Just in case Steffan doesn't answer soon 8-)

If you have 2 15 MB/s tape drives on Ultra SCSI-2 or whatever you call the
80MB/s standard then if both your drive achieve compression at say 3:1 then
you would be trying to push 90 MB/s down your SCSI, OK this might not be a
problem if it only happens for a short while or your backups always take
less than the available window but I have seen multiple drive doing
averages of 4:1 compression presumably with peaks higher than this and
during backups lasting several hours, this would throttle us down to 2/3rds
of the drive capabilities, putting each drive on its own SCSI port would
allow 5:1 compression not to bottleneck.
Having said that Ultra SCSI-3 goes at 160 MB/s and if you have suitably
rated cables the you would only have to worry about 2 drives per port if
you expect you compression to average higher than 5:1 ( I have seen one
100G tape with just over 600G of data on it !).

Bigasm is an "asm" that generates data rather than reading it from disk, it
is used to check limits in Legato performance, if you run it on the server
to a single drive then the drive not streaming will indicate there are SCSI
or I/O problems.
If you run it to all the drives on your server it will show is the total
I/O cababilities are adequate.
Running it on a client will indicate if the OS or network or disk subsystem
of the client is a bottleneck.

As I'm getting tired I'll just say search the archives for "bigasm" and
there should be enough examples on how to use it (unfortunately mostly for
Unix but the directives used should be similar).

PS please let the list know what sort of figures you get I know we have
managed to bottlneck on E450 network performance (much improved with
V880's) but have no clue as too W2K legato server performance.

On Thu, 17 Apr 2003 12:02:47 -0400, Michael Brooks
<michaelb AT SEQUOIAINS DOT COM> wrote:

>thank you all for your input. steffan, i will try a few of your
>suggestions, but i have a question for you: we have been told that daisy-
>chaining two drives on one adapter is ok. have you seen significant
>performance hits as a result of this? also, i am not familiar
>with "bigasm," can you provide a few details? thanx again.
>
>--
>Note: To sign off this list, send a "signoff networker" command via email
>to listserv AT listmail.temple DOT edu or visit the list's Web site at
>http://listmail.temple.edu/archives/networker.html where you can
>also view and post messages to the list.
>=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=