Re: CISC to RISC and Excessive Logical I/O

["Alexei Pytel" <pytel@xxxxxxxxxx>]

Logical I/O happens when HLL (or HLL runtime) requests next block of
records from a file.
If logical I/O jumps I would suspect that for some reason blocking was
suppressed - or blocking factor was reduced.
Watch for the message in joblogs. I do not remember message ID, but it
reads something like "FIle open changed to SEQONLY(*NO) for such and such
reason".

    Alexei Pytel

"The better is the worst enemy of the good"




                                                                                
                   
                    "Law, Tom"                                                  
                   
                    <actlaw@NMHG.com>         To:     MIDRANGE-L 
<MIDRANGE-L@midrange.com>         
                    Sent by:                  cc:                               
                   
                    owner-midrange-l@mi       Subject:     CISC to RISC and 
Excessive Logical I/O  
                    drange.com                                                  
                   
                                                                                
                   
                                                                                
                   
                    06/20/2001 08:08 AM                                         
                   
                    Please respond to                                           
                   
                    MIDRANGE-L                                                  
                   
                                                                                
                   
                                                                                
                   




We are doing a CISC to RISC conversion from version 3.2 to 4.5.  We are
having some slow response times in some transactions.  We were finding that
DB CPU was getting upwards of 100%.  When looking at the traces (PEX,
performance, smtrace, DBmonitor), we found that simple transactions were
getting upwards of 100,000 to over a 1,000,000 logicals I/O's with less
than
320 SYNC reads.  The same functions on the CISC systems were usually less
than a 1,000 and no problems.

Data was restored to the RISC systems using the backup created from the
UPGRADE2 function on the CISC box.  I used the RSTLIB command for the data
libraries.  The data looks correct after the restore.  The only thing I can
think of that is common among different transactions is that the logicals
are not correct.  Could the logicals be the culprit?

We have had IBM look at the problem, but so far it can't be explained.

Any info would be greatly appreciated.

Thanks

Thomas A. Law Jr.

Sr. Systems Programmer

NACCO Materials Handling Group, Inc.
Americas MIS - Danville
1813 E. Voorhees St.
Danville, IL. 61832
217-443-7622
217-443-7657 Fax
217-444-9936 Pager
E-mail: actlaw@nmhg.com

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to
MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator:
david@midrange.com
+---




+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---