因为操作系统的一个bug, 我们对一个两节点的RAC做了一次切换以升级操作系统,导致有些应用在两个节点上都有连接,随后的几天经常在两个节点上的 trace 文件中发现如下的信息:
/oracle/app/admin/pcendb/bdump/******_m000_1982704.trc
Oracle Database 10g Enterprise Edition Release 10.2.0.5.0 - 64bit Production
With the Partitioning, Real Application Clusters, Oracle Label Security, OLAP,
Data Mining and Real Application Testing options
ORACLE_HOME = /oracle/app/product/10.2.0
System name: AIX
Node name: *****
Release: 1
Version: 6
Machine: 00C23A554C00
Instance name: *****
Redo thread mounted by this instance: 1
Oracle process number: 42
Unix process pid: 1982704, image: oracle@***** (m000)
*** ACTION NAME:(Monitor Tablespace Thresholds) 2011-03-10 11:16:12.181
*** MODULE NAME:(MMON_SLAVE) 2011-03-10 11:16:12.181
*** SERVICE NAME:(SYS$BACKGROUND) 2011-03-10 11:16:12.181
*** SESSION ID:(1615.15357) 2011-03-10 11:16:12.181
SKGXPSEGRCV: MESSAGE TRUNCATED user data 56 bytes payload 2096 bytes
SKGXPSEGRCV: trucated message buffer data skgxpmsg meta data header 0xffffffffffed530 len 56 bytes
SKGXPLOSTACK: message truncation expected
SKGXPLOSTACK: data sent to port with no buffers queued from
SKGXPGPID ffffffffffed450 Internet address 192.168.174.2 UDP port number 41112
SKGXPLOSTACK: sent seq 32770 expecting 32771
SKGXPLOSTACK: lost ack detected retransmit ack
According to the message in Metalink thread 639798.999, "These SKGXPLOSTACK messages seen in your case are valid. They are informational messages and not error messages."
First letter S means OS layer. K: Oracle kernel. G: generic layer of Oracle. XP: cross process (i.e. Network IPC used by RAC). LOSTACK: sounds like low stack. Not sure what this function does. According to description for event 10402, turning on that event may give you more info.
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/12725904/viewspace-1050568/,如需转载,请注明出处,否则将追究法律责任。
转载于:http://blog.itpub.net/12725904/viewspace-1050568/