This is a multipart message in MIME format.
--=_alternative 00019158852569F3_=
Content-Type: text/plain; charset="us-ascii"
Kim, How many clients do you have hitting the server simultaneously? I
recently tested Sun's GigEther card on two 220R's (each has one GE card)
running Solaris-7. The best I could get was 400Mb/s between the two. I
talked with our networking folks here and they said that that's to be
expected given the gig protocol as far as the sending of data goes;
recieving should be different. We were told that if multiple clients were
to send data at the same time, we should see the full gig speed as far as
recieving goes. Unfortunately, when I did the test, I did not have enough
clients to fully stress the gig link.
-- Tony Guzzi
Solutions Engineer, AssuredRestore team
Storability, Inc.
> Message: 8
> From: "Pearson, Kim (STP)" <kim.pearson AT guidant DOT com>
> To: "'Veritas-bu AT mailman.eng.auburn DOT edu'"
> <Veritas-bu AT mailman.eng.auburn DOT edu>
> Date: Tue, 13 Feb 2001 16:16:08 -0600
> charset="iso-8859-1"
> Subject: [Veritas-bu] Sun Gigabit Performance
>
> What is the best way to benchmark network throughput? We are using 2
Gigabit
> ethernet ports on our NetBackup Master/Media server, and we are not
seeing
> the throughput we are expecting for backups. Since we stream multiple
> classes to each drive, I'm inclined to believe that it's the network,
not
> the 8 DLT drives. They sit idle a lot...shoe shining? We are using
crude
> copy scripts now to try to measure throughput - anyone got something
better?
>
> HW - Sun E450, 2 Sun Gb PCI cards dedicated to the backup segment...
>
> Thanks!
>
> Kim
> Kim Pearson
> UNIX Team Lead
> Guidant Corporation
> kim.pearson AT guidant DOT com
--=_alternative 00019158852569F3_=
Content-Type: text/html; charset="us-ascii"
<br><font size=2 face="Courier New">Kim, How many clients do you have
hitting the server simultaneously? I recently tested Sun's GigEther card
on two 220R's (each has one GE card) running Solaris-7. The best I could
get was 400Mb/s between the two. I talked with our networking folks here
and they said that that's to be expected given the gig protocol as far as the
sending of data goes; recieving should be different. We were told that if
multiple clients were to send data at the same time, we should see the full gig
speed as far as recieving goes. Unfortunately, when I did the test, I did not
have enough clients to fully stress the gig link.</font>
<br>
<br><font size=2 face="Courier New">-- Tony Guzzi</font>
<br><font size=2 face="Courier New">Solutions Engineer, AssuredRestore
team</font>
<br><font size=2 face="Courier New">Storability, Inc.</font>
<br>
<br><font size=2 face="Courier New">> Message: 8<br>
> From: "Pearson, Kim (STP)" <kim.pearson AT guidant DOT
com><br>
> To: "'Veritas-bu AT mailman.eng.auburn DOT edu'"<br>
> <Veritas-bu AT mailman.eng.auburn DOT edu><br>
> Date: Tue, 13 Feb 2001 16:16:08 -0600<br>
> charset="iso-8859-1"<br>
> Subject: [Veritas-bu] Sun Gigabit Performance<br>
> <br>
> What is the best way to benchmark network throughput? We are using 2
Gigabit<br>
> ethernet ports on our NetBackup Master/Media server, and we are not
seeing<br>
> the throughput we are expecting for backups. Since we stream multiple<br>
> classes to each drive, I'm inclined to believe that it's the network,
not<br>
> the 8 DLT drives. They sit idle a lot...shoe shining? We are using
crude<br>
> copy scripts now to try to measure throughput - anyone got something
better?<br>
> <br>
> HW - Sun E450, 2 Sun Gb PCI cards dedicated to the backup segment...<br>
> <br>
> Thanks!<br>
> <br>
> Kim<br>
> Kim Pearson<br>
> UNIX Team Lead<br>
> Guidant Corporation<br>
> kim.pearson AT guidant DOT com</font>
--=_alternative 00019158852569F3_=--
|