Dead-lock detected during TagService.tag operation

Bonjour

contexte: 5.8 + HF21 - Windows - DtB:H2 - Installation locale avec un seul utilisateur

J'ai créé une opération SetDocumentTag implémentée ainsi:

 public void run(DocumentModel input) throws Exception {
    TagService service = Framework.getLocalService(TagService.class);
    String docId = input.getId();
    log.error("SetDocumentTag> docId=" + docId + " tag=" + label + " username=" + username);

-> service.tag(session, docId, label, username);

    session.save();
}

De temps en temps j'ai l'erreur suivante dans l'appel service.tag:

Caused by: org.nuxeo.ecm.core.api.ConcurrentUpdateDocumentException: org.nuxeo.ecm.core.storage.ConcurrentUpdateStorageE
xception: org.h2.jdbc.JdbcBatchUpdateException: Deadlock detected. The current transaction was rolled back. Details:
Session #7 (user: SA) is waiting to lock PUBLIC.HIERARCHY while locking PUBLIC.CONTENT (exclusive), PUBLIC.FULLTEXT (exc
lusive).
Session #5 (user: SA) is waiting to lock PUBLIC.FULLTEXT while locking , PUBLIC.HIERARCHY (exclusive), PUBLIC.MISC (excl
usive), PUBLIC.RELATION (exclusive).; SQL statement:
INSERT INTO "FULLTEXT" ("ID", "JOBID", "SIMPLETEXT", "BINARYTEXT", "SIMPLETEXT_TITLE", "BINARYTEXT_TITLE", "SIMPLETEXT_D
ESCRIPTION", "BINARYTEXT_DESCRIPTION") VALUES (?, ?, ?, ?, ?, ?, ?, ?) [40001-114]
        at org.nuxeo.ecm.core.storage.sql.coremodel.SQLSession.save(SQLSession.java:159)
        at org.nuxeo.ecm.core.api.AbstractSession.save(AbstractSession.java:1908)
        ... 100 more
Caused by: org.nuxeo.ecm.core.storage.ConcurrentUpdateStorageException: org.h2.jdbc.JdbcBatchUpdateException: Deadlock d
etected. The current transaction was rolled back. Details:
Session #7 (user: SA) is waiting to lock PUBLIC.HIERARCHY while locking PUBLIC.CONTENT (exclusive), PUBLIC.FULLTEXT (exc
lusive).
Session #5 (user: SA) is waiting to lock PUBLIC.FULLTEXT while locking , PUBLIC.HIERARCHY (exclusive), PUBLIC.MISC (excl
usive), PUBLIC.RELATION (exclusive).; SQL statement:
INSERT INTO "FULLTEXT" ("ID", "JOBID", "SIMPLETEXT", "BINARYTEXT", "SIMPLETEXT_TITLE", "BINARYTEXT_TITLE", "SIMPLETEXT_D
ESCRIPTION", "BINARYTEXT_DESCRIPTION") VALUES (?, ?, ?, ?, ?, ?, ?, ?) [40001-114]
        at org.nuxeo.ecm.core.storage.sql.jdbc.JDBCConnection.checkConcurrentUpdate(JDBCConnection.java:323)
        at org.nuxeo.ecm.core.storage.sql.jdbc.JDBCRowMapper.insertSimpleRows(JDBCRowMapper.java:605)
        at org.nuxeo.ecm.core.storage.sql.jdbc.JDBCRowMapper.writeCreates(JDBCRowMapper.java:502)
        at org.nuxeo.ecm.core.storage.sql.jdbc.JDBCRowMapper.write(JDBCRowMapper.java:472)
        at org.nuxeo.ecm.core.storage.sql.SoftRefCachingRowMapper.write(SoftRefCachingRowMapper.java:427)
        at org.nuxeo.ecm.core.storage.sql.SessionImpl.flush(SessionImpl.java:391)
        at org.nuxeo.ecm.core.storage.sql.SessionImpl.flushAndScheduleWork(SessionImpl.java:365)
        at org.nuxeo.ecm.core.storage.sql.SessionImpl.save(SessionImpl.java:344)
        at org.nuxeo.ecm.core.storage.sql.ra.ConnectionImpl.save(ConnectionImpl.java:184)
        at org.nuxeo.ecm.core.storage.sql.coremodel.SQLSession.save(SQLSession.java:157)
        ... 101 more
Caused by: org.h2.jdbc.JdbcBatchUpdateException: Deadlock detected. The current transaction was rolled back. Details:
Session #7 (user: SA) is waiting to lock PUBLIC.HIERARCHY while locking PUBLIC.CONTENT (exclusive), PUBLIC.FULLTEXT (exc
lusive).
Session #5 (user: SA) is waiting to lock PUBLIC.FULLTEXT while locking , PUBLIC.HIERARCHY (exclusive), PUBLIC.MISC (excl
usive), PUBLIC.RELATION (exclusive).; SQL statement:
INSERT INTO "FULLTEXT" ("ID", "JOBID", "SIMPLETEXT", "BINARYTEXT", "SIMPLETEXT_TITLE", "BINARYTEXT_TITLE", "SIMPLETEXT_D
ESCRIPTION", "BINARYTEXT_DESCRIPTION") VALUES (?, ?, ?, ?, ?, ?, ?, ?) [40001-114]
        at org.h2.jdbc.JdbcPreparedStatement.executeBatch(JdbcPreparedStatement.java:1067)
        at org.apache.tomcat.dbcp.dbcp.DelegatingStatement.executeBatch(DelegatingStatement.java:297)
        at org.apache.tomcat.dbcp.dbcp.DelegatingStatement.executeBatch(DelegatingStatement.java:297)
        at org.nuxeo.ecm.core.storage.sql.jdbc.JDBCRowMapper.insertSimpleRows(JDBCRowMapper.java:590)
        ... 109 more
Caused by: org.h2.jdbc.JdbcSQLException: Deadlock detected. The current transaction was rolled back. Details:
Session #7 (user: SA) is waiting to lock PUBLIC.HIERARCHY while locking PUBLIC.CONTENT (exclusive), PUBLIC.FULLTEXT (exc
lusive).
Session #5 (user: SA) is waiting to lock PUBLIC.FULLTEXT while locking , PUBLIC.HIERARCHY (exclusive), PUBLIC.MISC (excl
usive), PUBLIC.RELATION (exclusive).; SQL statement:
INSERT INTO "FULLTEXT" ("ID", "JOBID", "SIMPLETEXT", "BINARYTEXT", "SIMPLETEXT_TITLE", "BINARYTEXT_TITLE", "SIMPLETEXT_D
ESCRIPTION", "BINARYTEXT_DESCRIPTION") VALUES (?, ?, ?, ?, ?, ?, ?, ?) [40001-114]
        at org.h2.message.Message.getSQLException(Message.java:105)
        at org.h2.message.Message.getSQLException(Message.java:116)
        at org.h2.message.Message.getSQLException(Message.java:75)
        at org.h2.table.TableData.doLock(TableData.java:426)
        at org.h2.table.TableData.lock(TableData.java:373)
        at org.h2.command.dml.Insert.update(Insert.java:100)
        at org.h2.command.CommandContainer.update(CommandContainer.java:72)
        at org.h2.command.Command.executeUpdate(Command.java:208)
        at org.h2.jdbc.JdbcPreparedStatement.executeUpdateInternal(JdbcPreparedStatement.java:139)
        at org.h2.jdbc.JdbcPreparedStatement.executeBatch(JdbcPreparedStatement.java:1050)
        ... 112 more

Comment protéger cet appel contre les transactions non terminées?

Il semble que des événements ne sont pas encore traités au moment de l'appel au TagService …

Merci de votre aide

Christian

0 votes

1 answers

964 views

ANSWER



La base embedded h2 gère très mal la concurrence et n'est donc pas adaptée à quelque sorte de production que ce soit. Nous vous conseillons d'utiliser PostgreSQL par exemple.

0 votes



Bonjour

Merci pour cette réponse. En effet c'est bien ce que je pensais, en effet, en tournant sur une autre plateforme (non locale) avec PostGres, je n'observe pas le problème

Merci Christian

10/27/2014

Désolé je reviens sur ce que je viens de dire:

j'observe cette erreur aussi sur la base utilisant PstGres.

C'est très rare: lors d'un import massif, créant 1000 documents, j'ai 2 occurrences de l'erreur

Christian

10/27/2014

Pouvez-vous mettre à jour la question initiale avec un lien (pastebin) sur le log d'erreur ?
10/27/2014