WAS_WRITE lock leaking with DB_DIRTY_READ & DB_AUTO_COMMIT turn on

sub.apache@gmail.com · *nix forums beginner Joined: 11 Jul 2006 Posts: 3

Hi, I has a hash based berkeley db (4.4) with DB_AUTO_COMMIT on
and it works fine. But once I turned on both DB_DIRTY_READ
and DB_AUTO_COMMIT, it generates a WAS_WRITE lock after every
DB->put() operation without releasing it. So after a while, it
ran out of all locks (default 64K) and reached error of
"BERKELEY DB: Lock table is out of available object entries".

I didn't generate any transaction explicitly in my application.
I didn't make any DB->get() or cursor call. I didn't share
the db with other process. The txn_stat output
didn't show any transaction holding those WAS_WRITE locks.
If I removed DB_DIRTY_READ in the open() call, this lock
leaking problem is gone.

Any idea why the db is holding WAS_WRITE lock after the
DB->put() is done ? How can I find out which transaction
is holding those locks ? Thanks in advanced. - John

Here the db create code: --------------------
rc = db_env->open( db_env,
my_home,
DB_CREATE |
DB_INIT_LOCK |
DB_INIT_LOG |
DB_INIT_TXN |
DB_INIT_MPOOL |
DB_PRIVATE ,
0 );

rc = db->open(db, NULL, my_file_name, NULL, DB_HASH, DB_CREATE,
DB_PERMS);
db->close(db, 0);

Here db open code: -----------------------
rc = db_env->set_flags(db_env, DB_AUTO_COMMIT, 1);

rc = my_db->open( db_env,
NULL, /*txn*/
my_file_name,
NULL,
DB_UNKNOWN,
DB_THREAD , DB_DIRTY_READ,
0664 );

Here's the lock_stat output: ---------------

Default locking region information:
3 Last allocated locker ID
0x7fffffff Current maximum unused locker ID
9 Number of lock modes
65536 Maximum number of locks possible
128 Maximum number of lockers possible
65536 Maximum number of lock objects possible
6 Number of current locks
7 Maximum number of locks at any one time
2 Number of current lockers
2 Maximum number of lockers at any one time
6 Number of current lock objects
7 Maximum number of lock objects at any one time
114 Total number of locks requested
108 Total number of locks released
0 Total number of locks upgraded
1 Total number of locks downgraded
0 Lock requests not available due to conflicts, for which we
waited
0 Lock requests not available due to conflicts, for which we did
not wait
0 Number of deadlocks
0 Lock timeout value
0 Number of locks that have timed out
0 Transaction timeout value
0 Number of transactions that have timed out
13MB 152KB The size of the lock region
0 The number of region locks that required waiting (0%)
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Lock REGINFO information:
Lock Region type
5 Region ID
/db/1/__db.005 Region name
0xbe3cad0 Original region address
0xbe3cad0 Region address
0xbe3cb28 Region primary address
13787136 Region maximum allocation
13787136 Region allocated
REGION_CREATE, REGION_CREATE_OK, REGION_JOIN_OK Region flags
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Lock region parameters:
2062 Lock region region mutex [0/241 0% 12274/3076336544]
131 locker table size
65537 object table size
199554752 obj_off
199478360 locker_off
0 need_dd
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Lock conflict matrix:
0 0 0 0 0 0 0 0 0

0 0 1 0 1 0 1 0 1

0 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0

0 1 1 0 0 0 0 1 1

0 0 1 0 0 0 0 0 1

0 1 1 0 0 0 0 1 1

0 0 1 0 1 0 1 0 0

0 1 1 0 1 1 1 0 1

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Locks grouped by lockers:
Locker Mode Count Status ----------------- Object
---------------
2 dd= 0 locks held 1 write locks 0 pid/thread
12274/3086829760
2 READ 1 HELD my_test.db4.4 handle
0
3 dd= 0 locks held 5 write locks 5 pid/thread
12274/3065846688
3 WAS_WRITE 1 HELD my_test.db4.4 page
83096
3 WAS_WRITE 1 HELD my_test.db4.4 page
62334
3 WAS_WRITE 1 HELD my_test.db4.4 page
24944
3 WAS_WRITE 1 HELD my_test.db4.4 page
28650
3 WAS_WRITE 1 HELD my_test.db4.4 page
72029
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Locks grouped by object:
Locker Mode Count Status ----------------- Object
---------------
3 WAS_WRITE 1 HELD my_test.db4.4 page
24944

3 WAS_WRITE 1 HELD my_test.db4.4 page
28650

3 WAS_WRITE 1 HELD my_test.db4.4 page
83096

3 WAS_WRITE 1 HELD my_test.db4.4 page
72029

2 READ 1 HELD my_test.db4.4 handle
0

3 WAS_WRITE 1 HELD my_test.db4.4 page
62334

Here is the txn_stat output -----------------

Default transaction region information:
1/68944 File/offset for last checkpoint LSN
Sun Jul 16 14:22:24 2006 Checkpoint timestamp
0x80000000 Last transaction ID allocated
20 Maximum number of active transactions configured
0 Active transactions
0 Maximum active transactions
0 Number of transactions begun
0 Number of transactions aborted
0 Number of transactions committed
0 Number of transactions restored
16KB Transaction region size
0 The number of region locks that required waiting (0%)
Active transactions:
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Transaction REGINFO information:
Transaction Region type
6 Region ID
/db/1/__db.006 Region name
0xab455c0 Original region address
0xab455c0 Region address
0xab45618 Region primary address
16384 Region maximum allocation
16384 Region allocated
REGION_CREATE, REGION_CREATE_OK, REGION_JOIN_OK Region flags
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
DB_TXNMGR handle information:
2065 DB_TXNMGR mutex [0/0 0% !Own]
0 Number of transactions discarded
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
DB_TXNREGION handle information:
2063 DB_TXNREGION region mutex [0/9 0% 10783/3076389792]
20 Maximum number of active txns
0x80000000 Last transaction ID allocated
0xffffffff Current maximum unused ID
2064 checkpoint mutex [0/2 0% !Own]
1/68944 Last checkpoint LSN
Sun Jul 16 14:22:24 2006 Last checkpoint timestamp
Flags
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
XA information:
0 XA RMID

ubell@sleepycat.com · *nix forums beginner Joined: 06 Jun 2005 Posts: 49

sub.apache@gmail.com · *nix forums beginner Joined: 11 Jul 2006 Posts: 3

Thanks for the update. That actual code is using db handle and the
program does work
for 64k put() calls. But in order to make the post shorter, I replaced
the original long db handle with the wrong typing.

Mike wrote:

Google