Re: FW: Performance Issues
Posted in 2000
Hi;
I'd like to put some input also in your performance problems..
First o all as you said your normal check point times normally 1- 2 secs.
but when there is a load on the system you're having larger
checkpoints.Looking through your onconfig physical log size seems kind a
big...Maybe you are flushing it at the the time of CKPTINTVL value
reached.This may cause larger chekpoints.Also it is not deatached from
rootdbs you might be suffering from disk i/o standpoint.You may lover your
interval time or smaller physical log file might give you less checkpoitn
times.Depending on your config also LRU's seems enough but you might need
more CLEANERS..
Like if you make a search on this list you'll see LRU's = CLEANERS formula
for some cases.
Do you really need that many locks on the procedure you're running...Maybe
you can slide some of resources allocated for LOCKS to BUFFERS which will
have you do more proccesing with your box.Like lovering LOCKS adding more
BUFFERS.NUMCPUVP equals NUM PHY PROC - 1 formula always worked better for me
for systems upto 4 CPU.And KAIO enabled.Have you tried that.?or is ti
already enabled..?
On update statistics ,I strongly recommend Art's dostats esql/c script to be
implemented on your system.
you can either make a search on list or go directly www.iiug.org/software
find utils_ak...
What is your NETTYPE configuration..
Did you remove that on pupose from your onconfig ...?Couldn't see it..:(
Hope these will help ....
Alkin TEZUYSAL
INFORMIX DBA
yener@my-deja.com wrote:
> In article <8itefh$15d$1@news.xmission.com>,
> "Martin, Wayne E." <WMartin@kmart.com> wrote:
> >
> >
> > Ps Always do update statistics on major row changes!!!
> We do update statistics for the whole system every week. It lasts 9
> hours to update whole database using medium.
> >
> > Min low, optimal high
> >
> > Wayne
> > -----Original Message-----
> > From: Martin, Wayne E. [mailto:WMartin@kmart.com]
> > Sent: Thursday, June 22, 2000 9:26 AM
> > To: 'yener@my-deja.com'
> > Cc: informix-list@iiug.org
> > Subject: RE: Performance Issues
> >
> > Increase the check point time interval; if there is a good UPS on the
> > system.
> Yes, there is one. What is your suggestion?
> >
> > Increase page cleaners, put a value into the page read ahead
> > threshold, how long are your check points and what is the out
> > put from onstat -F any foreground writes?
>
> During normal activity checpoint times are 0 or 1 seconds. However,
> when there is a batch processing such as inserting transactions to the
> system from a file, it differs from 20 to 40 seconds.
>
> onstat -F doesn't cases any foreground writes.> >
> > This batch process does what, only inserts, updates, deletes etc.
>
> You are right. It does just inserting, updating and deletes.
>
> Each day over 100000 records inserted to the system.
>
> Yener.
>
> >
> > Wayne E. Martin
> > Informix Database Administrator
> > Kmart Corp.
> >
> > -----Original Message-----
> > From: yener@my-deja.com [mailto:yener@my-deja.com]
> > Sent: Wednesday, June 21, 2000 3:01 PM
> > To: informix-list@iiug.org
> > Subject: Performance Issues
> >
> > Hi,
> >
> > We are currently running informix on E5000 system which 4 CPUs and 2
> MB
> > of RAM. We have got some performance problems on this system.
> >
> > The performance of the application running on this system is not
> > comparable what we had expected. The system is a batch processing
> > system.
> >
> > We analys the system and unable to find any suitable solution to
> > improve performance.
> >
> > The listing below shows the output of onstat -p.
> > ================================================
> >
> > dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached
> > 4043750739 59853281 604861800 0.00 122732622 70823866 1106184833
> > 88.90
> >
> > isamtot open start read write rewrite delete commit
> > rollbk
> > 1018217434 2365010653 400986740 1380105727 152884987 127463044 8803283
> > 39620880
> > 366903
> >
> > 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 2000759.78 364288.72 8828 18200
> >
> > bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress
> seqscans
> > 179603434 0 3227542476 0 0 18465 9537417
> > 11416813
> >
> > ixda-RA idx-RA da-RA RA-pgsused lchwaits
> > 1492613204 3357525 1959151075 3444017014 21784573
> >
> > and the following listing shows our onconfig file
> > ==================================================
> >
> >
> #***********************************************************************
> > ***
> > #
> > # INFORMIX SOFTWARE, INC.
> > #
> > # Title: onconfig.std
> > # Description: Informix Dynamic Server Configuration Parameters
> > #
> >
> #***********************************************************************
> > ***
> >
> > # Root Dbspace Configuration
> >
> > ROOTNAME rootdbs # Root dbspace name
> > ROOTPATH /export/home/ifmxdata/s01_01> > # Path for device containing root
> > dbspace
> > ROOTOFFSET 0 # Offset of root dbspace into device
> > (Kbytes)
> > ROOTSIZE 2000000 # Size of root dbspace (Kbytes)> >
> > # Disk Mirroring Configuration Parameters
> >
> > MIRROR 0 # Mirroring flag (Yes = 1, No = 0)
> > MIRRORPATH # Path for device containing mirrored> > root
> > MIRROROFFSET 0 # Offset into mirrored device (Kbytes)> >
> > # Physical Log Configuration
> >
> > PHYSDBS rootdbs # Location (dbspace) of physical log
> > PHYSFILE 1000000 # Physical log file size (Kbytes)> >
> > # Logical Log Configuration
> >
> > LOGFILES 600 # Number of logical log files
> > LOGSIZE 20000 # Logical log size (Kbytes)> >
> > # Diagnostics
> >
> > MSGPATH /export/home/informix/online.log # System message log
> > file path
> > CONSOLE /dev/console # System console message path
> > ALARMPROGRAM /export/home/informix/etc/log_full.sh # Alarm program> > path
> > SYSALARMPROGRAM /export/home/informix/etc/evidence.sh
> > # System Alarm program path
> > TBLSPACE_STATS 1> >
> > # System Archive Tape Device
> > TAPEDEV /export/home/ifmxdata/tape0 # Tape device path
> > TAPEBLK 16 # Tape block size (Kbytes)
> > TAPESIZE 90000000 # Maximum amount of data to put on
> > tape (Kbytes)> >
> > # Log Archive Tape Device
> >
> > LTAPEDEV /export/home/ifmxdata/tape1 # Log tape device
> path
> > LTAPEBLK 16 # Log tape block size (Kbytes)
> > LTAPESIZE 40000000 # Max amount of data to put on l