| Author |
networkIO and 100Mbps NIC
|
|
| Hassan 2006-02-28, 8:23 pm |
| I see a lot of networkIO in waitstats . Does that mean I need to increase my
NIC to a 1Gbps ?
Whats a better way to find out if I am pushing too much on my n/w interface
card ?
| |
| Andrew J. Kelly 2006-02-28, 8:23 pm |
| Network waits can be caused by lots of things not just slow networks. This
isn't to say that going to 1GB won't help some but it is probably not the
real issue. First are you pulling too much data across the network? Are you
doing something network intensive like replication, data loads, linked
servers etc? What is the client and how efficient are they? If you ask SQL
Server for 10,000 rows and the client can not read them fast enough you will
get network waits. I would start by running a trace to see where your
inefficiencies are so you can narrow down the scope.
--
Andrew J. Kelly SQL MVP
"Hassan" <Hassan@hotmail.com> wrote in message
news:%23GTbp$1OGHA.2040@TK2MSFTNGP14.phx.gbl...
>I see a lot of networkIO in waitstats . Does that mean I need to increase
>my NIC to a 1Gbps ?
>
> Whats a better way to find out if I am pushing too much on my n/w
> interface card ?
>
| |
| Hassan 2006-02-28, 8:23 pm |
| Start a trace and capture what and what do I look out for once I have some
events captured ?
"Andrew J. Kelly" < sqlmvpnooospam@shadh
awk.com> wrote in message
news:urZaYF6OGHA.2268@TK2MSFTNGP09.phx.gbl...
> Network waits can be caused by lots of things not just slow networks. This
> isn't to say that going to 1GB won't help some but it is probably not the
> real issue. First are you pulling too much data across the network? Are
> you doing something network intensive like replication, data loads, linked
> servers etc? What is the client and how efficient are they? If you ask
> SQL Server for 10,000 rows and the client can not read them fast enough
> you will get network waits. I would start by running a trace to see where
> your inefficiencies are so you can narrow down the scope.
>
> --
> Andrew J. Kelly SQL MVP
>
>
> "Hassan" <Hassan@hotmail.com> wrote in message
> news:%23GTbp$1OGHA.2040@TK2MSFTNGP14.phx.gbl...
>
>
| |
| Andrew J. Kelly 2006-02-28, 8:23 pm |
| Look for events with long durations, high I/O especially ones that return a
lot of rows. To get the rows affected you will need to trace at the
statement level and use the IntegerData column. The higher these are the
more likely they will be an issue if the client or network is hampered.
--
Andrew J. Kelly SQL MVP
"Hassan" <Hassan@hotmail.com> wrote in message
news:eP3Ysc%23OGHA.344@TK2MSFTNGP11.phx.gbl...
> Start a trace and capture what and what do I look out for once I have some
> events captured ?
>
> "Andrew J. Kelly" < sqlmvpnooospam@shadh
awk.com> wrote in message
> news:urZaYF6OGHA.2268@TK2MSFTNGP09.phx.gbl...
>
>
|
|
|
|