IDS Forum

Re: Buffer Turnover Ratio (BTR)

Posted By: Art Kagel
Date: Thursday, 23 April 2009, at 8:17 a.m.

In Response To: Re: Buffer Turnover Ratio (BTR) (ANEES AHMAD)

You obviously have something other than a buffer cache problem going on in
the instance#1. You can try posting other information and we'll try to
help, but I suspect that this is going to take a deeper analysis. If you
are interested in having someone look into this for you, contact me directly
to my email (art@oninit.com) and we'll work that out.

Meanwhile, post the following:

onstat -p
onstat -P
onstat -D
onstat -d
onstat -g iov
onstat -g glo
onstat -c
onstat -g iof

The community will take a look and see if there's anything obvious there and
try to help.

Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)

Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Oninit, the IIUG, nor any other organization
with which I am associated either explicitly or implicitly. Neither do
those opinions reflect those of other individuals affiliated with any entity
with which I am affiliated nor those of the entities themselves.

On Thu, Apr 23, 2009 at 7:17 AM, ANEES AHMAD <aanees@i2cinc.com> wrote:

> Art Thanks for the information.
>
> Consider the following two blocks of outputs from the instance-1 and
> instance-2 two.
>
> --------------------
> Instance-1
>
> BR=0.565
> RAU=98.6
> BTR=7.36
>
> BUFFERPOOL
>
> size=2K,buffers=120000,lrus=46,lru_min_dirty=50.000000,lru_max_dirty=60.000000
>
> --------------------
>
> Instance 2
>
> BR=1.99
> RAU=99.88
> BTR=2137
>
> BUFFERPOOL
> size=2K,buffers=6000,lrus=8,lru_min_dirty=50.000000,lru_max_dirty=60.000000
>
> ---------------------
>
> The point which is disturbing me a lot is that the Instance-2 is showing
> very
> good query/procedural performance where as the Instance-1 is showing very
> bad
> performance and DB requests are getting timed out.
> Does it mean that the bufferpool and LRU settings of Insatnce-1 are still
> not
> tuned? Why the Insatace-2 with too few buffers is showing good performance?
> The load of activities is very high at Instance-2 as compared to
> Instance-1.
> Say if Instance-2 is executing 15000 requests a day then Insartance-1 is
> just
> executing less than 500 requests a day.
> Where might be the bottle neck? Same application is deployed on both
> instances.
> The difference is only that
> Instance-1 is running with 4 databases with 821 maximum concurrent
> Instance-2 is running with only 1 database with 308 maximum concurrent.
>
> See the following outputs of onstat -p for both instances
>
> Instance-1
> IBM Informix Dynamic Server Version 11.50.FC3W1 -- On-Line (Prim) -- Up 19
> days 21:40:28 -- 1548288 Kbytes
>
> Profile
> dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached
> 62227661 362878768 5326958137 98.83 9279279 19347443 59636292 87.01
>
> isamtot open start read write rewrite delete commit rollbk
> 4369875770 204876627 363233569 2906055279 7540938 3295430 506526 3119073
> 492
>
> gp_read gp_write gp_rewrt gp_del gp_alloc gp_free gp_curs
> 0 0 0 0 0 0 0
>
> ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes
> 0 0 0 347692.07 42132.09 5704 5735
>
> bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
> 2390527 2412 4432863050 0 0 614 3508280 13477435
>
> ixda-RA idx-RA da-RA RA-pgsused lchwaits
> 10370317 53200 15847828 25901168 1221946
>
>
> -----------------------------------------------------------------------------
>
> Instance-2
>
> IBM Informix Dynamic Server Version 11.50.FC3W1 -- On-Line (Prim) -- Up 34
> days 20:46:59 -- 1527808 Kbytes
>
> Profile
> dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached
> 7668590970 10289375428 42917152205 82.13 45635154 136785986 442875348 90.43
>
> isamtot open start read write rewrite delete commit rollbk
> 39605469811 351650681 818427987 34798426922 40611296 15311936 6025699
> 19016271
> 6137
>
> gp_read gp_write gp_rewrt gp_del gp_alloc gp_free gp_curs
> 0 0 0 0 0 0 0
>
> ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes
> 0 0 0 3634949.86 865140.29 10836 11129
>
> bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
> 213692299 14652 42599333339 0 0 3572 30374193 103282201
>
> ixda-RA idx-RA da-RA RA-pgsused lchwaits
> 1445033543 1565761 701870740 2145951638 130851848
>
>
>
> *******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>

--001636c5b31486ae03046837dc20

Messages In This Thread

• Buffer Turnover Ratio (BTR)
  ANEES AHMAD -- Wednesday, 22 April 2009, at 3:28 a.m.
  • Re: Buffer Turnover Ratio (BTR)
    Art Kagel -- Wednesday, 22 April 2009, at 11:13 a.m.
    • Re: Buffer Turnover Ratio (BTR)
      ANEES AHMAD -- Thursday, 23 April 2009, at 7:17 a.m.
      • Re: Buffer Turnover Ratio (BTR)
        ANEES AHMAD -- Thursday, 23 April 2009, at 7:30 a.m.
      • Re: Buffer Turnover Ratio (BTR)
        Art Kagel -- Thursday, 23 April 2009, at 8:17 a.m.
        • Re: Buffer Turnover Ratio (BTR)
          ANEES AHMAD -- Thursday, 23 April 2009, at 10:37 a.m.
          • Re: Buffer Turnover Ratio (BTR)
            ANEES AHMAD -- Thursday, 23 April 2009, at 10:40 a.m.

IDS Forum is maintained by Administrator with WebBBS 5.12.