Veritas-bu

[Veritas-bu] Netbackup clients with gigabit interface

2005-07-28 00:18:58
Subject: [Veritas-bu] Netbackup clients with gigabit interface
From: Dwayne.Brzozowski AT mail.va DOT gov (Brzozowski, Dwayne)
Date: Wed, 27 Jul 2005 23:18:58 -0500
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C5932B.79D485D8
Content-Type: text/plain

Also, another way to improve performance is to add another gig card and
install Sun Trunking software, assuming the switch you are connected to has
ether channel cabapility. My NB master is a Sun 880 , failover master is
also 880, and 2 media servers are also 880's. All 3 have with 4 X 1.2 GHz
procs and 8 gigs of ram. Other media server has 4 X 900 MHz procs and 8 gigs
of ram. Also,  4 have dual gig cars going to a Cisco 6509 with Sun Trunking
on the server side, to have a huge pipe to push data to tape, and ether
channel enabled on the cisco, to push data to master/media servers as fast
as the clients can feed it. I have 1-10 drive LTO2 STK library, 1-10 drive
STK DLT-7k library, and 8 -9940b drives in a STK Powderhorn silo. All
clients, if possible, have a dedicated gig/quad,hme card directly connected
to the Cisco. LTO's are fc going to dual brocades. Depending on how
optimized the client connection is plays a huge factor in backup speeds.
When I notice a client running slower than I think I should, the problem is
always the same, the SA who owns the server forgot to hard code the ndd
settings into a startup script. 
 
 
-Dwayne
 
 
Dwayne J. Brzozowski 
Department of Veterans Affairs 
Austin Automation Center 
Team Lead-Open Systems Support 
email:dwayne.brzozowski AT mail.va DOT gov 
phone:512-326-6728 


  _____  

From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf Of
dmoynahan AT lic.co DOT nz
Sent: Sunday, July 24, 2005 11:02 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: RES: [Veritas-bu] Netbackup clients with gigabit interface 



I've just joined the group. 
Have just completed the 5.1 5 day course. 

i upgraded most servers to gigabit  as many of my backups were maxing out at
11000 kb/sec on 100 mbit ethernet. 
Putting gigabit in releived the bottle neck and I get upto 56000 kb/sec
witha max of 77000 kb/sec and two steams going to two 9840C drives. 
The SAN disk is StorageTek D178 disk array. 
Even Netware backups reduced to one quarter of the elapsed time when the
file system was moved to the D178. 
As soon as you remove one constraint, you will find another. 
No matter what, gigabit is a big pipe, even when shifting 100+Gb files
around. 

/opt/openv/netbackup# cat NET_BUFFER_SZ 
131078 

/opt/openv/netbackup/db/config# cat NUMBER_DATA_BUFFERS 
16 

/opt/openv/netbackup/db/config# cat SIZE_DATA_BUFFERS   
262144 

Hope these figures help. 

The Master/media server is a Sun 280R with two processors and 1Gb memory. 

Any suggestions on the above sizes? 


Thanks

------_=_NextPart_001_01C5932B.79D485D8
Content-Type: text/html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">


<META content="MSHTML 6.00.2900.2668" name=GENERATOR></HEAD>
<BODY>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2>Also, another way to improve performance is to add another 
gig card and install Sun Trunking software, assuming the switch you are 
connected to has ether channel cabapility.&nbsp;My NB master is a Sun 880 , 
failover master is also 880, and 2 media servers are also 880's. All 3 have 
with 
4 X 1.2 GHz procs and 8 gigs of ram. Other media server has 4 X 900 MHz procs 
and 8 gigs of ram. Also, &nbsp;4 have dual gig cars going to a Cisco 6509 with 
Sun Trunking on the server side, to have a huge pipe to push data to tape, and 
ether channel enabled on the cisco, to push data to master/media servers as 
fast 
as the clients can feed it. I have 1-10 drive LTO2 STK library, 1-10 drive STK 
DLT-7k library, and 8 -9940b drives in a STK Powderhorn silo. All clients, if 
possible, have a dedicated gig/quad,hme card directly connected to the Cisco. 
LTO's are fc going to dual brocades. Depending on how optimized the client 
connection is plays a huge factor in backup speeds. When I notice a client 
running slower than I think I should, the problem is always the same, the SA 
who 
owns the server forgot to hard code the ndd settings into a startup script. 
</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2>-Dwayne</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><FONT face=Arial 
color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=ltr align=left><SPAN class=778005503-28072005><!-- Converted from 
text/rtf format -->
<P><SPAN lang=en-us><FONT face=Arial size=2>Dwayne J. Brzozowski</FONT></SPAN> 
<BR><SPAN lang=en-us><FONT face=Arial size=2>Department of Veterans 
Affairs</FONT></SPAN> <BR><SPAN lang=en-us><FONT face=Arial size=2>Austin 
Automation Center</FONT></SPAN> <BR><SPAN lang=en-us><FONT face=Arial 
size=2>Team Lead-Open Systems Support</FONT></SPAN> <BR><SPAN lang=en-us><FONT 
face=Arial size=2>email:dwayne.brzozowski AT mail.va DOT gov</FONT></SPAN> 
<BR><SPAN 
lang=en-us><FONT face=Arial size=2>phone:512-326-6728</FONT></SPAN> 
</P></SPAN></DIV><BR>
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma size=2><B>From:</B> veritas-bu-admin AT mailman.eng.auburn 
DOT edu 
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] <B>On Behalf Of 
</B>dmoynahan AT lic.co DOT nz<BR><B>Sent:</B> Sunday, July 24, 2005 11:02 
PM<BR><B>To:</B> veritas-bu AT mailman.eng.auburn DOT edu<BR><B>Subject:</B> 
RES: 
[Veritas-bu] Netbackup clients with gigabit interface <BR></FONT><BR></DIV>
<DIV></DIV><BR><FONT face=sans-serif size=2>I've just joined the group.</FONT> 
<BR><FONT face=sans-serif size=2>Have just completed the 5.1 5 day 
course.</FONT> <BR><BR><FONT face=sans-serif size=2>i upgraded most servers to 
gigabit &nbsp;as many of my backups were maxing out at 11000 kb/sec on 100 mbit 
ethernet.</FONT> <BR><FONT face=sans-serif size=2>Putting gigabit in releived 
the bottle neck and I get upto 56000 kb/sec witha max of 77000 kb/sec and two 
steams going to two 9840C drives.</FONT> <BR><FONT face=sans-serif size=2>The 
SAN disk is StorageTek D178 disk array. </FONT><BR><FONT face=sans-serif 
size=2>Even Netware backups reduced to one quarter of the elapsed time when the 
file system was moved to the D178.</FONT> <BR><FONT face=sans-serif size=2>As 
soon as you remove one constraint, you will find another.</FONT> <BR><FONT 
face=sans-serif size=2>No matter what, gigabit is a big pipe, even when 
shifting 
100+Gb files around.</FONT> <BR><BR><FONT face=sans-serif 
size=2>/opt/openv/netbackup# cat NET_BUFFER_SZ </FONT><BR><FONT face=sans-serif 
size=2>131078</FONT> <BR><BR><FONT face=sans-serif 
size=2>/opt/openv/netbackup/db/config# cat NUMBER_DATA_BUFFERS </FONT><BR><FONT 
face=sans-serif size=2>16</FONT> <BR><BR><FONT face=sans-serif 
size=2>/opt/openv/netbackup/db/config# cat SIZE_DATA_BUFFERS &nbsp; 
</FONT><BR><FONT face=sans-serif size=2>262144</FONT> <BR><BR><FONT 
face=sans-serif size=2>Hope these figures help.</FONT> <BR><BR><FONT 
face=sans-serif size=2>The Master/media server is a Sun 280R with two 
processors 
and 1Gb memory.</FONT> <BR><BR><FONT face=sans-serif size=2>Any suggestions on 
the above sizes?</FONT> <BR><BR><BR><FONT face=sans-serif 
size=2>Thanks</FONT></BODY></HTML>

------_=_NextPart_001_01C5932B.79D485D8--