Cynthia,

We fought the same issue.  We had servers whose network connection was "100 Mbs"
all the way to the ADSM server backing up at speeds that looked that there was a
modem in the middle of the chain.  That chain included a FDDI loop.

We eventually traced it to a combination of:
1) auto speed detections vs. manually setting the network speed;
2) half-duplex vs. full-duplex at EVERY network interface;
3) protocols installed on the source server.

Eventually we got all the equipment in the chain reading from the same page and
throughput between that server and our ADSM server went up 50-fold.

Tab







"CANNAM, CYNTHIA L (SBCSI)" <CC5657 AT MOMAIL.SBC DOT COM> on 02/16/2000 
03:40:38 PM

Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

To:   ADSM-L AT VM.MARIST DOT EDU
cc:    (bcc: Tab Trepagnier/Corporation/Laitram/US)
Subject:  larger ADSM clients




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