背景：
XX系統實施一段時間之后，出現
數據庫連接池滿，第一次通過修改if(con!=null && con.isClosed()){con.close();}這樣的邏輯
錯誤解決部分
問題。第二次通過徹底復查
代碼，修改了connection、session沒有釋放的問題，基本上保證我們自己寫的代碼沒有數據庫連接不釋放的問題。但是臨近近期還是出現連接池滿的問題。。。
過程：
從日志看，除了有大量工作流
報錯之外程序很少有異常，類似如下：
引用:
2009-06-12 15:44:34,187 [http-80-Processor44] [org.hibernate.event.def.AbstractFlushingEventListener] [ERROR] - Could not synchronize database state with session
org.hibernate.StaleObjectStateException: Row was updated or deleted by another transaction (or unsaved-value mapping was incorrect): [org.jbpm.graph.exe.Token#35000000000033432]
..............................................
at org.jbpm.persistence.db.DbPersistenceService.close(DbPersistenceService.java:180)
..............................................
2009-06-12 15:44:34,187 [http-80-Processor44] [org.jbpm.svc.Services] [ERROR] - problem closing service 'persistence'
org.jbpm.persistence.JbpmPersistenceException: couldn't flush hibernate session
at org.jbpm.persistence.db.DbPersistenceService.close(DbPersistenceService.java:182)
Caused by: org.hibernate.StaleObjectStateException: Row was updated or deleted by another transaction (or unsaved-value mapping was incorrect): [org.jbpm.graph.exe.Token#35000000000033432]
at org.jbpm.persistence.db.DbPersistenceService.close(DbPersistenceService.java:180)
... 54 more
最開始基本確定了是工作流報錯導致數據庫連接池不釋放，理由：
a、上面的錯和hibernate的session有關
b、在
sybase執行sp_who發現大量不釋放連接所占用的庫為DB_LC，而這個庫就是工作流相關的庫。
c、從sybase的sysprocesses表查看，不釋放連接是每天新增10-30不等，隨機統計了日志某天的如前所述的異常為27個，而從數據庫端統計該天新增連接也是27個。
因為自己對工作流不熟悉，所以每次都是把情況反映給相關人員處理。前幾天去客戶現場正好抓取了一下不釋放連接正在執行的
sql，基本都是
亂碼，如下：
引用:
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
SQL Text: !
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
(1 row affected)
47
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
SQL Text: *
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
(1 row affected)
49
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
SQL Text:
DBCC execution completed. If DBCC printed error messages, contact a user with System Administrator (SA) role.
(1 row affected)
這個結果用處不大，很好奇這個問題，所以找了一份工作流的源碼，找到報錯的類DbPersistenceService.close方法，如下：
復制內容到剪貼板
代碼:
public void close() {
if ( (session!=null)
&& (transaction==null)
&& (isRollbackOnly)
) {
throw new JbpmException("setRollbackOnly was invoked while configuration specifies user managed transactions");
}
if (messagingSession!=null) {
messagingSession.closeOpenIterators();
}
if (schedulerSession!=null) {
schedulerSession.closeOpenIterators();
}
if ( (isTransactionEnabled)
&& (transaction!=null)
) {
if (isRollbackOnly) {
try {
log.debug("rolling back hibernate transaction");
mustSessionBeFlushed = false; // flushing updates that will be rolled back is not very clever :-)
transaction.rollback();
} catch (Exception e) {
throw new JbpmPersistenceException("couldn't rollback hibernate session", e);
}
} else {
try {
log.debug("committing hibernate transaction");
mustSessionBeFlushed = false; // commit does a flush anyway
transaction.commit();
} catch (Exception e) {
try {
// if the commit fails, we must do a rollback
transaction.rollback();
} catch (Exception e2) {
// if the rollback fails, we did what we could and you're in
// deep shit :-(
log.error("problem rolling back after failed commit", e2);
}
throw new JbpmPersistenceException("couldn't commit hibernate session", e);
}
}
}
if (mustSessionBeFlushed) {
try {
log.debug("flushing hibernate session");
session.flush();
} catch (Exception e) {
throw new JbpmPersistenceException("couldn't flush hibernate session", e);
}
}
if (mustSessionBeClosed) {
try {
log.debug("closing hibernate session");
session.close();
} catch (Exception e) {
throw new JbpmPersistenceException("couldn't close hibernate session", e);
}
}
if (mustConnectionBeClosed) {
try {
log.debug("closing jdbc connection");
connection.close();
} catch (Exception e) {
throw new JbpmPersistenceException("couldn't close jdbc connection", e);
}
}
}
一看真是嚇一跳，程序在執行到session.flush();時候報錯的話，如果mustSessionBeClosed為true根本不能執行到后面的session.close()，會導致數據庫連接不釋放的問題……基本確定問題所在了，就在本地試著復現一下問題（因為前面所描述的異常在
開發環境無法復現，所以只能強制在flush后拋異常），果然不出意料。
因為這個是jbpm3.1.2版本，覺得應該是jbpm的
bug吧，就又下載了一份jbpm3.3.0GA源碼，找到DbPersistenceService.close()方法：
復制內容到剪貼板
代碼:
public void close() {
if ( (session!=null)
&& !isTransactionActive()
&& (isRollbackOnly())
) {
throw new JbpmException("setRollbackOnly was invoked while configuration specifies user managed transactions");
}
if ( (isTransactionEnabled)
&& (transaction!=null)
) {
if (! isRollbackOnly()) {
Exception commitException = commit();
if (commitException!=null) {
rollback();
closeSession();
closeConnection();
throw new JbpmPersistenceException("hibernate commit failed", commitException);
}
} else { // isRollbackOnly==true
Exception rollbackException = rollback();
if (rollbackException!=null) {
closeSession();
closeConnection();
throw new JbpmPersistenceException("hibernate rollback failed", rollbackException);
}
}
}
Exception flushException = flushSession();
if (flushException!=null) {
// JBPM-1465 transaction has been either committed or rolled back at this point
// on the other hand, it is possible that no transaction is underway
// hence rolling back here is redundant and possibly dangerous
closeSession();
closeConnection();
throw new JbpmPersistenceException("hibernate flush failed", flushException);
}
Exception closeSessionException = closeSession();
if (closeSessionException!=null) {
closeConnection();
throw new JbpmPersistenceException("hibernate close session failed", closeSessionException);
}
Exception closeConnectionException = closeConnection();
if (closeConnectionException!=null) {
throw new JbpmPersistenceException("hibernate close connection failed", closeConnectionException);
}
}
果然在3.3.0版本中，當flush、close等操作出現異常時候，都會調用closeSession()和closeConnection()以保證連接正常釋放。照貓畫虎在該方法寫了關閉session和connection的方法，準備月底發布新版本試試。
結論：
XX系統工作流jbpm3.1.2存在連接不釋放的bug，當然前提是程序執行數據庫操作報錯的情況下（如session.flush）。雖然解決了連接不釋放的問題，但是這個關于這個報錯的深層原因還沒搞清楚。另外和相關人員確認，工作流的這些異常可以cacth掉，到目前為止除了引起連接不釋放之外，沒有發現其他問題。
posted on 2009-06-22 17:38
歲月如歌 閱讀(1865)
評論(0) 編輯 收藏 所屬分類:
java