Sponsored Content
Operating Systems AIX Network performance problem between AIX and SUN ethernet gigabit Post 47727 by chuckw on Tuesday 17th of February 2004 02:23:47 PM
Old 02-17-2004
Tools Network performance problem between AIX and SUN ethernet gigabit

I am posting this as an FYI. I burned a lot of time on digging for this solution. The problem was performance between newer AIX p series boxes and a sun 4500 with a fiber nic. I tried many misc. solutions but this turned out to fix the issue. My bandwidth benchmark went from around 20MB/sec to 60MB/sec. after the change. (the other symptom I had was on the my cisco catalyst 6500 series switch the RxPause grew to 272448910)

I made the change from the console without rebooting.

Hope this helps someone! -Chuck

IBM guys comments:

There have been some reported problems with FC 5700, 5701, 5706 and 5707 which are the newer Gigabit Ethernet adapters we support in the pSeries systems. The devices have two TCP offload features, large_send (TCP Resegmentation) and chksum_offload (Checksum Offload), that are intended
to be performance features. In some instances the devices work as designed, but there have been issues with some software combinations where the offload features have caused problems. It is not difficult to disable the two features, but you want to make sure you have them either on or off for all devices that used the feature. To disable the large_send and chksum_offload you will detach the IP interfaces, put the device into a defined state, change the device configuration then reinitialized the
device and its interfaces. Sometimes it is easier to reboot the server as their may be application issues when IP interfaces are temporarily removed. The example shows how to change one device, ent1, and if you have other devices of the same type you will want to modify them as well.

1. If you can reboot:

chdev -l ent1 -a large_send=no -a chksum_offload=no -P
shutdown -Fr

2. If you can't reboot, but can take a very short communications outage
Note: if you get a device busy error when running the rmdev command then
try the ifconfig sequences again. If
you continue to get device busy error then use proceedure 1.

#!/bin/ksh
ifconfig en1 down detach
ifconfig et1 down detach
rmdev -l ent1
chdev -l ent1 -a large_send=no -a chksum_offload=no
# Checks to make sure device driver is unloaded
genkex | grep /usr/lib/drivers/pci/goentdd
# If above command returned no output then the driver is unloaded. If not,
use proceedure 1.
mkdev -l ent1
mkdev -l en1
mkdev -l inet0
exit
 

9 More Discussions You Might Find Interesting

1. UNIX for Advanced & Expert Users

AIX server performance problem!

Hello, I have a performance problem on an AIX box. I'm not sure what is causing this and hoping someone may have suggestions. Currently I'm noticing that cpu's are waiting while the box is in a idle state. I checked the disks and none of them are at 100%. If they were then I would understand... (3 Replies)
Discussion started by: ctcuser
3 Replies

2. Linux

SUSE9.1 Peculiar Gigabit Lan performance.

I'm having some peculiar performance issues with my Gigabit Lan. I have some 100Mb devices so I can't do the necessary "jumbo Frame" tweaks for absolute optimum performance as I'd prevent them access. I'm getting appauling transfer rates sending files to the linux machine, around 10 Mbps 3%... (0 Replies)
Discussion started by: Mark Ward
0 Replies

3. SCO

Configuring Broadcom BCM5721 NetXtreme Gigabit Ethernet

Hi, I have my server which has LAN with above model build onboard my Dell Server, I got this driver from SCO Unix website but yet to identify the following option: 1. PCI Slot- 2. Bus- 3. Device- 4. Function- Would anyone please kindly advice what should the... (0 Replies)
Discussion started by: spiderweb
0 Replies

4. UNIX for Dummies Questions & Answers

Network performance problem

I have a Teradata Machine, using MP-RAS Unix, with a 1000 Intel Ethernet card and a Cisco switch. If I configure the ethernet card and the switch to auto, so they negotiate to 1000, or configure the ethernet card and switch manually to 1000Full or 100Full, the velocity is very very low. Only... (2 Replies)
Discussion started by: cuatrodos
2 Replies

5. Programming

send mail performance problem in AIX

Hi All, Our application is designed to send the emails. we use standard C lib to send the mails using send() function. This works fine on Solaris,HP UX and windows, but on AIX the performance is very slow. it takes around 32 sec to send a simple mail whereas in Solaris 2-3 sec. I would like to... (2 Replies)
Discussion started by: shrihari_s
2 Replies

6. Solaris

SUN GIGABIT ETHERNET CARD

Hi, we have SUN Gigabit ethernet cards at the server, both have 2 lines connected but hiw can I found out which cable is on which port before I configure it: the cards has 4 ports , from right to left the cables are connected at the first right then 2 empty then the last at the left side. ... (1 Reply)
Discussion started by: manni2
1 Replies

7. Hardware

Need Sun 10-Gigabit Ethernet Adapter CD or ixge driver

Hi, Does anybody have a "Sun 10-Gigabit Ethernet Adapter installation CD" or driver ixge for Sun 10-Gigabit Ethernet PCI-X adapter? I need it but can't find it anywhere. So if you know where I can download it or purchase it, please let me know. (Already tried Oracle's website, don't have it) ... (5 Replies)
Discussion started by: Lhao
5 Replies

8. AIX

AIX VIOS IVM Ethernet problem

Hi All, I have a p520 box running 3 LPARs, 1 x VIOS, 1 x AIX 6.1 and 1 x AIX 7.1 and no HMC, all configuration is performed via VIOS/IVM Until recently the box did not have the AIX7.1 LPAR and everything was fine with the box. We needed to use AIX 7.1 for a new project, so I created a... (11 Replies)
Discussion started by: KevB
11 Replies

9. AIX

Newbie - AIX LPAR performance problem

Running into performance issues with WAS application servers on two of LPAR's or like configuration under high load web pages crawl. Please forgive me I'm new to AIX and most my expertise is in the Linux space. Thanks for your help!! Here's the run down: The problem appears to be CPU... (3 Replies)
Discussion started by: margeson
3 Replies
SMC91C92_CS(4)						     Kernel Interfaces Manual						    SMC91C92_CS(4)

NAME
smc91c92_cs - SMC 91cxx device driver SYNOPSIS
insmod smc91c92_cs.o [pc_debug=n] [if_port=n] [irq_list=i,j,...] DESCRIPTION
Smc91c92_cs is the low-level Card Services driver for SMC 91c92, 91c94, 91c96, and 91c100 based PC Card ethernet adapters. When this driver is attached to a card, it allocates the next available ethernet device (eth0..eth#). This device name will be reported in the ker- nel log file, and passed on to cardmgr(8). With the default transceiver setting, the driver will attempt to continuously autodetect the transceiver type (10base2 or 10baseT). There should normally only be a one or two second lag before the correct transceiver is selected. On particularly noisy or busy networks, it is possible for the detection logic to incorrectly toggle the transceiver type. If this becomes a problem, it may be necessary to explicitly select the interface type when the module is loaded or with the ifport command. PARAMETERS
pc_debug=n Selects the PCMCIA debugging level. This parameter is only available if the module is compiled with debugging enabled. A non-zero value enables debugging. if_port=n Selects the transceiver type. 0 is autodetect (10baseT or 10base2), 1 is 10baseT (twisted pair), and 2 is 10base2 (BNC, or thin net). The default is 0 (autodetect). irq_list=i,j,... Specifies the set of interrupts that may be allocated by this driver. DIAGNOSTICS
eth#: interrupt(s) dropped! Indicates that the driver did not receive an interrupt notification for some reason. The driver will poll the card (with a signifi- cant performance penalty) if the problem persists. The most likely cause is an interrupt conflict and/or host bridge configuration problem. DIAGNOSTICS
eth#: interrupt(s) dropped! Indicates that the driver did not receive an interrupt notification for some reason. The driver will poll the card (with a signifi- cant performance penalty) if the problem persists. The most likely cause is an interrupt conflict and/or host bridge configuration problem. AUTHOR
David Hinds - dahinds@users.sourceforge.net SEE ALSO
cardmgr(8), pcmcia(5), ifport(8). pcmcia-cs 2000/06/12 21:24:48 SMC91C92_CS(4)
All times are GMT -4. The time now is 06:45 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy