Just to compare we backup approx. 2.5GB a night from 2 different nodes that
have 1,000,000,000+ files each on them. Each node runs to completion in less
than 4 hours. Not great but better that what you are seeing. Our experience
is the client size/speed does play a role in the overall throughput. The
same data was recently migrated from two Sun machines. Although they were
faster processors they had slower I/O. The backup of the same data was
running around 8 hours per night.

Server is a RS/6000 J40 (4.2.1 / 3.1.1.3)
Clients are RS/6000 990's (4.2.1 / 3.1.05 & 3.1.06)

Curt Magura
Lockheed Martin
Enterprise Information Systems
Gaithersburg Md.
301-240-6305


                -----Original Message-----
                From:   Sean Stecker [mailto:Stecker.Sean AT ORBITAL-LSG DOT 
COM]
                Sent:   Monday, March 08, 1999 1:31 PM
                To:     ADSM-L AT VM.MARIST DOT EDU
                Subject:        Slow backup due to large amount of files???

                All,

                We have a node that has approximately 600,000 to 700,000
files on it and are
                receiving extremely poor backup performance.  Is there
anyone out there who has
                experienced this or is experiencing this now?  If it is past
tense,  any ideas
                on how to resolve this?  We do not see this problem on any
other client in the
                building.  I would be interested to hear other people's
performance vs. our own.
                On this one node, we had backed up just over 1 Gb in 24
hours.

                Any responses would be greatly appreciated.

                Thanks alot.

                Sean Stecker
                stecker.sean AT orbital-lsg DOT com