Originally posted by: SystemAdmin

Hi

We have p 570 ibm aix 5.3 server .it is being used for running oracle 9i database for oracle e business suite 11i

Aix server is having ram 32 gb with 8 processors

The problem is as database is running the paging space keeps on increasing

Pl advice how to avoid this problem

Koti Reddy

Aix and db admin

Originally posted by: MarkTaylor

depending on you level of aix 5.3 you should make sure that you protect computational pages first

ref: http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TD103148

Once this is done and paging space usage is still an issue, then you need to work out if you are just overcommiting memory, so paste the following info when you system is under load:

svmon
vmstat 1 20
vmstat -v
vmo -a | egrep "lru|max"

Basically, you need to make sure the AVM is smaller than realmem or you will page.

AVM == active virtual memory i.e. Oracle SGA / PGA, mbufs, kernel, app procs etc..

HTH
Mark Taylor

Originally posted by: SystemAdmin

It might also help checking the below.

• No mirror on paging space volumes
• Try use single hdisk for paging space. Even if your sever does paging, it will not page-in/out across different hdisks.

Interested to see your VMO settings..

Originally posted by: SystemAdmin

I have similar concern but my lsps -a maintaining around %38 all the time and I/O fsbuf increasing from 5960 a week agon to 6907 of today. Here are the results of those check on my 5.3L, with 4G memory, around 10 AM, kind of busy hours here.

1. svmon

size inuse free pin virtual
memory 970752 963831 6921 223994 1137564
pg space 1605632 622554

work pers clnt
pin 223994 0 0
in use 685166 0 278665

1. vmstat 1 20

System configuration: lcpu=4 mem=3792MB

kthr memory page faults cpu

---

-----------

---

------------

---

r b avm fre re pi po fr sr cy in sy cs us sy id wa
0 0 1137666 6806 0 0 0 0 0 0 34 1395 122 0 0 99 1
0 0 1137667 6801 0 2 0 0 0 0 63 2361 532 3 0 96 1
0 0 1137667 6801 0 0 0 0 0 0 38 1256 179 1 0 99 0
0 0 1137667 6801 0 0 0 0 0 0 46 1100 138 0 0 99 1
0 0 1137757 6709 0 0 0 0 0 0 35 3877 1070 4 1 94 1
0 0 1137757 6705 0 0 0 0 0 0 58 1037 145 1 0 96 3
1 0 1137757 6705 0 0 0 0 0 0 58 1214 171 3 1 96 1
0 0 1137757 6705 0 0 0 0 0 0 32 1169 114 1 0 99 0
1 0 1138235 6213 0 2 0 0 0 0 64 67848 212 11 3 84 2
0 0 1137832 8417 0 5 43 1807 4238 0 122 5294 678 8 7 83 3
0 0 1137804 8443 0 0 0 0 0 0 54 3462 1052 3 1 96 0
0 0 1137831 8415 0 0 0 0 0 0 25 58966 151 12 2 86 0
2 0 1137822 8424 0 0 0 0 0 0 46 12513 163 8 5 86 1
0 0 1137822 8424 0 0 0 0 0 0 39 1215 168 1 0 99 0
0 0 1137822 8424 0 0 0 0 0 0 36 1299 215 1 0 99 0
0 0 1144032 2214 0 0 0 0 0 0 56 2146 120 1 5 93 1
0 0 1137822 8424 0 0 0 0 0 0 64 2808 143 3 1 96 0
0 0 1137867 8379 0 0 0 0 0 0 60 1418 218 3 0 97 0
0 0 1137867 8379 0 0 0 0 0 0 30 1167 190 0 0 98 1
0 0 1137867 8379 0 0 0 0 0 0 45 1155 144 0 0 99 0

1. vmstat -v

970752 memory pages
929021 lruable pages
8342 free pages
0 memory pools
223994 pinned pages
80.0 maxpin percentage
20.0 minperm percentage
80.0 maxperm percentage
28.4 numperm percentage
264124 file pages
0.0 compressed percentage
0 compressed pages
29.8 numclient percentage
80.0 maxclient percentage
276979 client pages
0 remote pageouts scheduled
42289 pending disk I/Os blocked with no pbuf
18124227 paging space I/Os blocked with no psbuf
2740 filesystem I/Os blocked with no fsbuf
83 client filesystem I/Os blocked with no fsbuf
6907 external pager filesystem I/Os blocked with no fsbuf

1. vmo -a | egrep "lru|max"

lru_file_repage = 1
lru_poll_interval = 10
lrubucket = 131072
maxclient% = 80
maxfree = 1088
maxperm = 743216
maxperm% = 80
maxpin = 784258
maxpin% = 80
npsrpgmax = 100352
npsscrubmax = 100352
strict_maxclient = 1
strict_maxperm = 0
with the 1904 M/512 M SGA/PGA on my Oracle 10g R2.
Which one above is the realmem?