逻辑STANDBY报错

jxq0210

一台逻辑STANDBY  alert老是报如下错误：Wed Mar 14 10:03:13 2012
LSP0: rolling back apply server 3
LSP0: apply server 3 rolled back
LSP0: can't recover from rollback of multi-chunk txn, aborting..
LOGSTDBY Analyzer process P014 pid=55 OS id=6812 stopped
LOGSTDBY Apply process P015 pid=56 OS id=6814 stopped
LOGSTDBY Apply process P022 pid=63 OS id=6828 stopped
LOGSTDBY Apply process P017 pid=58 OS id=6818 stopped
LOGSTDBY Apply process P023 pid=64 OS id=6830 stopped
LOGSTDBY Apply process P019 pid=60 OS id=6822 stopped
LOGSTDBY Apply process P018 pid=59 OS id=6820 stopped
LOGSTDBY Apply process P020 pid=61 OS id=6824 stopped
LOGSTDBY Apply process P016 pid=57 OS id=6816 stopped
LOGSTDBY Apply process P024 pid=65 OS id=6832 stopped
LOGSTDBY Apply process P021 pid=62 OS id=6826 stopped
Wed Mar 14 10:03:13 2012
LOGSTDBY status: ORA-16246: User initiated abort apply successfully completed
LOGSTDBY status: ORA-16222: automatic Logical Standby retry of last action
LOGSTDBY status: ORA-16111: log mining and apply setting up
Wed Mar 14 10:03:13 2012
LOGMINER: Parameters summary for session# = 1
LOGMINER: Number of processes = 3, Transaction Chunk Size = 10
LOGMINER: Memory Size = 2048M, Checkpoint interval = 10240M
Wed Mar 14 10:03:13 2012
LOGMINER: krvxpsr summary for session# = 1
LOGMINER: StartScn: 0 (0x0000.00000000)
LOGMINER: EndScn: 0 (0x0000.00000000)
LOGMINER: HighConsumedScn: 33348354845 (0x0007.c3b7831d)
LOGMINER: session_flag 0x1
LOGMINER: session# = 1, reader process P000 started with pid=37 OS id=6806
LOGMINER: session# = 1, builder process P001 started with pid=53 OS id=6808
LOGMINER: session# = 1, preparer process P002 started with pid=54 OS id=6810
Wed Mar 14 10:03:14 2012
LOGMINER: Begin mining logfile for session 1 thread 1 sequence 30005, /archpay/S_STDBY/1_30005_703177248.arc

。。。。很多这样can't recover from rollback of multi-chunk txn, aborting..

同时后台trc文件一直在增长
[ bdump]$ ll | grep 6808
-rw-r-----   1 oracle     dba        2315045139 Mar 15 21:11 orapay_p001_6808.trc
[bdump]$ ll | grep 6808
-rw-r-----   1 oracle     dba        2315185055 Mar 15 21:14 orapay_p001_6808.trc
[ bdump]$ ll | grep 6808
-rw-r-----   1 oracle     dba        2315186633 Mar 15 21:14 orapay_p001_6808.trc

谢谢解答下。

Maclean Liu(刘相兵

orapay_p001_6808.trc TRACE 文件打包后上传

jxq0210

参照.........................

Maclean Liu(刘相兵

Cannot resolve kxidsqn for lcr:
++  LCR Dump Begin: 0xc0000009ccf89e70 - COMMIT
     op: 7, Original op: 7, baseobjn: 0, objn: 0, objv: 0
     DF: 0x00000002, MF: 0x00200000, MF2: 0x00000000
     PF: 0x00000000, PF2: 0x00000000
     MergeFlag: 0x03, FilterFlag: 0x00
     Id: 0, iotPrimaryKeyCount: 0, numChgRec: 0
     NumCrSpilled: 0
     RedoThread#: 1, rba: 0x00757f.0000b7e3.0010
scn: 0x0007.c594cbab
xid: 0x0002.02d.ffffffff parent xid: 0x0002.02d.ffffffff
     ncol: 0newcount: 0, oldcount: 0
     LUBA: 0x0.0.0.0.0


kxidsqn       XID SQN

ODM FINDING:

1. Starting Captures produces trace: Cannot Resolve Kxidsqn For Lcr
   
2. 
   
3. Applies to:
   
4. Oracle Server - Enterprise Edition
   
5. Information in this document applies to any platform.
   
6. Symptoms
   
7. 
   
8. When starting a capture process, trace files are being generated. The contents looks like this:
   
9. 
   
10. Cannot resolve kxidsqn for lcr:
    
11. ++ LCR Dump Begin: 0x700000108320d98
    
12. operation: 26, bobjn: 0 objn: 12986, objv: 0,
    
13. flags: 0xa10000, flags2: 0x00, MergeFlag: 0x3, Id : 0 numChgRec = 0
    
14. Thread#: 1 rba: 0x20852.000a653a.4c
    
15. scn: 0x0001.1f774ae2
    
16. xid: 0x012e.02c.ffffffff
    
17. newcount = 0, oldcount = 0
    
18. LUBA: 0x46.1180a19c.2c.50.32ba
    
19. 
    
20. 
    
21. 
    
22. 
    
23. 
    
24. 
    
25. Changes
    
26. Cause
    
27. 
    
28. This is an informational warning on capture restart from Unpublished Bug 8502190
    
29. The message can be ignored or suppressed by setting an event in the init.ora
    
30. 
    
31. Solution
    
32. 
    
33. To turn this particular warning trace off ,set the event 1349 in the init.ora file.
    
34. 
    
35. 
    
36.    event="1349 trace name context forever, level 1048576"
    
37. 
    
38. 
    
39. 
    
40. Then bounce the instance.

复制代码

如果LOGICAL STANDBY的使用正常的话 ，一般可以忽略这些警告 ，也可以通过 设置1349 EVENT来 关闭这些告警的产生：

ALTER SYSTEM SET event='1349 trace name context forever, level 1048576'   scope=spfile;

REBOOT INSTANCE