Re: How fast should an IO be on a SYMMETRIX?

From: Sal L <member46676_at_dbforums.com>
Date: Sun, 02 Nov 2003 14:31:42 -0500
Message-ID: <3550820.1067801502@dbforums.com>

Hello Jonathan,

this is more complete citiation from VC post:

"On my home PC with a single IDE disk, I got these numbers:

 create table t1 as select * from all_objects;

 insert /*+ append */ into t1 select * from t1;

.................

commit;

select count(*) from t1;

count(*)

---

    811040

create index t1_idx on t1(object_name) nologging;

alter session set events '10046 trace name context forever, level 8';

select count(*) from t1 where object_name between 'B' and 'T';

select count(*)

from

 t1 where object_name between 'B' and 'T'

call count cpu elapsed disk query current

rows

• ------ -------- ---------- ---------- ---------- ----------

  ---

  ---

Parse 1 0.00 0.00 0 0 0

Execute 1 0.00 0.00 0 0 0

Fetch 2 0.27 0.54 375 376 0

1

• ------ -------- ---------- ---------- ---------- ----------

  ---

  ---

total 4 0.27 0.54 375 376 0

1

• ---------------------------------------------------

      1 SORT AGGREGATE (cr=376 r=375 w=0 time=548486 us)

  94016 INDEX RANGE SCAN OBJ#(48017) (cr=376 r=375 w=0 time=444504

us)(object id 48017)

Elapsed times include waiting on following events:

  Event waited on                             Times   Max. Wait  Total

Waited

---

 Waited ---------- ------------

  SQL*Net message to client                       2        0.00

00

  db file sequential read                       375        0.03

29

  SQL*Net message from client                     2        5.31

5.31

******************************************************************-

---

---

Average wait time for a single 8K block was about 1 ms (0.29/375 ) . I do

realize it's not quite apples-to-apples but surely a Symmetrics could do

better than it did in your case."

You can see there is no multi-block read, thus 1 ms timing can be explained

by using of file system cache. But what is wrong with Symmetrix cache?

And could you comment the statement regarding tkprof disk statistics taken

from Knowledge Xpert for Oracle Administration:

"disk Number of data blocks read from disk for each parse, fetch,

or execute. If a single multiblock read returns eight blocks,

this figure is incremented only once for the physical read."

According this one you shouldn't divide 375/8 to get read request count.

--

SL


--
Posted via http://dbforums.com

Received on Sun Nov 02 2003 - 13:31:42 CST