17

我试图了解我们在挂起的 Java 进程中遇到的问题。这个过程已经在生产中运行了大约 4 个月,本周早些时候它开始挂起。当我查看进程的线程转储时,所有相关线程 (3) 都有如下堆栈:

    "TxnParser_1" prio=6 tid=0x69bd3400 nid=0x2534 runnable [0x6aa2f000]
   java.lang.Thread.State: RUNNABLE
        at java.net.SocketInputStream.socketRead0(Native Method)
        at java.net.SocketInputStream.read(SocketInputStream.java:129)
        at oracle.net.ns.Packet.receive(Unknown Source)
        at oracle.net.ns.DataPacket.receive(Unknown Source)
        at oracle.net.ns.NetInputStream.getNextPacket(Unknown Source)
        at oracle.net.ns.NetInputStream.read(Unknown Source)
        at oracle.net.ns.NetInputStream.read(Unknown Source)
        at oracle.net.ns.NetInputStream.read(Unknown Source)
        at oracle.jdbc.driver.T4CMAREngine.unmarshalUB1(T4CMAREngine.java:1099)
        at oracle.jdbc.driver.T4CMAREngine.unmarshalSB1(T4CMAREngine.java:1070)
        at oracle.jdbc.driver.T4C8Oall.receive(T4C8Oall.java:478)
        at oracle.jdbc.driver.T4CStatement.doOall8(T4CStatement.java:207)
        at oracle.jdbc.driver.T4CStatement.executeForDescribe(T4CStatement.java:790)
        at oracle.jdbc.driver.OracleStatement.executeMaybeDescribe(OracleStatement.java:1039)
        at oracle.jdbc.driver.T4CStatement.executeMaybeDescribe(T4CStatement.java:830)
        at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1132)
        at oracle.jdbc.driver.OracleStatement.executeInternal(OracleStatement.java:1687)
        at oracle.jdbc.driver.OracleStatement.execute(OracleStatement.java:1653)
        - locked <0x40e22f88> (a oracle.jdbc.driver.T4CStatement)
        - locked <0x28f8d398> (a oracle.jdbc.driver.T4CConnection)
        at com.gcg.data.LogParsingInfo.initFromDB(LogParsingInfo.java:262)
        at com.gcg.om.OmQueueEntry.initParseInfoFromDB(OmQueueEntry.java:104)
        at com.gcg.om.GenericQueueEntry.run(GenericQueueEntry.java:237)
        at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
        at java.lang.Thread.run(Thread.java:619)

没有等待锁的线程,因此进程不会死锁。这 3 个正在执行工作的线程只是被阻塞等待 Oracle 的响应,至少在我看来是这样的。

查看 Oracle,当我查询 v$session 时,看起来与这些线程关联的连接之一当前正在执行查询,尽管我看不到 sql。

select ... from v$session where ...;
SQL_ADDRESS      SQL_HASH_VALUE SQL_ID        SQL_CHILD_NUMBER SQL_EXEC_START SQL_EXEC_ID PREV_SQL_ADDR    PREV_HASH_VALUE PREV_SQL_ID   PREV_CHILD_NUMBER PREV_EXEC_START PREV_EXEC_ID
---------------- -------------- ------------- ---------------- -------------- ----------- ---------------- --------------- ------------- ----------------- --------------- ------------
              00              0                                                           0000000239F59EE8      1483377872 fqr8pndc6p36h                 5 26-JUL-12           32080545
              00              0                                                           0000000239F59EE8      1483377872 fqr8pndc6p36h                 5 26-JUL-12           32080546
0000000148CABD88     1784444892 a16hxxtp5sxyw                                             0000000239F59EE8      1483377872 fqr8pndc6p36h                 5 26-JUL-12           32080544

select * from v$sql where sql_id = 'a16hxxtp5sxyw';

no rows selected

我的问题是:

  1. 我的分析是否正确,该进程只是被阻止等待 Oracle 的响应?
  2. 我应该在 Oracle 中寻找什么来理解为什么这个进程被阻塞?

更新:

基于关于查看 DBA_WAITERS 和 DBA_LOCKS 的评论

select * from dba_waiters;

no rows selected

select * from dba_locks where BLOCKING_OTHERS <> 'Not Blocking';

no rows selected 

dba_locks 中有 98 行,但由于所有行都“不阻塞”,我认为这不是锁定问题吗?有问题的进程已处于此状态超过 3 小时,因此现在已检测到任何死锁。

我认为 Oracle 实例不是“健康的”,但我不知道该看什么。我有一个重新启动 Oracle 服务器的请求,但尚未完成。

追问:v$session包含一个在v$sql中不存在的sql_id是否正常,如果是,在什么条件下?

4

2 回答 2

9

问题解决了,答案就在 v$session 表中。显然,Oracle 会话可能会因为锁定以外的原因而阻塞。请注意列 FINAL_BLOCKING_SESSION - 它标识了作为阻塞根本原因的会话。我们调查了会话 845,发现客户端进程(由 MACHINE 和 PORT 标识)不再存在。DBA 终止了会话 845 并全部恢复正常。

SID     SERIAL# STATUS    PROGRAM          TYPE SQL_ID        PREV_SQL_ID    BLOCKING_SESSION_STATUS BLOCKING_INSTANCE BLOCKING_SESSION FINAL_BLOCKING_SESSION_STATUS FINAL_BLOCKING_INSTANCE FINAL_BLOCKING_SESSION EVENT
------- ------- --------- ---------------- ---- ------------- -------------- ----------------------- ----------------- ---------------- ----------------------------- ----------------------- ---------------------- ----------------------------
 108    22447   ACTIVE    Gcg log parser 1 USER               fqr8pndc6p36h  VALID                   1                 1581             VALID                         1                       845                    library cache: mutex X
 639    40147   ACTIVE    Gcg log parser 3 USER               fqr8pndc6p36h  VALID                   1                 1581             VALID                         1                       845                    library cache: mutex X
 742    34683   ACTIVE    Gcg log parser 2 USER a16hxxtp5sxyw fqr8pndc6p36h  VALID                   1                 1581             VALID                         1                       845                    library cache: mutex X
于 2012-07-27T14:42:54.007 回答
2

我最近也遇到了这个问题,并使用此查询查找 Oracle 中的锁定/锁定会话:

select 
   inst_id||' '||sid||','||serial# inst_sid_s#, 
   username,
   row_wait_obj#||','||row_wait_block#||','||row_wait_row# obj_lck,
   blocking_session_Status||' '||blocking_instance||','||blocking_session blk_info,
   final_blocking_session_Status||' '||final_blocking_instance||','||final_blocking_session f_blk_info,
   event, 
   seconds_in_wait 
from 
   gv$session 
where 
   lockwait is not null
order by 
   inst_id;

来源: http ://www.dba-oracle.com/t_final_blocking_session_final_blocking_instance.htm

于 2016-04-11T04:50:03.877 回答