For both the generator and distributor the default value is every 60
seconds. That means that every 60 seconds the generator (or distributor)
fires to see if there's work to be done.
You can change the value to decrease (or increase for that matter) the
period. The smaller the number to more frequently the generator (or
distributor) fires. Of course there is an inverse relationship between
the quantum duration and the resources consumed by SQLNS - the smaller
the on duration the more resources that are consumed.
Since you are using the default value of one minute, I would doubt that
this is the problem behind your delayed notifications.
I would start digging into Performance Monitor looking for clues that
would lead you to the bottle neck.
It would also be worthwhile to manually submit a notification through
the delivery channel to see if the channel itself is the problem.
HTH....
Joe
JM wrote:
> I am currently using the default values. What does the Quantum do and
> what would be a value that I could set to make the messages
> instantanious?
>
> "Patrice Truong [MS]" <patricet DeleteThis @online.microsoft.com> wrote in message news:<OmAN7RmQEHA.132 DeleteThis @TK2MSFTNGP09.phx.gbl>...
>
>>You might also want to check your QuantumDuration. Are you using the default
>>values or did you specify your own QuantumDuration for both the generator
>>and the distributor?
>>
>>Patrice.
>>
>>--
>>This posting is provided "AS IS" with no warranties, and confers no rights.
>>Use of included script samples are subject to the terms specified at
>>http://www.microsoft.com/info/cpyright.htm.
>>
>>
>>
>>"Joe Webb" <joew DeleteThis @webbtechsolutions.com> wrote in message
>>news:%23Yzjk2lQEHA.3476@tk2msftngp13.phx.gbl...
>>
>>>JM -
>>>
>>>Are you using the standard edition of the enterprise edition? The
>>>enterprise edition is more scalable and allows more threads for the
>>>generator and distributor.
>>>
>>>What deployment architecture are you using? Is everything installed on
>>>one computer?
>>>
>>>How many notification are going through your system?
>>>
>>>Have you check with Perfmon to see what the bottleneck is? Here's an
>>>excerpt from BOL that may help with Perfmon:
>>>
>>>
>>>Component-Level Objects
>>>Notification Services provides the component-level performance objects
>>>shown in the following table.
>>>
>>>Topic Description
>>>NS$instance_name: Delivery Channels Object Provides information about an
>>>individual delivery channel.
>>>NS$instance_name: Distributors Object Provides information about an
>>>individual distributor.
>>>NS$instance_name: Event Providers Object Provides information about an
>>>individual event provider.
>>>NS$instance_name: Generator Object Provides information about the
>>>generator component for the instance.
>>>
>>>These component-level objects monitor the components as they are updated
>>>on the system where the component runs.
>>>
>>>Application-Level Objects
>>>Notification Services provides the application-level performance objects
>>>shown in the following table.
>>>
>>>Topic Description
>>>NS$instance_name: Events Object Provides information about events for an
>>>application.
>>>NS$instance_name: Notifications Object Provides information about
>>>notifications created by an application.
>>>NS$instance_name: Subscriptions Object Provides information about the
>>>subscriptions for an application.
>>>NS$instance_name: Vacuumer Object Provides information about the
>>>application vacuuming process.
>>>
>>>These application-level objects query the application databases, and can
>>>be run on any server that has the NS$instance_name service installed and
>>>running, but the counter data reflects only the activity that takes
>>>place on the local server.
>>>
>>>Instance-Level Objects
>>>Notification Services provides the instance-level performance object
>>>shown in the following table.
>>>
>>>Topic Description
>>>NS$instance_name:Subscribers Object Provides information about the
>>>subscribers for an instance of Notification Services.
>>>
>>>The instance-level object queries the application databases, and can be
>>>run on any server that has the NS$instance_name service installed and
>>>running, but the counter data reflects only the activity that takes
>>>place on the local server.
>>>
>>>
>>>
>>>
>>>HTH....
>>>
>>>Joe Webb
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>JM wrote:
>>>
>>>
>>>>I have a NS instance running that sends messages throigh MSN Messenger
>>>>using LCS. It takes about 30 secs for the message to be recieved. Is
>>>>there any timing settings in NS that could effect this?<!-- ~MESSAGE_AFTER~ -->
>> Stay informed about: Slow Notifications?? 
FAQ
Profile
Private Messages
Log in
