ADSM-L

Re: larger ADSM clients

2015-10-04 17:33:43
Subject: Re: larger ADSM clients
From: Joel Fuhrman [mailto:joelf AT CAC.WASHINGTON DOT EDU]
To: ADSM-L AT VM.MARIST DOT EDU
Here are the non-default settings which I use on AIX:
        /usr/sbin/no -o rfc1323=1
        /usr/sbin/no -o sb_max=524288
        /usr/sbin/no -o tcp_mssdflt=4096
        /usr/sbin/no -o tcp_sendspace=131072
        /usr/sbin/no -o tcp_recvspace=131072
        /usr/sbin/no -o thewall=131072
I'm not at work so I can't look up the name of the Redbook which covers
this subject.  The title is something like Configuring ADSM for Performance.
If I remember, the rfc1323 and sb_max have the biggest impact on fddi
operations.


On Wed, 16 Feb 2000, CANNAM, CYNTHIA L (SBCSI) wrote:

> To all:
>
> I have two very large (nearing 1 TB each) HP boxes that are really giving
> the network fits these days. Although there are other equally large HP and
> other platform clients in our environment and spread throughout the
> geographical area of support, these two seem to be more trouble than all
the
> rest combined. I think it may be that these two, unlike all the others,
are
> attached to a FDDI backbone where the others are all 100 Mb Ethernet. I've
> tried tuning up and down the scale, and even though the machines do
> eventually back up (they each average about 5-8 GBs per evening), the
> throughput is AWFUL (sometimes about 125 MB/hour)!
>
> Does anyone have any opinions on this theory? Thanks!
>
> Cindy Cannam
>
> Cynthia L. (Cindy) Cannam, MBA
> Midrange Planning/TSM Technical Support
> Systems Manager/SBC Services, Inc.
> (314)340-0722 office
> (314)879-1696 pager
> (314)340-0992 fax
> Nearly Native Floridian Freezing in the Midwest
>
<Prev in Thread] Current Thread [Next in Thread>