Skip to Main Content

Berkeley DB Family

Announcement

For appeals, questions and feedback, please email oracle-forums_moderators_us@oracle.com

locktimeout on secondary database inserts

3480917Oct 9 2018

Hello all,

My current setup uses BDBJE 4.1.10. I have 10 threads trying to write a Keyword object to the database.

The nature of the data is that the secondary keys - stores and items have a lot of duplicates. Meaning, many Keyword objects map to the same store and items entries.

LOCK_N_LOCK_TABLES is 7

@Entity

public class Keyword implements Serializable {

    @PrimaryKey

    private String key;

    @SecondaryKey(relate = Relationship.MANY_TO_MANY)

    private List<Long> stores = new ArrayList<Long>();

    @SecondaryKey(relate = Relationship.MANY_TO_MANY)

    private List<Long> items = new ArrayList<Long>();

}

About 10% (out of 100,000) of the secondary database inserts fail with a LockTimeoutException.

Looking at the stacktrace, I understand that the thread 6 waited to obtain the lock for 2000ms before giving up and there are 3 more threads waiting.

Essentially, the operation that the Owner thread (thread 10) is performing is taking more than 2000ms. Thread 6 is trying to insert a record in the secondary database items.

Stacktrace:

com.sleepycat.je.LockTimeoutException: (JE 4.1.10) Lock expired.

Locker 786527582 -1_fileChannelTaskExecutor-6_ThreadLocker: waited for lock on database=persist#faceted_store#com.xyzq.seo.bdb.entity.Keyword#items

LockAddr:43409842 node=281964 type=WRITE grant=WAIT_NEW timeoutMillis=2000 startTime=1538759059393 endTime=1538759061393

Owners: [<LockInfo locker="786774600 -1_fileChannelTaskExecutor-10_ThreadLocker" type="WRITE"/>]

Waiters:

[

<LockInfo locker="1270863098 -1_fileChannelTaskExecutor-9_ThreadLocker" type="WRITE"/>,

<LockInfo locker="1244207474 -1_fileChannelTaskExecutor-4_ThreadLocker" type="WRITE"/>,

<LockInfo locker="1391625295 -1_fileChannelTaskExecutor-1_ThreadLocker" type="WRITE"/>

]

at com.sleepycat.je.txn.LockManager.newLockTimeoutException(LockManager.java:608

at com.sleepycat.je.txn.LockManager.makeTimeoutMsgInternal(LockManager.java:567

at com.sleepycat.je.txn.SyncedLockManager.makeTimeoutMsg(SyncedLockManager.java:75

at com.sleepycat.je.txn.LockManager.lockInternal(LockManager.java:385

at com.sleepycat.je.txn.LockManager.lock(LockManager.java:272

at com.sleepycat.je.txn.BasicLocker.lockInternal(BasicLocker.java:134

at com.sleepycat.je.txn.Locker.lock(Locker.java:453

at com.sleepycat.je.dbi.CursorImpl.lockDupCountLN(CursorImpl.java:2768

at com.sleepycat.je.tree.Tree.insertDuplicate(Tree.java:2847

at com.sleepycat.je.tree.Tree.insert(Tree.java:2488

at com.sleepycat.je.dbi.CursorImpl.put(CursorImpl.java:1209

at com.sleepycat.je.Cursor.putAllowPhantoms(Cursor.java:1799

at com.sleepycat.je.Cursor.putNoNotify(Cursor.java:1756

at com.sleepycat.je.Cursor.putNotify(Cursor.java:1689

at com.sleepycat.je.Cursor.putInternal(Cursor.java:1626

at com.sleepycat.je.SecondaryDatabase.insertKey(SecondaryDatabase.java:984

at com.sleepycat.je.SecondaryDatabase.updateSecondary(SecondaryDatabase.java:909

at com.sleepycat.je.SecondaryTrigger.databaseUpdated(SecondaryTrigger.java:41

at com.sleepycat.je.Database.notifyTriggers(Database.java:2016

at com.sleepycat.je.Cursor.putNotify(Cursor.java:1702

at com.sleepycat.je.Cursor.putInternal(Cursor.java:1626

at com.sleepycat.je.Database.putInternal(Database.java:1186

at com.sleepycat.je.Database.put(Database.java:1058

at com.sleepycat.persist.PrimaryIndex.putNoReturn(PrimaryIndex.java:479

at com.sleepycat.persist.PrimaryIndex.putNoReturn(PrimaryIndex.java:442

at com.xyzq.bdb.cache.da.impl.BDBDataAccessor.create(BDBDataAccessor.java:77)

I tried the following

- Reduced the lock time. Hoping that the winner would release the lock as soon as its done insering records into

- Catch LockTimeOutException and retry inserting the keyword later. This works but it's manual and takes time. 10% of keywords fail because of a LockTimeOutException. ~ 10k out of 100k.

I have a few questions:

Is there a limit on the number of threads that can write to the DB or does it depend on the data?

Would it help if I model the entities as pure key value pairs instead of having secondary databases?

@Entity

Keyword

    @PrimaryKey

    - key

@Entity

Item

    -id

    -key [reference to Keyword Obj]

@Entity

Store

    - id

    - key [reference to Keyword obj.]

Thanks!

- K

Comments
Post Details
Added on Oct 9 2018
0 comments
281 views