|
Did you try changing the max frame size?
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Friday, February 08, 2013 1:47 PM
To: midrange-l@xxxxxxxxxxxx
Subject: FTP Time trials
I wanted to do some time trials, before and after changing key FTP and/or TCP settings.
I used a guest lpar, GDI, and it's host, MAIL3. The guest lpar was not optimally set up. GDI only runs on two storage spaces (with a wide disparity in size between the two). I picked these because I could bring down TCP/IP without incurring the users wrath.
GDI had the least amount of free space, just 88GB. Neither system was too busy.
GDI
Sys Pool Reserved Max ----DB----- --Non-DB---
Pool Size M Size M Act Fault Pages Fault Pages
1 729.73 248.09 +++++ .0 .0 1.4 1.7
2 5935.38 51.72 203 .0 .0 3.7 8.1
3 162.37 .00 41 .0 .0 .0 .0
4 81.18 .00 5 .0 .0 .0 .0
5 1210.31 13.25 21 .0 .0 .0 .0
Size % I/O Request Read Write Read Write %
Unit Type (M) Used Rqs Size (K) Rqs Rqs (K) (K) Busy
1 6B22 233020 64.6 95.2 4.5 .0 95.2 .0 4.5 1
2 6B22 93212 93.6 5.3 4.5 1.3 3.9 4.5 4.5 1
MAIL3
System Pool Reserved Max -----DB----- ---Non-DB---
Pool Size (M) Size (M) Active Fault Pages Fault Pages
1 836.98 428.39 +++++ .0 .0 1.8 2.5
2 14079.37 40.51 1055 .1 4.7 55.4 273.9
3 1657.39 .00 9 .0 .0 2.7 5.4
4 .25 .00 5 .0 .0 .0 .0
Size % I/O Request Read Write Read Write %
Unit Type (M) Used Rqs Size (K) Rqs Rqs (K) (K) Busy
1 433C 139594 59.3 3.8 21.4 1.0 2.8 62.0 6.7 3
1 433C 139594 59.3 4.6 10.6 2.4 2.1 13.4 7.5 2
2 4328 123488 59.3 2.9 29.3 1.0 1.9 67.5 9.0 0
3 4328 141129 59.3 5.5 29.5 2.6 2.8 39.2 20.3 1
4 4328 123488 59.3 1.6 46.4 1.0 .6 48.5 43.2 0
5 4328 123488 59.3 3.9 21.4 1.6 2.3 40.6 7.5 0
6 4328 123488 59.3 4.7 32.5 2.3 2.4 44.0 21.6 1
7 4328 123488 59.3 1.7 25.4 .7 1.0 47.3 9.0 0
8 4328 123488 59.3 3.4 28.5 1.1 2.3 44.0 20.8 0
9 4328 123488 59.3 5.3 38.8 2.6 2.6 43.6 34.0 1
10 4328 123488 59.3 4.1 26.8 1.9 2.1 48.8 7.5 1
11 4328 141129 59.3 3.7 24.2 1.6 2.0 43.3 8.7 0
12 4328 141129 59.3 8.2 15.8 6.1 2.0 14.2 20.7 2
13 4328 141129 59.3 11.0 15.1 7.8 3.2 13.9 18.0 1
14 4328 141129 59.3 7.0 17.8 5.1 1.9 15.1 25.0 0
15 4328 141129 59.3 10.7 14.1 6.9 3.8 14.5 13.3 2
16 4328 141129 59.3 1.9 13.0 1.2 .6 7.6 24.0 0
17 4328 141129 59.3 4.3 25.8 1.6 2.6 21.5 28.5 0
18 4327 52923 59.3 1.5 10.0 .1 1.4 12.0 9.8 0
19 4327 52923 59.3 1.9 6.9 .7 1.1 4.0 8.8 0
20 4327 52923 59.3 1.7 8.2 1.2 .5 4.0 19.0 0
21 4327 70564 59.3 3.9 29.8 1.9 2.0 29.3 30.2 0
22 4327 70564 59.3 1.5 25.3 .2 1.2 128.0 4.8 0
23 4327 52923 59.3 1.0 14.5 .7 .2 18.0 4.0 0
To create a large chuck of data to FTP I performed the following:
I saved a DB2 PF of this size
Size . . . . . . . . . . . . . . . . : 13186801664
Offline size . . . . . . . . . . . . : 13186760704
into a save file. The resultant size of the save file was 25580260992.
(Big huge difference!)
Knowing that FTPing save files creates a high level of inefficiency I used CPY to copy it to a stream file in a stream file directory just off the root of the IFS. Size of the file is 25580260992.
GDI has an IP address of 10.17.6.xxx and MAIL3 has an IP address of 10.17.7.xxx. Basically they are not on the same subnet and must pass through a switch.
This is a match on both lpars: CHGTCPA TCPRCVBUF(8192) TCPSNDBUF(8192) And this matches on both lpars:
DSPLIND
Line speed . . . . . . . . . . . . : *AUTO
Current line speed . . . . . . . . : 1G
Duplex . . . . . . . . . . . . . . : *AUTO
Current duplex . . . . . . . . . . : *FULL
Serviceability options . . . . . . : *NONE
Maximum frame size . . . . . . . . : 1496
Did a transfer from GDI to MAIL3:
25580260992 bytes transferred in 1374.158 seconds. Transfer rate 18615.223 KB/sec.
Deleted in on GDI and sent it back:
25580260992 bytes transferred in 2403.694 seconds. Transfer rate 10642.061 KB/sec.
Deleted it on MAIL3 and sent it back:
25580260992 bytes transferred in 1342.492 seconds. Transfer rate 19054.309 KB/sec.
Deleted it on GDI and sent it back:
25580260992 bytes transferred in 2086.183 seconds. Transfer rate 12261.756 KB/sec.
So far the guest lpar looks better at sending then receiving. Could be a WRKDSKSTS thing.
Brought systems down into restricted state. Ran CHGTCPA TCPRCVBUF(262144) TCPSNDBUF(262144).
IPLed. Did a few things to let system settle.
Did a transfer from GDI to MAIL3:
25580260992 bytes transferred in 1877.477 seconds. Transfer rate 13624.810 KB/sec.
Deleted in on GDI and tried to send it back but after about 10G I got an ABOR.
Deleted what was there on GDI and tried again:
25580260992 bytes transferred in 2165.111 seconds. Transfer rate 11814.756 KB/sec.
Deleted it on MAIL3 and sent it back:
25580260992 bytes transferred in 1715.588 seconds. Transfer rate 14910.493 KB/sec.
Deleted it on GDI and tried to send it back. After about 10G I got:
TCP12B9 Unable to send data to server.
Deleted it on GDI again and tried sending it again:
25580260992 bytes transferred in 2158.348 seconds. Transfer rate 11851.779 KB/sec,
For some reason that doesn't seem like an improvement.
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a moment to review the archives at http://archive.midrange.com/midrange-l.
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.
midrange.com
