补丁不生效例子
aix 7202 11.2.0.4 rac 180417 PSU
实例2 例行重启,实例1发生了意外重启
实例1告警日志
-
Fri Sep 16 00:34:57 2022
-
Set master node info
-
Submitted all remote-enqueue requests
-
Dwn-cvts replayed, VALBLKs dubious
-
All grantable enqueues granted
-
Fri Sep 16 00:34:59 2022
-
minact-scn: Master returning as live inst:2 has inc# mismatch instinc:0 cur:8 errcnt:0
-
Fri Sep 16 00:36:16 2022
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms2_38732932.trc (incident=1488512):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488512/orcl1_lms2_38732932_i1488512.trc
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Fri Sep 16 00:36:17 2022
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms0_19269850.trc (incident=1488496):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488496/orcl1_lms0_19269850_i1488496.trc
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Fri Sep 16 00:36:18 2022
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms1_5505754.trc (incident=1488504):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488504/orcl1_lms1_5505754_i1488504.trc
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Fri Sep 16 00:36:25 2022
-
Dumping diagnostic data in directory=[cdmp_20220916003625], requested by (instance=1, osid=19269850 (LMS0)), summary=[incident=1488496].
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms0_19269850.trc (incident=1488497):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488497/orcl1_lms0_19269850_i1488497.trc
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms2_38732932.trc (incident=1488513):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488513/orcl1_lms2_38732932_i1488513.trc
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Fri Sep 16 00:36:27 2022
-
Sweep [inc][1488513]: completed
-
Sweep [inc][1488512]: completed
-
Sweep [inc][1488504]: completed
-
Sweep [inc][1488497]: completed
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms1_5505754.trc (incident=1488505):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms1_5505754.trc (incident=1488506):
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
Use ADRCI or Support Workbench to package the incident.
-
See Note 411.1 at My Oracle Support for error and packaging details.
-
Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms1_5505754.trc (incident=1488507):
lms2、lms0、lms1三个进程都突然发生ORA-4030异常,内存不足,看相关trc文件
-
Dump file /u01/app/oracle/diag/rdbms/orcl/orcl1/incident/incdir_1488512/orcl1_lms2_38732932_i1488512.trc
-
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
-
With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,
-
Data Mining and Real Application Testing options
-
ORACLE_HOME = /u01/app/oracle/product/11.2.0/db_1
-
System name: AIX
-
Node name: db1
-
Release: 2
-
Version: 7
-
Machine: 00FB76794C00
-
Instance name: orcl1
-
Redo thread mounted by this instance: 1
-
Oracle process number: 15
-
Unix process pid: 38732932, image: oracle@db1 (LMS2)
-
-
-
*** 2022-09-16 00:36:16.929
-
*** SESSION ID:(1892.1) 2022-09-16 00:36:16.929
-
*** CLIENT ID:() 2022-09-16 00:36:16.929
-
*** SERVICE NAME:(SYS$BACKGROUND) 2022-09-16 00:36:16.929
-
*** MODULE NAME:() 2022-09-16 00:36:16.929
-
*** ACTION NAME:() 2022-09-16 00:36:16.929
-
-
Dump continued from file: /u01/app/oracle/diag/rdbms/orcl/orcl1/trace/orcl1_lms2_38732932.trc
-
ORA-04030: out of process memory when trying to allocate 195248 bytes (kjc pga smbuf,KJC PGA send buffer)
-
-
========= Dump for incident 1488512 (ORA 4030) ========
-
----- Beginning of Customized Incident Dump(s) -----
-
=======================================
-
TOP 10 MEMORY USES FOR THIS PROCESS
-
---------------------------------------
-
-
*** 2022-09-16 00:36:22.939
-
99% 31 GB, 172610 chunks: "KJC PGA send buffer " <--------------这里很异常
-
kjc pga smbuf ds=110733230 dsprt=110005120
-
1% 213 MB, 172724 chunks: "free memory "
-
pga heap ds=110005120 dsprt=0
-
0% 149 MB, 145 chunks: "permanent memory "
-
pga heap ds=110005120 dsprt=0
-
0% 75 KB, 3 chunks: "permanent memory "
-
session heap ds=1109d66e0 dsprt=11011f5d8
-
0% 54 KB, 1 chunk : "free memory "
-
top call heap ds=11011f3b8 dsprt=0
-
0% 38 KB, 1 chunk : "Fixed Uga "
-
Fixed UGA heap ds=1108bad28 dsprt=110005120
-
0% 20 KB, 2 chunks: "free memory "
-
session heap ds=1109d66e0 dsprt=11011f5d8
-
0% 16 KB, 7 chunks: "permanent memory "
-
diag pga ds=1107324d0 dsprt=110005120
-
0% 16 KB, 2 chunks: "dbgeInitProcessCtx:InvCtx "
-
diag pga ds=1107324d0 dsprt=110005120
-
0% 16 KB, 1 chunk : "kgh stack "
-
pga heap ds=110005120 dsprt=0
-
-
=======================================
-
PRIVATE MEMORY SUMMARY FOR THIS PROCESS
-
---------------------------------------
-
******************************************************
-
PRIVATE HEAP SUMMARY DUMP
-
32 GB total:
-
31 GB commented, 149 MB permanent
-
213 MB free (0 KB in empty extents),
-
32 GB, 1 heap: "kjc pga smbuf " 212 MB free held
-
------------------------------------------------------
-
Summary of subheaps at depth 1
-
31 GB total:
-
31 GB commented, 98 KB permanent
-
31 KB free (0 KB in empty extents),
-
31 GB, 172610 chunks: "KJC PGA send buffer "
-
-
*** 2022-09-16 00:36:23.380
-
-
=========================================
-
REAL-FREE ALLOCATOR DUMP FOR THIS PROCESS
-
-----------------------------------------
-
-
Dump of Real-Free Memory Allocator Heap [0x1109a4810]
-
mag=0xfefe0001 flg=0x5000003 fds=0x0 blksz=65536
-
blkdstbl=0x1109a4820, iniblk=521216 maxblk=524288 numsegs=320
-
In-use num=172723 siz=4028628992, Freeable num=1 siz=65536, Free num=310 siz=20316160
-
-
==========================================
-
INSTANCE-WIDE PRIVATE MEMORY USAGE SUMMARY
-
------------------------------------------
-
-
Dumping Work Area Table (level=1)
-
=====================================
-
-
Global SGA Info
-
---------------
-
-
global target: 115507 MB
-
auto target: 11333 MB
-
max pga: 2048 MB
-
pga limit: 4096 MB
-
pga limit known: 0
-
pga limit errors: 0
-
-
pga inuse: 102955 MB
-
pga alloc: 105793 MB
-
pga freeable: 742 MB
-
pga freed: 43967956 MB
-
pga to free: 0
-
broker request: 0
-
-
pga auto: 41 MB
-
pga manual: 0 MB
-
-
pga alloc (max): 105793 MB
-
pga auto (max): 5444 MB
-
pga manual (max): 1 MB
-
-
# workareas : 43
-
# workareas(max): 224
-
-
-
-
================================
-
PER-PROCESS PRIVATE MEMORY USAGE
-
--------------------------------
-
-
Private memory usage per Oracle process
-
-
-------------------------
-
Top 10 processes:
-
-------------------------
-
(percentage is of 103 GB total allocated memory)
-
31% pid 13: 32 GB used of 32 GB allocated (128 KB freeable)
-
31% pid 14: 32 GB used of 32 GB allocated
-
31% pid 15: 32 GB used of 32 GB allocated (64 KB freeable) <= CURRENT PROC
-
0% pid 12: 14 MB used of 113 MB allocated (98 MB freeable)
-
0% pid 10: 45 MB used of 56 MB allocated (7680 KB freeable)
-
0% pid 86: 42 MB used of 45 MB allocated
-
0% pid 87: 42 MB used of 45 MB allocated
-
0% pid 89: 42 MB used of 45 MB allocated
-
0% pid 973: 42 MB used of 43 MB allocated (1024 KB freeable)
-
0% pid 1029: 42 MB used of 43 MB allocated (1024 KB freeable)
-
-
-------------------------
-
All processes:
-
-------------------------
-
(session detail when over 5290 MB allocated)
-
-
pid 2: 2119 KB used of 2350 KB allocated
-
pid 3: 1060 KB used of 1182 KB allocated
-
pid 4: 1056 KB used of 1182 KB allocated
-
pid 5: 1066 KB used of 1182 KB allocated
-
pid 6: 6161 KB used of 6494 KB allocated (128 KB freeable)
-
pid 7: 2121 KB used of 2718 KB allocated (320 KB freeable)
-
...
三个进程每个都32G,搜索MOS,发现有个bug很像
检查当前的补丁,opatch lsinv|grep 18316692 已经包含了此修复,但是LMS的pga还是很大,看来没有解决。
怎么办呢?
监控,发现内存占用太大就重启实例(LMS进程杀不得)
-
select pid,spid,program,pga_used_mem,pga_alloc_mem,pga_freeable_mem,pga_max_mem from v$process where program like '%LMS%';
lms进程的工作机制
参考:
-
https://orainternals.files.wordpress.com/2008/03/riyaj_battle_of_nodes_rac_perf_myths_ppt.pdf
阅读(1415) | 评论(0) | 转发(0) |