Hi folks,
wondered if anyone can help me understand some FICON performance issues?
I'm seeing huge DISC times (58 - 70%) for a device on LCU 6571. The manual
says that I can get an good estimate of the MB/SSCH by taking the MB/SEC
from the IOQ activity report and dividing them by the Total SSCH/SEC from
the CHANNEL PATH activity report. Unfortunately, it then states; for each
LCU that a single FICON channel is connected to.
The IOQueueing report appears to show 7 channel paths......backed up by
devserv command that shows 1 chpid as inop from total of 8.
22:45 CHPID TAKEN
F5 6571 035E 178.17
EE 6571 035E 330.33
E5 6571 035E 112.89
F2 6571 035E 313.28
ED 6571 035E 163.84
E0 6571 035E 124.17
E4 6571 035E 284.76
035E 0.0 0.00 2.7 1507.4
So I've taken the 7 channels from the Channel Path Activity report.
read(B/s) write(B/s)
part total part total
F5 2 FC_S Y 13.2 23.5 18.4 9M 14M 7M 9M
EE 2 FC_S Y 12.4 19.0 15.9 8M 10M 6M 8M
E5 2 FC_S Y 7.9 14.7 13.6 6M 9M 4M 4M
F2 2 FC_S Y 10.1 13.3 13.4 6M 8M 5M 6M
ED 2 FC_S Y 12.2 22.1 17.4 8M 13M 6M 7M
E0 2 FC_S Y 6.8 15.1 13.6 5M 9M 4M 5M
E4 2 FC_S Y 8.7 13.1 12.9 5M 7M 4M 5M
Do I now take the total of all these 7 and divide by 1507.4 or do i just
divide the individual channels (from channel path) by the individual
channels on the ioq report.
Am I right in assuming that the LCU is pretty busy?
Many thanks to anyone who can spare the time to help.
JJ
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
