One of our customer have network access paths. We require to specify how the end branch DC should communicate to which another site DC (SCOM agent installed) and then inturn that monitoring date to reach head office SCOM managemnet server. We have 400 branches containing only two servers at each branch. Can't afford or feasible to have management server at each branch either. In the current possiblities of GUI i could find the agent proxying is possible but the behavior as to which end machine should connect to which next hop (agent) and then make up full path is not possible to consider to network paths as well as bandwidth considerations. I think it's a big problem in that case, if the design really does not allow thisShahid Roofi

There needs to be a direct path - with DNS set up at the management server. What is it you are trying to do at each branch? There are only two servers in each place - so what is your monitoring scenario. If you keep it minimal, you might try having gateways owning individual agents. Don't think proxy - agent proxy is not about path to server.Microsoft Corporation

Apprciate you input Dan. Let me try to elabore more. My customer is big bank with lots of small branches with single AD server only. On top of that, all sub branches are not directly accessible from one single Head Office, rather some branches have some hub site which indirectly connects to another core site and ultimately reaches to Head Office. Some network links are around 16K which we need to protect at best from unusal traffic. There would be only one SCOM Management Server and would be at Head Office. Can't afford another server just to monitor 250 Domain controllers in all branches. We require that agents at each location should ideally have [change proxy agent] setting so that we can specify the path to complete monitoring to reach ultimately at Head Office. What we've found is the [Enable Agent Proxy] which we understand to enable at HUB/Core sites but does not suffice to our requirements as possible with prior strategy i mentioned.Shahid Roofi

Apprciate you input Dan. Let me try to elabore more. My customer is big bank with lots of small branches with single AD server only. On top of that, all sub branches are not directly accessible from one single Head Office, rather some branches have some hub site which indirectly connects to another core site and ultimately reaches to Head Office. Some network links are around 16K which we need to protect at best from unusal traffic. There would be only one SCOM Management Server and would be at Head Office. Can't afford another server just to monitor 250 Domain controllers in all branches. We require that agents at each location should ideally have [change proxy agent] setting so that we can specify the path to complete monitoring to reach ultimately at Head Office. What we've found is the [Enable Agent Proxy] which we understand to enable at HUB/Core sites but does not suffice to our requirements as possible with prior strategy i mentioned.Shahid Roofi

Hi I don't think you've correctly understood the agent proxy setting in OpsMgr. This is a setting that is required for some MPs to function - Active Directory, Exchange, the monitoring of clusters. Basically an agent is discoverying resources that are not local on the agent ... AD Topology, Exchange Topology, cluster resources on the cluster virtual server. It has nothing to do with one agent monitoring another agent or in setting a path from an agent to a management server. Agent proxy is irrelevant to your design. And if you are looking at deploying the more complex management packs such as AD then agentless monitoring isn't supported. So you'll need an agent on each server you want to monitor and that will need to be able to communicate reliably with the Management Server. If you can't do that then OpsMgr can't help you for the branch offices. Cheers Graham View OpsMgr tips and tricks at http://systemcentersolutions.wordpress.com/

Hi I don't think you've correctly understood the agent proxy setting in OpsMgr. This is a setting that is required for some MPs to function - Active Directory, Exchange, the monitoring of clusters. Basically an agent is discoverying resources that are not local on the agent ... AD Topology, Exchange Topology, cluster resources on the cluster virtual server. It has nothing to do with one agent monitoring another agent or in setting a path from an agent to a management server. Agent proxy is irrelevant to your design. And if you are looking at deploying the more complex management packs such as AD then agentless monitoring isn't supported. So you'll need an agent on each server you want to monitor and that will need to be able to communicate reliably with the Management Server. If you can't do that then OpsMgr can't help you for the branch offices. Cheers Graham View OpsMgr tips and tricks at http://systemcentersolutions.wordpress.com/

Thanx Graham for correcting. Note however that we are concluding a big limitation of SCOM that it cannot indirectly monitor. Atleast for my current customer, we cannot move forward with single MG Server at HO and some branches not directly accessible !Shahid Roofi