The bottleneck is hdisk0 which is swap, no datafiles or anything else there.
The rest of the IO is spread out pretty nicely, archiver interferes with
redo at times but nothing I can do about it until I get some disk moved from
dev to prod. shared pool is 100M, from experience this is the right size
for J.E. Edwards Oneworld, free space runs low and no misses or request
failures. Buffer pool is 2 GB with 500 MB in a keep pool, I have some large
freqently accessed tables that will be going there (address book, chart of
accounts and perhaps a few others.) A small amount is configured for recycle
and the rest standard buffer cache. It is likely more than I need a the
moment but the box has 4 GB and Oracle is the only large app on it. There
should be plenty of memory free. I am trying to hammer the system now and
can not produced the poor results I had last night. I have seen the same
problem on my dev box when user run large address book reports. In that
case the app is on the same box as the db. If it continues I will get the
admin to put swap across more disks. I think the increase may have helped
though I don't really understand how. It seems this is one of those things
that is really a black art. Check aix newsgroups, docs, here and there and
you can find a lot of people who experience the same things but not a lot of
clear anwers. I know the contention is swap and hdisk0 I just can't get
anyone to tell me why the heck we even need to go to swap so quickly and why
is swap show 60% free? I am having my admins put a call into IBM since they
can not answer these questions and I feel they should be able to.
- Ethan
-----Original Message-----
Sent: Thursday, December 20, 2001 1:25 PM
To: Multiple recipients of list ORACLE-L
One thing that jumps right out is 50% IO waits.
You need to find out which disk(s) are so overloaded.
Analyze the stats from v$filestat to see what's going
on at the database level.
Q: Why is your SGA 2 gig, and how is it partitioned?
db_block_buffers:
shared_pool:
other pools:
Jared
"Post, Ethan"
<Ethan.Post@ps To: Multiple recipients of list
ORACLE-L <[EMAIL PROTECTED]>
.net> cc:
Sent by: Subject: Excessive Paging - AIX
Related?
[EMAIL PROTECTED]
om
12/20/01 10:35
AM
Please respond
to ORACLE-L
Sorry for the cross post if you are the other big list.
I am having some trouble finding a solution to a problem. This is
specifically AIX related. Here is the gist.
* Any large sort or import appears to cause paging to take place. I only
say large because of the problems, in reality these are quite small objects
for the machine we are running, 200,000 - 600,000 rows in the 200-400 MB
range. Any way last night I was importing one of these objects.
* CPU wio goes through the roof almost right away. Spending up to 50% of
time waiting on IO.
* vmstat shows huge numbers in po, fr, sr and a few block processes in b
column.
* hdisk0 goes 100% (swap space is here)
* lsps -a shows swap space is only 40% utilized, never changes. There is
only 2GB of swap, admin just increase to 4 GB but still on same disk. Not
sure this will really help since it never was 100% anyway.
* Setup- RS/6000 7026-M80 rack mount running AIX with 4-500Mhz processors,
4GB memory, 3-10/100 Ethernet cards, 2 fiber ports, and 2-18.2GB internal
drives with RAID. The 2 fiber ports will attach to a storage area network
providing access to EMC Clarion storage (see below)
* Oracle is the only app on the box and I have configured about 2GB for the
SGA. There should be plenty of free memory.
I am pretty sure we hosed something up with are memory configuration. I
see
a lot of things about using vmtune to configure various AIC related
parameters as too how much memory is can use for various things. Can
anyone
provide me with the direction to head in examining our configuration and
what settings should/could be changed to improve the situation?
Thanks,
Ethan Post
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Post, Ethan
INET: [EMAIL PROTECTED]
Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051
San Diego, California -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author:
INET: [EMAIL PROTECTED]
Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051
San Diego, California -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Post, Ethan
INET: [EMAIL PROTECTED]
Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051
San Diego, California -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
