Re: S85 not quite right !!

• Previous message: John Jolet: "Re: jfs2 & 5.2 any performance issues?"
• Maybe in reply to: Spencer Johnson: "S85 not quite right !!"
• Next in thread: Hunter, Mark: "Re: S85 not quite right !!"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Date:         Mon, 30 Jun 2003 11:47:21 -0400
To: aix-l@Princeton.EDU

Spencer-

Couple suggestions-

    run "pstat -A | grep aio | wc -l"
    this will show how many Async io servers you are using.
    check "smitty aio" to see if you are using the max number defined.
    you may want to increase the number of aio servers available. I've seen
S80s with Oracle need 300-1000 aio servers.

    run "vmtune -a" ; sleep 20 ; "vmtune -a"
    if fsbufcnt is increasing, then you need to add more file system buffers
by increasing the numfsbufs (-b) setting.
  (problem here is filesystems have to be unmounted and remounted for the
change to help).

-pablo

-----Original Message-----
From: Spencer Johnson [mailto:Spencer.Johnson@2020LOG.COM]
Sent: Monday, June 30, 2003 11:34 AM
To: aix-l@Princeton.EDU
Subject: Re: S85 not quite right !!

Thanks for replies Mark / Simon,

To answer some of your questions :

- Disks are 50% active on average is quite good as this means layout of
database is good.

- Our vmtune settings are currently :

vmtune: current values:
  -p -P -r -R -f -F -N -W
minperm maxperm minpgahead maxpgahead minfree maxfree pd_npages
maxrandwrt
 104855 209711 2 64 128 144 524288 0

  -M -w -k -c -b -B -u -l
-d
maxpin npswarn npskill numclust numfsbufs hd_pbuf_cnt lvm_bufcnt lrubucket
defps
1677691 100608 25152 1 93 1584 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

 -t
maxclient
1676868

number of valid memory pages = 2097113 maxperm=10.0% of real memory
maximum pinable=80.0% of real memory minperm=5.0% of real memory
number of file memory pages = 942297 numperm=44.9% of real memory

number of compressed memory pages = 0 compressed=0.0% of real memory
number of client memory pages = 0 numclient=0.0% of real memory
# of remote pgs sched-pageout = 0 maxclient=80.0% of real memory

-------- IBM have also suggest turning sync_release_ilock (s) to 1 and
strict_maxperm (h) to 1.

I will be shortly upgrading our Dev server to ML 11, will let you know how I
get on.

Spencer

-----Original Message-----
From: Green, Simon [mailto:Simon.Green@EU.ALTRIA.COM]
Sent: 30 June 2003 15:29
To: aix-l@Princeton.EDU
Subject: Re: S85 not quite right !!

The only problems I have experienced - other than an actual shortage of
resources - have been due to the wrong vmtune settings. (And remember that
the default settings are NOT good for a database.)
It's worth considering AIO as well, but I've never had a problem with that.

The sort of settings I use are:
/usr/samples/kernel/vmtune -p3 -P8 -b400

That's taken from an 8-way S7A with 8GB memory, running SAP with a 750GB
database. It's quite busy, but has almost no paging. It's based on our
experience and recommendations from SAP.

If you don't have something similar - particularly the -p/-P - then take
care of that before tackling anything else.

When you say the disks are "only 50% busy" what do you mean? If that's
either I/O Wait or % active then I'd say it's on the high side. What sort
of disk subsystem do you have?

Simon Green
Altria ITSC Europe Ltd

AIX-L Archive at http://marc.theaimsgroup.com/?l=aix-l
<http://marc.theaimsgroup.com/?l=aix-l&r=1&w=2> &r=1&w=2
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
<http://www.faqs.org/faqs/aix-faq/>

N.B. Unsolicited email from vendors will not be appreciated.

-----Original Message-----
From: Spencer Johnson [mailto:Spencer.Johnson@2020LOG.COM]
Sent: 30 June 2003 14:49
To: aix-l@Princeton.EDU
Subject: S85 not quite right !!

Hi list,

I have a slight problem / issue, this is related to an 7017-S85 server,
connected to F20 ESS. The server is a 6W - 8 GB RAM running Oracle 8.1.7.4
with
database instance of approx. 1 TB.

The problem is that most operations (coping files, etc) cause the system to
page and send the CPU's into Wait state constantly, I have logged a call
with IBM and they have
come back with a few suggestions of changing vmtune parameters.

Oracle is tuned to allow for 6W processors but I feel that the O/S is being
restricted by something. The disks do not perform that much work (only 50%
busy at max).

The O/S is currently at 4.3.3 ML 10, but am considering going to ML 11.

Has anyone out there had similar experiences and how have they been
overcome.

***************************************************************
Disclaimer :-
This email is confidential and intended solely for the use of the
individual to whom it is addressed. Any views or opinions are solely
those of the author and do not necessarily represent those of
20:20 Logistics Ltd, Dextra Accessories Ltd,
The Discovery Store Ltd and The Mobile Phone Repair Company Ltd.
If you are not the intended recipient, be advised that you have
received this email in error and that any use, dissemination,
forwarding, printing or copying is strictly prohibited. If you
have received this email in error please delete it from your system
and notify the sender immediately.
***************************************************************

Blue Cross Blue Shield of Florida, Inc., and its subsidiary and affiliate
companies are not responsible for errors or omissions in this e-mail message.
Any personal comments made in this e-mail do not reflect the views of Blue
Cross Blue Shield of Florida, Inc. The information contained in this document
may be confidential and intended solely for the use of the individual or
entity to whom it is addressed. This document may contain material that is
privileged or protected from disclosure under applicable law. If you are not
the intended recipient or the individual responsible for delivering to the
intended recipient, please (1) be advised that any use, dissemination,
forwarding, or copying of this document IS STRICTLY PROHIBITED; and (2) notify
sender immediately by telephone and destroy the document. THANK YOU.

• Previous message: John Jolet: "Re: jfs2 & 5.2 any performance issues?"
• Maybe in reply to: Spencer Johnson: "S85 not quite right !!"
• Next in thread: Hunter, Mark: "Re: S85 not quite right !!"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]