none
Prevent Use Of Client Access Server On DR Site RRS feed

  • Question

  • I've recently deployed Exchange 2013 in our organisation (upgraded from 2010). With our 2010 environment I had 3 mailbox servers in the DAG, 2 at head office (SITE1) and one at our DR site (SITE2). With this being a DR site the bandwidth usage was quite high throughout the day.

    I have tried to recreate this setup with 2013 but, from the articles I've read, each site with a mailbox server role also requires the client access role. This is where the problem seems to start!

    Due to the high bandwidth usage (not just Exchange) to the DR site the ECP webpage becomes extremely slow at generating pages, property windows, etc. I have seen this issue reported by other users in a similar situation. I'm guessing it's because all Exchange servers need to send data and it's having to wait longer to receive the required data from SITE2.

    We also get issues with Outlook becoming unresponsive. Users at SITE1 appear to have no issues, SITE2 I'm not aware of any reported issues. But other sites appear to select whichever client access server they like.

    I can cope to a point with the ECP slowness but is there any way we can prevent clients from selecting the SITE2 CAS over SITE1?

    Would setting the DR site CAS role setting 'isOutOfService' to True (until required in a DR situation) make a difference or would this cause more issues as then technically there is no CAS role in SITE2. My other thought was to somehow add a cost to the CAS, but I'm not sure if that would be possible.

    Thanks for any help!
    Monday, January 4, 2016 4:57 PM

Answers

  • As Ed Crowley says, your Load Balancer(s) and DNS should control which CAS servers your clients connect to.  However, Outlook clients may still connect randomly to different CAS servers to retrieve their Autodiscover settings. 

    Internal domain joined clients query Active Directory for a Service Connection Point (SCP) to retrieve their Autodiscover settings from.   If the client access servers are not configured for a particular site, the client will choose a client access server at random. 

    You can ensure the clients connect to the nearest available client access server by configuring the AutodiscoverSiteScope parameter.  Look at the following link for details and the correct syntax

    https://technet.microsoft.com/en-us/library/aa998575(v=exchg.141).aspx

    __________________________________________

    Please mark as Answer if this answers your question

    Regards,

    Shane Jackson

    Blog: https://shanejacksonitpro.wordpress.com/

    Twitter: https://twitter.com/shane00jackson

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    • Marked as answer by AndyH16 Friday, January 8, 2016 2:05 PM
    Tuesday, January 5, 2016 3:57 PM
  • To configured a preferred CAS server for Autodiscover for Outlook clients in Site 3, add Site 3 to the site scope for the preferred CAS server

    __________________________________________

    Please mark as Answer if this answers your question

    Regards,

    Shane Jackson

    Blog: https://shanejacksonitpro.wordpress.com/

    Twitter: https://twitter.com/shane00jackson

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    • Marked as answer by AndyH16 Friday, January 8, 2016 2:05 PM
    Friday, January 8, 2016 11:08 AM

All replies

  • Just because you create a CAS doesn't mean that it has to be used.  Your load balancer configuration and/or DNS entries, coupled with the URLs in the various virtual directory settings, dictate where your client traffic goes.

    Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."
    Celebrating 20 years of providing Exchange peer support!

    Tuesday, January 5, 2016 6:34 AM
    Moderator
  • Hi AndyH,

    Thank you for your question.

    We want to learn more details which is about your issue. For example:

    What is Exchange and AD infrastructure?

    Which site open ECP get slower? (As I am concerned, not site 1 and site 2, right?)

    Are there any others issues? (For example, send email become slower)

    As I am concerned, you could not configure the CAS server in DR site with “IsOutOfServer”, because the CAS server in DR site will not be considered as a proxy endpoint after we configure it. We could refer to the following link:

    http://blogs.technet.com/b/exchange/archive/2014/03/12/client-connectivity-in-an-exchange-2013-coexistence-environment.aspx 

    If there are high bandwidth usage between site 1 and site2, we could active database copy on Site 1 which includes user mailbox on site 1 and active database copy on site 2 which includes user mailbox on site 2.

    In addition, please check if there are any related errors in application log and post them for our troubleshooting.

    Best Regard,

    Jim Xu

    Please remember to mark the replies as answers if they help, and unmark the answers if they provide no help. If you have feedback for TechNet Support, contact tnmff@microsoft.com.

    Jim Xu
    TechNet Community Support

    Tuesday, January 5, 2016 6:41 AM
  • Hi Andy,

    If you run this cmdlet

    Get-ServerComponentState –Identity <ServerID>

    The majority of the components have a name like “*Proxy”. They are specific for the CAS Role. In addition to the single components which can be managed individually, there’s also a component called “ServerWideOffline”, which is used to manage the state of all components together.

    You might try this cmdlet to make the DR site CAS inactive.

    Set-ServerComponentState CAS-DR -Component ServerWideOffline -State Inactive -Requester Maintenance

    This cmdlet also applies to Client Access Servers which, when placed into maintenance mode, stops the server from acknowledging load balancer heartbeats and disables all proxy services.

    When should you change Server Component States manually?

    Perhaps the two most important scenarios where manual changes of Server Component States come into play are:

    1. Planned server maintenance
    2. Temporary isolation of some server components, so that they are not targets for proxy requests from other CAS servers any more.


    Regards,

    Satyajit

    Please “Vote As Helpful” if you find my contribution useful or “Mark As Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.

    Tuesday, January 5, 2016 9:22 AM
  • If there are high bandwidth usage between site 1 and site2, we could active database copy on Site 1 which includes user mailbox on site 1 and active database copy on site 2 which includes user mailbox on site 2.



    Best Regard,

    Jim Xu

    Just keeping Active mailboxes separated on Sites doesn't ensure site specific CAS would be used. Site2 CAS can still equally proxy Site1 Mailboxes as Site1 CAS.

    Andy needs to be switch to Bound CAS model, by having individual site specific CAS URLs.


    Regards,

    Satyajit

    Please “Vote As Helpful” if you find my contribution useful or “Mark As Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.

    Tuesday, January 5, 2016 9:28 AM
  • As Ed Crowley says, your Load Balancer(s) and DNS should control which CAS servers your clients connect to.  However, Outlook clients may still connect randomly to different CAS servers to retrieve their Autodiscover settings. 

    Internal domain joined clients query Active Directory for a Service Connection Point (SCP) to retrieve their Autodiscover settings from.   If the client access servers are not configured for a particular site, the client will choose a client access server at random. 

    You can ensure the clients connect to the nearest available client access server by configuring the AutodiscoverSiteScope parameter.  Look at the following link for details and the correct syntax

    https://technet.microsoft.com/en-us/library/aa998575(v=exchg.141).aspx

    __________________________________________

    Please mark as Answer if this answers your question

    Regards,

    Shane Jackson

    Blog: https://shanejacksonitpro.wordpress.com/

    Twitter: https://twitter.com/shane00jackson

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    • Marked as answer by AndyH16 Friday, January 8, 2016 2:05 PM
    Tuesday, January 5, 2016 3:57 PM
  • Thanks all for your help and suggestions. It certainly seems that a load balancer would be the way to go, unfortunately money is tight here so I'll have to ask nicely - I know we could benefit from using one for other services.

    The Server Component State options is an interesting one but I'm guessing it's not intended to leave it in this mode for anything other than short maintenance periods.

    As asked by Jim, the ECP is slow on both sites as they are both querying Exchange servers outside of their own site. The way I tend to work is from the Exchange Shell on that particular server, as changing settings directly on the server works fine - querying which would involve other Exchange servers (from other sites) would also cause the slowness.

    Friday, January 8, 2016 10:19 AM
  • Hi, thanks for the reply. We already have the SiteScope parameter set, 'Site1' and 'Site2' respectively.

    I'm assuming this means that when an Outlook client from Site3 queries AD it will return whichever it feels is the better link? Which goes back to my original post, if we can add an increased cost to 'Site2' this might be enough, for the short term at least.

    Friday, January 8, 2016 10:22 AM
  • To configured a preferred CAS server for Autodiscover for Outlook clients in Site 3, add Site 3 to the site scope for the preferred CAS server

    __________________________________________

    Please mark as Answer if this answers your question

    Regards,

    Shane Jackson

    Blog: https://shanejacksonitpro.wordpress.com/

    Twitter: https://twitter.com/shane00jackson

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    • Marked as answer by AndyH16 Friday, January 8, 2016 2:05 PM
    Friday, January 8, 2016 11:08 AM
  • I didn't realise the AutodiscoverSiteScope parameter could have more than one value, but that makes sense as the property is bracketed as an array would be.

    So for Site1 if I set the CAS property as "Site1,Site3,Site4,Site5..." and Site CAS as "Site2", this would in theory force clients to use Site1 (apart from Site2)? What would happen in a DR situation? Or would it just be a case of modifying this value to include all sites?

    Thanks

    Friday, January 8, 2016 12:01 PM
  • In a DR scenario, if an Outlook client cannot contact a CAS within its own site for autodiscover, it will automatically look for a CAS in the other sites

    __________________________________________

    Please mark as Answer if this answers your question

    Regards,

    Shane Jackson

    Blog: https://shanejacksonitpro.wordpress.com/

    Twitter: https://twitter.com/shane00jackson

    Note: Posts are provided “AS IS” without warranty of any kind, either expressed or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

    Friday, January 8, 2016 12:44 PM
  • That makes sense, thanks again for the help and guidance.

    • Proposed as answer by jim-xu Monday, January 11, 2016 2:27 AM
    • Unproposed as answer by jim-xu Monday, January 11, 2016 2:27 AM
    Friday, January 8, 2016 2:07 PM