NOSYNC |
181 | |
217 | |
|
|
218 | =head4 db_txn_finish (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef) |
|
|
219 | |
|
|
220 | This is not actually a Berkeley DB function but a BDB module |
|
|
221 | extension. The background for this exytension is: It is very annoying to |
|
|
222 | have to check every single BDB function for error returns and provide a |
|
|
223 | codepath out of your transaction. While the BDB module still makes this |
|
|
224 | possible, it contains the following extensions: |
|
|
225 | |
|
|
226 | When a transaction-protected function returns any operating system |
|
|
227 | error (errno > 0), BDB will set the C<TXN_DEADLOCK> flag on the |
|
|
228 | transaction. This flag is also set by Berkeley DB functions themselves |
|
|
229 | when an operation fails with LOCK_DEADLOCK, and it causes all further |
|
|
230 | operations on that transaction (including C<db_txn_commit>) to fail. |
|
|
231 | |
|
|
232 | The C<db_txn_finish> request will look at this flag, and, if it is set, |
|
|
233 | will automatically call C<db_txn_abort> (setting errno to C<LOCK_DEADLOCK> |
|
|
234 | if it isn't set to something else yet). If it isn't set, it will call |
|
|
235 | C<db_txn_commit> and return the error normally. |
|
|
236 | |
|
|
237 | How to use this? Easy: just write your transaction normally: |
|
|
238 | |
|
|
239 | my $txn = $db_env->txn_begin; |
|
|
240 | db_get $db, $txn, "key", my $data; |
|
|
241 | db_put $db, $txn, "key", $data + 1 unless $! == BDB::NOTFOUND; |
|
|
242 | db_txn_finish $txn; |
|
|
243 | die "transaction failed" if $!; |
|
|
244 | |
|
|
245 | That is, handle only the expected errors. If something unexpected happens |
|
|
246 | (EIO, LOCK_NOTGRANTED or a deadlock in either db_get or db_put), then the remaining |
|
|
247 | requests (db_put in this case) will simply be skipped (they will fail with |
|
|
248 | LOCK_DEADLOCK) and the transaction will be aborted. |
|
|
249 | |
|
|
250 | You can use the C<< $txn->failed >> method to check wether a transaction |
|
|
251 | has failed in this way and abort further processing (excluding |
|
|
252 | C<db_txn_finish>). |
|
|
253 | |
182 | =head3 DB_ENV/database environment methods |
254 | =head3 DB_ENV/database environment methods |
183 | |
255 | |
184 | Methods available on DB_ENV/$env handles: |
256 | Methods available on DB_ENV/$env handles: |
185 | |
257 | |
186 | DESTROY (DB_ENV_ornull *env) |
258 | DESTROY (DB_ENV_ornull *env) |
… | |
… | |
196 | $int = $env->set_flags (U32 flags, int onoff) |
268 | $int = $env->set_flags (U32 flags, int onoff) |
197 | $env->set_errfile (FILE *errfile = 0) |
269 | $env->set_errfile (FILE *errfile = 0) |
198 | $env->set_msgfile (FILE *msgfile = 0) |
270 | $env->set_msgfile (FILE *msgfile = 0) |
199 | $int = $env->set_verbose (U32 which, int onoff = 1) |
271 | $int = $env->set_verbose (U32 which, int onoff = 1) |
200 | $int = $env->set_encrypt (const char *password, U32 flags = 0) |
272 | $int = $env->set_encrypt (const char *password, U32 flags = 0) |
201 | $int = $env->set_timeout (NV timeout, U32 flags) |
273 | $int = $env->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT) |
202 | $int = $env->set_mp_max_openfd (int maxopenfd); |
274 | $int = $env->set_mp_max_openfd (int maxopenfd); |
203 | $int = $env->set_mp_max_write (int maxwrite, int maxwrite_sleep); |
275 | $int = $env->set_mp_max_write (int maxwrite, int maxwrite_sleep); |
204 | $int = $env->set_mp_mmapsize (int mmapsize_mb) |
276 | $int = $env->set_mp_mmapsize (int mmapsize_mb) |
205 | $int = $env->set_lk_detect (U32 detect = DB_LOCK_DEFAULT) |
277 | $int = $env->set_lk_detect (U32 detect = DB_LOCK_DEFAULT) |
206 | $int = $env->set_lk_max_lockers (U32 max) |
278 | $int = $env->set_lk_max_lockers (U32 max) |
207 | $int = $env->set_lk_max_locks (U32 max) |
279 | $int = $env->set_lk_max_locks (U32 max) |
208 | $int = $env->set_lk_max_objects (U32 max) |
280 | $int = $env->set_lk_max_objects (U32 max) |
209 | $int = $env->set_lg_bsize (U32 max) |
281 | $int = $env->set_lg_bsize (U32 max) |
210 | $int = $env->set_lg_max (U32 max) |
282 | $int = $env->set_lg_max (U32 max) |
|
|
283 | $int = $env->mutex_set_increment (U32 increment) |
|
|
284 | $int = $env->mutex_set_tas_spins (U32 tas_spins) |
|
|
285 | $int = $env->mutex_set_max (U32 max) |
|
|
286 | $int = $env->mutex_set_align (U32 align) |
211 | |
287 | |
212 | $txn = $env->txn_begin (DB_TXN_ornull *parent = 0, U32 flags = 0) |
288 | $txn = $env->txn_begin (DB_TXN_ornull *parent = 0, U32 flags = 0) |
213 | flags: READ_COMMITTED READ_UNCOMMITTED TXN_NOSYNC TXN_NOWAIT TXN_SNAPSHOT TXN_SYNC TXN_WAIT TXN_WRITE_NOSYNC |
289 | flags: READ_COMMITTED READ_UNCOMMITTED TXN_NOSYNC TXN_NOWAIT TXN_SNAPSHOT TXN_SYNC TXN_WAIT TXN_WRITE_NOSYNC |
214 | |
290 | |
215 | =head4 Example: |
291 | =head4 Example: |
… | |
… | |
294 | DESTROY (DB_TXN_ornull *txn) |
370 | DESTROY (DB_TXN_ornull *txn) |
295 | CODE: |
371 | CODE: |
296 | if (txn) |
372 | if (txn) |
297 | txn->abort (txn); |
373 | txn->abort (txn); |
298 | |
374 | |
299 | $int = $txn->set_timeout (NV timeout, U32 flags) |
375 | $int = $txn->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT) |
300 | flags: SET_LOCK_TIMEOUT SET_TXN_TIMEOUT |
376 | flags: SET_LOCK_TIMEOUT SET_TXN_TIMEOUT |
|
|
377 | |
|
|
378 | $bool = $txn->failed |
|
|
379 | # see db_txn_finish documentation, above |
301 | |
380 | |
302 | |
381 | |
303 | =head3 DBC/cursor methods |
382 | =head3 DBC/cursor methods |
304 | |
383 | |
305 | Methods available on DBC/$dbc handles: |
384 | Methods available on DBC/$dbc handles: |
… | |
… | |
349 | |
428 | |
350 | =head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION |
429 | =head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION |
351 | |
430 | |
352 | =over 4 |
431 | =over 4 |
353 | |
432 | |
|
|
433 | =item $msg = BDB::strerror [$errno] |
|
|
434 | |
|
|
435 | Returns the string corresponding to the given errno value. If no argument |
|
|
436 | is given, use C<$!>. |
|
|
437 | |
354 | =item $fileno = BDB::poll_fileno |
438 | =item $fileno = BDB::poll_fileno |
355 | |
439 | |
356 | Return the I<request result pipe file descriptor>. This filehandle must be |
440 | Return the I<request result pipe file descriptor>. This filehandle must be |
357 | polled for reading by some mechanism outside this module (e.g. Event or |
441 | polled for reading by some mechanism outside this module (e.g. Event or |
358 | select, see below or the SYNOPSIS). If the pipe becomes readable you have |
442 | select, see below or the SYNOPSIS). If the pipe becomes readable you have |
… | |
… | |
396 | interactiveness when perl is not fast enough to process all requests in |
480 | interactiveness when perl is not fast enough to process all requests in |
397 | time. |
481 | time. |
398 | |
482 | |
399 | For interactive programs, values such as C<0.01> to C<0.1> should be fine. |
483 | For interactive programs, values such as C<0.01> to C<0.1> should be fine. |
400 | |
484 | |
401 | Example: Install an Event watcher that automatically calls |
485 | Example: Install an EV watcher that automatically calls |
402 | BDB::poll_cb with low priority, to ensure that other parts of the |
486 | BDB::poll_cb with low priority, to ensure that other parts of the |
403 | program get the CPU sometimes even under high AIO load. |
487 | program get the CPU sometimes even under high load. |
404 | |
488 | |
405 | # try not to spend much more than 0.1s in poll_cb |
489 | # try not to spend much more than 0.1s in poll_cb |
406 | BDB::max_poll_time 0.1; |
490 | BDB::max_poll_time 0.1; |
407 | |
491 | |
408 | # use a low priority so other tasks have priority |
492 | my $bdb_poll = EV::io BDB::poll_fileno, EV::READ, \&BDB::poll_cb); |
409 | Event->io (fd => BDB::poll_fileno, |
|
|
410 | poll => 'r', nice => 1, |
|
|
411 | cb => &BDB::poll_cb); |
|
|
412 | |
493 | |
413 | =item BDB::poll_wait |
494 | =item BDB::poll_wait |
414 | |
495 | |
415 | If there are any outstanding requests and none of them in the result |
496 | If there are any outstanding requests and none of them in the result |
416 | phase, wait till the result filehandle becomes ready for reading (simply |
497 | phase, wait till the result filehandle becomes ready for reading (simply |
… | |
… | |
428 | |
509 | |
429 | BDB::poll_wait, BDB::poll_cb |
510 | BDB::poll_wait, BDB::poll_cb |
430 | |
511 | |
431 | =item BDB::flush |
512 | =item BDB::flush |
432 | |
513 | |
433 | Wait till all outstanding AIO requests have been handled. |
514 | Wait till all outstanding BDB requests have been handled. |
434 | |
515 | |
435 | Strictly equivalent to: |
516 | Strictly equivalent to: |
436 | |
517 | |
437 | BDB::poll_wait, BDB::poll_cb |
518 | BDB::poll_wait, BDB::poll_cb |
438 | while BDB::nreqs; |
519 | while BDB::nreqs; |
… | |
… | |
443 | |
524 | |
444 | =over 4 |
525 | =over 4 |
445 | |
526 | |
446 | =item BDB::min_parallel $nthreads |
527 | =item BDB::min_parallel $nthreads |
447 | |
528 | |
448 | Set the minimum number of AIO threads to C<$nthreads>. The current |
529 | Set the minimum number of BDB threads to C<$nthreads>. The current |
449 | default is C<8>, which means eight asynchronous operations can execute |
530 | default is C<8>, which means eight asynchronous operations can execute |
450 | concurrently at any one time (the number of outstanding requests, |
531 | concurrently at any one time (the number of outstanding requests, |
451 | however, is unlimited). |
532 | however, is unlimited). |
452 | |
533 | |
453 | BDB starts threads only on demand, when an AIO request is queued and |
534 | BDB starts threads only on demand, when an BDB request is queued and |
454 | no free thread exists. Please note that queueing up a hundred requests can |
535 | no free thread exists. Please note that queueing up a hundred requests can |
455 | create demand for a hundred threads, even if it turns out that everything |
536 | create demand for a hundred threads, even if it turns out that everything |
456 | is in the cache and could have been processed faster by a single thread. |
537 | is in the cache and could have been processed faster by a single thread. |
457 | |
538 | |
458 | It is recommended to keep the number of threads relatively low, as some |
539 | It is recommended to keep the number of threads relatively low, as some |
… | |
… | |
463 | Under most circumstances you don't need to call this function, as the |
544 | Under most circumstances you don't need to call this function, as the |
464 | module selects a default that is suitable for low to moderate load. |
545 | module selects a default that is suitable for low to moderate load. |
465 | |
546 | |
466 | =item BDB::max_parallel $nthreads |
547 | =item BDB::max_parallel $nthreads |
467 | |
548 | |
468 | Sets the maximum number of AIO threads to C<$nthreads>. If more than the |
549 | Sets the maximum number of BDB threads to C<$nthreads>. If more than the |
469 | specified number of threads are currently running, this function kills |
550 | specified number of threads are currently running, this function kills |
470 | them. This function blocks until the limit is reached. |
551 | them. This function blocks until the limit is reached. |
471 | |
552 | |
472 | While C<$nthreads> are zero, aio requests get queued but not executed |
553 | While C<$nthreads> are zero, aio requests get queued but not executed |
473 | until the number of threads has been increased again. |
554 | until the number of threads has been increased again. |
… | |
… | |
576 | |
657 | |
577 | =head2 FORK BEHAVIOUR |
658 | =head2 FORK BEHAVIOUR |
578 | |
659 | |
579 | This module should do "the right thing" when the process using it forks: |
660 | This module should do "the right thing" when the process using it forks: |
580 | |
661 | |
581 | Before the fork, IO::AIO enters a quiescent state where no requests |
662 | Before the fork, BDB enters a quiescent state where no requests |
582 | can be added in other threads and no results will be processed. After |
663 | can be added in other threads and no results will be processed. After |
583 | the fork the parent simply leaves the quiescent state and continues |
664 | the fork the parent simply leaves the quiescent state and continues |
584 | request/result processing, while the child frees the request/result queue |
665 | request/result processing, while the child frees the request/result queue |
585 | (so that the requests started before the fork will only be handled in the |
666 | (so that the requests started before the fork will only be handled in the |
586 | parent). Threads will be started on demand until the limit set in the |
667 | parent). Threads will be started on demand until the limit set in the |
587 | parent process has been reached again. |
668 | parent process has been reached again. |
588 | |
669 | |
589 | In short: the parent will, after a short pause, continue as if fork had |
670 | In short: the parent will, after a short pause, continue as if fork had |
590 | not been called, while the child will act as if IO::AIO has not been used |
671 | not been called, while the child will act as if BDB has not been used |
591 | yet. |
672 | yet. |
592 | |
673 | |
593 | =head2 MEMORY USAGE |
674 | =head2 MEMORY USAGE |
594 | |
675 | |
595 | Per-request usage: |
676 | Per-request usage: |
… | |
… | |
609 | temporary buffers, and each thread requires a stack and other data |
690 | temporary buffers, and each thread requires a stack and other data |
610 | structures (usually around 16k-128k, depending on the OS). |
691 | structures (usually around 16k-128k, depending on the OS). |
611 | |
692 | |
612 | =head1 KNOWN BUGS |
693 | =head1 KNOWN BUGS |
613 | |
694 | |
614 | Known bugs will be fixed in the next release. |
695 | Known bugs will be fixed in the next release, except: |
|
|
696 | |
|
|
697 | If you use a transaction in any request, and the request returns |
|
|
698 | with an operating system error or DB_LOCK_NOTGRANTED, the internal |
|
|
699 | TXN_DEADLOCK flag will be set on the transaction. See C<db_txn_finish>, |
|
|
700 | above. |
615 | |
701 | |
616 | =head1 SEE ALSO |
702 | =head1 SEE ALSO |
617 | |
703 | |
618 | L<Coro::AIO>. |
704 | L<Coro::BDB>, L<IO::AIO>. |
619 | |
705 | |
620 | =head1 AUTHOR |
706 | =head1 AUTHOR |
621 | |
707 | |
622 | Marc Lehmann <schmorp@schmorp.de> |
708 | Marc Lehmann <schmorp@schmorp.de> |
623 | http://home.schmorp.de/ |
709 | http://home.schmorp.de/ |