locked
Need help to understand if I should setup and configure a DP at the Central Office for Scheduling and Throttling if the remote one goes down? RRS feed

  • Question

  • Q1: It is my understanding that when a DP role is co-located (installed) on a Site Server, NOT a Site System, then Scheduling and Throttling is NOT available. For instance, if I install the DP role on a Primary Site Server, then Scheduling and Throttling are NOT available. However, if I install the DP role on a server that is also being used as a SUP for example, then Scheduling and Throttling WILL BE available. It this correct?

    Q2: If the answer to question 1 is YES, then if a remote DP does decide to "go down" and since the server at the Central Office has both the Site Server and DP roles installed, should a separate standalone DP at the Central Office be used in case a remote DP does decide to "go down" and then configure Scheduling and Throttling on it for this worse case scenario since Scheduling and Throttling are NOT available on the DP that is installed on the Site Server?

    Thanks

    Friday, October 10, 2014 8:33 AM

Answers

    1. Correct.
    2. Not sure what you're trying to achieve. Those throttling settings are related to the packages being send to the DP and not to the clients downloading from the DP. So, I don't see the link between installing more remote DPs for throttling... 

    My Blog: http://www.petervanderwoude.nl/
    Follow me on twitter: pvanderwoude

    • Marked as answer by Daniel JiSun Tuesday, October 14, 2014 6:18 AM
    Friday, October 10, 2014 8:44 AM
  • Correct, if all your DPs are configured with a boundary group then you have to configure the fallback source location to allow the clients to download the content from a different DP when their own DP is not available (or does not contain the content).

    I think I would configure the central DP as fallback source location, then going for the BranchCache option. Of course it does depend on things like the network connection, if it would be a valid fallback scenario.


    My Blog: http://www.petervanderwoude.nl/
    Follow me on twitter: pvanderwoude

    Saturday, October 11, 2014 6:02 AM
  • Configuring both for a location is overkill and I've never seen anyone do this. Also, it won't really work as a fallback method. The use of BranchCache will effectively override the clients and only the first two clients requiring content within that subnet will ever use that DP -- the rest will use BranchCache so what you end up with is a DP sitting there doing almost nothing most of the time.

    If you are concerned about DP availability, then put two DPs at the location or use hardware that is itself fault tolerant. If you are concerned about WAN bandwidth, then use QoS.


    Jason | http://blog.configmgrftw.com | @jasonsandys

    Saturday, October 11, 2014 12:03 PM

All replies

    1. Correct.
    2. Not sure what you're trying to achieve. Those throttling settings are related to the packages being send to the DP and not to the clients downloading from the DP. So, I don't see the link between installing more remote DPs for throttling... 

    My Blog: http://www.petervanderwoude.nl/
    Follow me on twitter: pvanderwoude

    • Marked as answer by Daniel JiSun Tuesday, October 14, 2014 6:18 AM
    Friday, October 10, 2014 8:44 AM
  • Thanks for the response. 

    What I would like to probably do will be to setup a DP at the remote site, then enable BranchCache on the remote clients so they do not try to communicate with the Primary Site Server and DP located at the Central Office simultaneously. Then IF the remote site DP goes down for whatever reason, the clients will still be able to use BranchCache to help minimize bandwidth. 

    What I am trying to figure out is since I already have a DP installed on the Primary Site Server and since Scheduling and Throttling are NOT available for this DP, would it behoove me to simply setup a separate DP at the Central Office in case the remote DP goes down, and then configure Scheduling and Throttling on the separate DP, then remove the DP installed on the Primary Site Server?

    This way the clients will be able to use BranchCache and a separate dedicated DP that can utilize Scheduling and Throttling. 

    My entire plan is to do whatever it takes to ensure that WAN link does NOT become saturated.

    Thanks


    Friday, October 10, 2014 10:52 AM
  • Why do you want to enable scheduling and throtteling on the DP that's located on the site server itself? Peter already exlained it: those options are only used by the site server when it distributes content to DPs. It has nothing to do with clients at all.

    Torsten Meringer | http://www.mssccmfaq.de

    Friday, October 10, 2014 11:12 AM
  • I DONT want to enable Scheduling and Throttling on the DP that is on the Primary Site Server, because from how I understand it, this CANNOT happen.

    Ahh okay, I see. So Scheduling and Throttling are ONLY from the MP -> DP, NOT DP -> Clients?

    What do DPs use then to deliver content to the clients, if clients in a remote location were pulling from a Central Office DP in a scenario where the local DP went down? Compression and BITS? 

    I'm trying to come up with a fallback plan IN CASE of an emergency where the remote DP actually goes offline, for whatever reason. 

    What would be the best solution to configure to ensure that if a remote DP goes down, the WAN link will not become bombarded by client requests or package downloads from the Central Office DP?

    Thanks






    Friday, October 10, 2014 11:16 AM
  • The scheduling and throttling settings are only for site server > DP. 

    The client will use BITS to download packages. Also, before a client can use a remote DP, the DP needs to be configured as a fallback source location.


    My Blog: http://www.petervanderwoude.nl/
    Follow me on twitter: pvanderwoude

    Friday, October 10, 2014 11:32 AM
  • Another note here is that if you have a DP at a location, then BranchCache provides no value. BranchCache is for locations that have no local DP to use.

    Jason | http://blog.configmgrftw.com | @jasonsandys

    Friday, October 10, 2014 6:47 PM
  • Peter,

    So if and when a DP does decide to go down in a remote location, the clients will basically be screwed if the fallback source location is NOT configured for them to communicate with the Central Office DP?

    Jason,

    Okay, I thought you could use both a DP and BranchCache simultaneously. It was my thinking that you could have a remote DP setup and have the clients enabled to use BranchCache. Then when one client pulls from its local DP, other clients will just grab the content from that client rather than get it from that local DP. I know this may seem superfluous, because I would be using both technologies (ad DP and BranchCache), but again, it is just for the sake of ensuring the WAN link does NOT become saturated if the remote DP goes down. 

    Would it make more sense to just NOT enable BranchCache UNTIL the DP goes down? I want to be more proactive than reactive. This is why I would configure both a remote DP and enable BranchCache at the remote location. 

    Does this make any sense?

    Thanks


    Friday, October 10, 2014 7:21 PM
  • Correct, if all your DPs are configured with a boundary group then you have to configure the fallback source location to allow the clients to download the content from a different DP when their own DP is not available (or does not contain the content).

    I think I would configure the central DP as fallback source location, then going for the BranchCache option. Of course it does depend on things like the network connection, if it would be a valid fallback scenario.


    My Blog: http://www.petervanderwoude.nl/
    Follow me on twitter: pvanderwoude

    Saturday, October 11, 2014 6:02 AM
  • Configuring both for a location is overkill and I've never seen anyone do this. Also, it won't really work as a fallback method. The use of BranchCache will effectively override the clients and only the first two clients requiring content within that subnet will ever use that DP -- the rest will use BranchCache so what you end up with is a DP sitting there doing almost nothing most of the time.

    If you are concerned about DP availability, then put two DPs at the location or use hardware that is itself fault tolerant. If you are concerned about WAN bandwidth, then use QoS.


    Jason | http://blog.configmgrftw.com | @jasonsandys

    Saturday, October 11, 2014 12:03 PM
  • Great! Thanks you guys, I appreciate your help.

    Since I already have a WSUS downstream server in the remote location, I would like to keep using for SCCM so I will just turn that into a local DP for the remote site for clients to use. Then I guess configure scheduling and throttling from the site server to that remote DP and as Peter suggested, configure the central office DP as the fallback for the remote clients if the remote DP ever decides to go belly up. If the DP does go offline, then I will probably just enable BranchCache for the remote clients at that point. 

    Monday, October 13, 2014 5:55 AM