Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations SkipVought on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

vmtune for Informix

Status
Not open for further replies.

MoshiachNow

IS-IT--Management
Feb 6, 2002
1,851
0
0
IL
HI,
Hving a 6 cpu 6F1 with 4GB memory,plus heavy Informix DB reporting.
What would be the recommended vmtune changes ?
Below are my current setting.


-p -P -r -R -f -F -N -W
minperm maxperm minpgahead maxpgahead minfree maxfree pd_npages maxrandwrt
209505 838020 16 128 120 248 524288 0

-M -w -k -c -b -B -u -l -d
maxpin npswarn npskill numclust numfsbufs hd_pbuf_cnt lvm_bufcnt lrubucket defps
838841 42496 10624 1 93 272 9 131072 1

-s -n -S -L -g -h
sync_release_ilock nokilluid v_pinshm lgpg_regions lgpg_size strict_maxperm
0 0 0 0 0 0

Long live king Moshiach !
 
Hi

What is your AIX Release ?
What Type of Filesystem do you use ? ( JFS, JFS2 )
Can you send the following Data:

vmstat 5 12
vmstat -v
lsps -a
iostat 1 1 ( Is I/O Stat active on sys0 ? )
 
Hi,

It's aaaaaix 4.3.3,ML10. jfs ,not jfs2.

vmstat 5 12
kthr memory page faults cpu
----- ----------- ------------------------ ------------ -----------
r b avm fre re pi po fr sr cy in sy cs us sy id wa
2 3 359223 136 0 2 2 1602 4403 0 2543 7549 6145 15 11 59 16
1 4 359232 540 0 0 0 695 1028 0 2659 36173 7514 6 8 56 31
1 5 359232 471 0 0 0 180 426 0 3486 42189 8256 6 8 52 33
1 4 359501 147 0 0 0 308 434 0 3432 51884 12312 7 8 52 33
2 5 359448 186 0 0 0 1811 5073 0 3624 54796 10371 6 11 43 40
1 5 359289 334 0 0 0 3170 5945 0 2801 61000 10860 10 18 37 36
1 5 359289 142 0 0 3 2811 4426 0 2456 73564 9182 8 17 40 35
1 4 359289 165 0 0 2 2497 3347 0 1598 60892 6447 7 12 51 30
1 4 359289 197 0 0 0 3304 4900 0 1592 46615 5794 5 9 53 33
1 3 359637 198 0 0 0 2687 13447 0 1644 50377 7609 10 12 52 26
2 3 359553 0 0 0 0 2380 3513 0 1602 64274 11250 18 15 52 15
3 3 359631 145 0 0 0 7792 9368 0 1562 68122 8352 18 18 49 15

vmstat -v

lsps -a

iostat 1 1 root@SMP:/scitex/version/logs 100 % vmstat -v
Usage: vmstat [ -fsi ] [Drives] [Interval] [Count]

root@SMP:/scitex/version/logs 101 % lsps -a

Page Space Physical Volume Volume Group Size %Used Active Auto Type
hd6 hdisk0 rootvg 5312MB 5 yes yes lv
root@SMP:/scitex/version/logs 102 % iostat 1 1 iostat 1 1

tty: tin tout avg-cpu: % user % sys % idle % iowait
0.1 777.2 14.7 10.6 58.9 15.8

Disks: % tm_act Kbps tps Kb_read Kb_wrtn
hdisk0 16.2 793.9 24.3 54443876 203989209
hdisk5 15.2 771.4 22.1 48134213 202964613
hdisk4 0.0 0.1 0.0 37677 1620
hdisk1 2.7 344.8 8.8 110005731 2220600
hdisk2 5.4 211.5 34.7 65248662 3605204
hdisk3 0.6 88.0 19.2 26381505 2262576
cd0 0.0 0.0 0.0 0 0
hdisk6 4.5 359.7 19.8 90085433 27020516
hdisk7 1.6 222.0 5.9 60517777 11749920
hdisk8 7.6 521.6 26.5 146300745 23493352
hdisk9 8.1 727.3 33.8 189275941 47485356
hdisk12 2.2 32.9 7.5 7715545 3000528
hdisk13 2.2 33.9 7.9 9232173 1793368
hdisk10 6.6 683.9 24.6 189614789 33023428
hdisk11 4.5 434.2 18.7 115199685 26139320
root@SMP:/scitex/version/logs 103 %

Long live king Moshiach !
 
Hi

Try minperm to 10%, maxperm to 20% if it will be better, you can try minperm to 5%, maxperm to 10%
I always wait for a day to see what the new value will bring.
Are you happy with your minpgahead, maxpgahead Settings ?
They are high and will improve the I/O Performance on big File I/O. Maybe you can increase the numfsbufs to 200

Good Luck
 
Thanks.
A question - why should I decrease minperm and maxperm (originally were 20% and 80% repectfully) - if I do not have any more paging to the paging space,with a given 4GB memory ?
Regarding the minpageahead and maxpageahead - yes,I have increased these to ipmrove massive sequentual reading form the RAID volumes.Do you think decreasing these back may improve my current situation ?

Long live king Moshiach !
 
Hi

Your CPU is Waiting, so you can decrease the Values of minperm,maxperm to force the fsbufs to write down the Data efficintly....
Its difficult to say what is to do with your pgaheads, you need to test some Values.

Do you have hot LVs ?
I prefer to monitor the sytem with the Nmon collector and after one Day i see a good overview where my problem is.
search in the Web for nmon and nmon Analyser
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top