AIX 的内存使用诊断问题?

从监控上发现数据库服务器内存已经被使用完,AIX接触的不多,分析思路不是很明确,请各位大神帮忙提供一下思路方法,谢谢!oslevel -s6100-08-03-1339topas:on Aug 14 17:03:48 2017 Interval: 2 Cswitch 23073 Readch 22.0M ...显示全部

从监控上发现数据库服务器内存已经被使用完,AIX接触的不多,分析思路不是很明确,请各位大神帮忙提供一下思路方法,谢谢!

oslevel -s

6100-08-03-1339
topas:
on Aug 14 17:03:48 2017 Interval: 2 Cswitch 23073 Readch 22.0M

                                            Syscall   25390  Writech   16.1M

CPU User% Kern% Wait% Idle% Reads 8213 Rawin 0
ALL 0.5 0.7 0.1 98.8 Writes 285 Ttyout 462

                                            Forks         0  Igets         0

Network KBPS I-Pack O-Pack KB-In KB-Out Execs 0 Namei 52
Total 12.8K 8969.4 2319.4 12.7K 137.7 Runqueue 1.0 Dirblk 0

                                            Waitqueue   0.5

Disk Busy% KBPS TPS KB-Read KB-Writ MEMORY
Total 0.4 16.6K 284.0 0.0 16.6K PAGING Real,MB 513280

                                            Faults       25  % Comp     55

FileSystem KBPS TPS KB-Read KB-Writ Steals 771 % Noncomp 44
Total 25.8K 704.4 9.7K 16.1K PgspIn 0 % Client 44

                                            PgspOut       0

Name PID CPU% PgSp Owner PageIn 17 PAGING SPACE
oracle 6227900 0.2 73.0 orapor PageOut 4224 Size,MB 32768
oracle 25757322 0.1 66.9 oraxxdb Sios 4243 % Used 2
zabbix_a 17891428 0.1 0.6 zabbix % Free 98
oracle 40568054 0.1 8.1 oraxxdb NFS (calls/sec)
aioserve 5833816 0.0 0.4 oraxxdb SerV2 0 WPAR Activ 0
oracle 29819212 0.0 8.1 oraxxdb CliV2 0 WPAR Total 0
oracle 36307426 0.0 8.1 oraxxdb SerV3 0 Press: "h"-help
aioserve 13631546 0.0 0.4 oraxxdb CliV3 0 "q"-quit

svmon -uP -t 3|grep -p Pid|grep '^.*[0-9] '

36897384 oracle 72196683 9040 0 57758967 Y N N
6358142 oracle 72081330 9040 0 57760495 Y N N
35325796 oracle 72078864 9040 0 57760526 Y N N

ps vx |head -1 ;ps vx |grep -v PID |sort -rn +6 |head -10

  PID    TTY STAT  TIME PGIN  SIZE   RSS   LIM  TSIZ   TRS %CPU %MEM COMMAND

1640360 - A 17:58 587 216408 217484 xx 556 1076 0.0 0.0 /usr/s
2228494 - A 222:16 2890 93032 93320 xx 79 288 0.0 0.0 /var/o
1377170 - A 242:07 0 49792 49792 xx 0 0 0.0 0.0 j2pg
2164504 - A 0:19 455 30476 30548 xx 47 72 0.0 0.0 [cimserve]
1311094 - A 0:00 0 16832 16832 xx 0 0 0.0 0.0 dog
2359528 - A 8:30 110 1452 16104 xx 488 14652 0.0 0.0 /usr/s
1836656 - A 2:01 12 12812 12916 xx 74 104 0.0 0.0 /opt/f
1246414 - A 1:24 66 11112 11184 xx 47 72 0.0 0.0 /usr/b
1378202 - A 1:20 343 9500 9624 xx 85 124 0.0 0.0 /usr/l
1509358 - A 0:32 393 6996 7072 xx 20 76 0.0 0.0 /opt/i

nmon:
| Memory ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Physical PageSpace | pages/sec In Out | FileSystemCache |
|% Used 100.0% 1.8% | to Paging Space 0.0 0.0 | (numperm) 44.8% |
|% Free 0.0% 98.2% | to File System 30.5 331.5 | Process 50.3% |
|GB Used 501.0GB 0.6GB | Page Scans 0.0 | System 4.8% |
|GB Free 0.2GB 31.4GB | Page Cycles 0.0 | Free 0.0% |
|Total(GB) 501.2GB 32.0GB | Page Steals 0.0 | ------ |
| | Page Faults 34.0 | Total 100.0% |
|------------------------------------------------------------ | numclient 44.8% |
|Min/Maxperm 14956MB( 3%) 448695MB( 90%) <--% of RAM | maxclient 90.0% |
|Min/Maxfree 960 1088 Total Virtual 533.2GB | User 92.3% |
|Min/Maxpgahead 2 8 Accessed Virtual 276.0GB 51.8%| Pinned 7.5% |
| | lruable pages 127628768.0

收起
参与13

查看其它 2 个回答jxnxsdengyu的回答

jxnxsdengyujxnxsdengyu课题专家组系统工程师江西农信

这个是正常的啊,计算内存由Process 50.3% +System 4.8% 才55.1%啊,总内存中有(numperm) 44.8% 是要用于非计算内存,提升文件系统的IO效率的。

银行 · 2017-08-15

回答者

jxnxsdengyu
系统工程师江西农信
擅长领域: 存储灾备双活

jxnxsdengyu 最近回答过的问题

回答状态

  • 发布时间:2017-08-15
  • 关注会员:4 人
  • 回答浏览:1584
  • X社区推广