How to set up high availability CAS?

Hi,
I have a domain, let's call it a.com with 2 sites, let's call the sites siteA and siteB. I have 2 CAS-servers installed, one in each site. Autodiscover is defined with the two CAS-addresses and the internal URL's are the same for both servers, registered
with both addresses in DNS.
The config works OK, BUT I would like to have control of which CAS a clients connects to, I want the clients to connect to the CAS in the same site as the client itselv. Now the clients sometimes connects to the CAS in the opposite site, depending on which
address is returned first from DNS.
How can I control this so the client connects to the nearest CAS if this is running? Only if the CAS is down I want to connect to the other CAS.
Thor-Egil

Hi,
For internal clients, yes, your requirement can be met by setting the autodiscover site affinity by using the “AutoDiscoverSiteScope” parameter in Set-ClientAccessServer command:
https://technet.microsoft.com/zh-cn/library/bb125157(v=exchg.150).aspx
For external clients, I am afraid “No” because the “autodiscover.domain.com” only points to one CAS server.
Thanks,
Simon Wu
TechNet Community Support

Similar Messages

  • Outlook Anywhere, high availability cas issue.

    We are beginning a deployment of a 3 node exchange 2013 dag/cas group and have hit a snag with outlook anywhere.
    the configuration:
    SIDHEX01A - DAG member and CAS
    SIDHEX02A - DAG member and CAS
    SIDHEX03A - DAG member and CAS
    We have one live database running on SIDHEX02A with passive copies on 1 and 3. We have 5 live users functioning as a test group.
    DNS: we have internal DNS set to resolve mail.domain.net as all three servers. We have public DNS set to resolve mail.domain.net to an external ip on our firewall. From there we forward ports 25, 80, and 443 to server SIDHEX02A.
    All 3 servers have Certs for mail.domain.net applied to iis and smtp.
    We have set the external ( and internal names) of all the exchange services to mail.domain.net including owa, ecp, outlook anywhere, oab, activesync, and ews.
    Our next step plan was to replace the direct port forward with a load balancer completing the high availability cas, when we hig hthe following snag:
    The DAG works prefectly.
    OWA, activesync, ecp, oab, ews all function fine.incomming and outgoing mail works perfectly. we simulate cas load balancing by changing the port forwarding to direct to a different cas server.
    We are however experiencing some weird behavior with outlook anywhere.
    We have a large number of users that work completely off-prem and will be using outlook with out ever connecting to the internal network at all. When we try to configue outlook to connect to the CAS we are trying to use the following:
    server: mail.domain.net
    exchange proxy: mail.domain.net
    This doesn't work, the outlook client never finds the server.
    testexchange reports that it is unable to ping the proxy server (it is pingable) after corectly checking certs and doing all the other communications.
    Attempting to ping RPC proxy mail.domain.net.
      RPC Proxy can't be pinged.
     Additional Details
    A Web exception occurred because an HTTP 404 - NotFound response was received from Unknown.
    Headers received:
    request-id: e9462008-75e2-4990-9d0f-a6c5514f5bce
    X-CasErrorCode: EndpointNotFound
    Persistent-Auth: true
    X-FEServer: SIDHEX02A
    Content-Length: 0
    Cache-Control: private
    Date: Mon, 30 Dec 2013 02:46:34 GMT
    Server: Microsoft-IIS/8.0
    X-AspNet-Version: 4.0.30319
    X-Powered-By: ASP.NET
    Elapsed Time: 21308 ms. 
    Now if we tell the outlook client to use SIDHEX02A as the server name outlook will resolve the mailbox and everything apears great, except for two issues.
    With a load balancer implimented we wouldnt know what the target server name was.
    Secondly, we see some sort of link in outlook to the internal server name. On a cas failover outlook functions perfectly until it is closed and reopens with around a 60 second "trying to connect" before finally resolving the mailbox. This apears to
    be permanent, if we start an outlook profile on sidhex03a for example, it will have 60 sec delays on 01 or 02, but never 03. Its as if the client is looking for its "home server" before considering alternitives.
    Its posible that the load balancer would resolve some of this and we are happy to do that, we have also looked at ARR since this blog post seems to imply its required.
    http://blogs.technet.com/b/exchange/archive/2013/07/19/reverse-proxy-for-exchange-server-2013-using-iis-arr-part-1.aspx but we dont want to add complexity over a problem until we understand why its needed.
    Also of note: we dont have a cert for autodiscover.domain.net, nor is it publicly resolvable.
    we dont have any 2010 servers, this is a clean 2013 envroment.
    Thank you for taking the time to review this issue, I will provide any follow up info I can.

    Hi,
    Firstly, I’d like to explain, the host name autodiscover.domain.com should be in the certificate to ensure Autodiscover service can work externally.
    And I’d like to confirm if the test users are used internally and how about directly accessing the following URL:
    https://autodiscover.domain.com/autodiscover/autodiscover.xml
    Additionally, according to your description, you forward ports 25, 80, and 443 to server SIDHEX02A.
    Then to understand more about the issue, I’d like to confirm if you set all SMTP,HTTP and HTTPS requests forward to the server SIDHEX02A and how to achieve that. Maybe this is the key point of the issue.
    Thanks,
    Angela Shi
    TechNet Community Support

  • How to perform up high availability upgrade with database changes?

    Hi,
    Can someone point me to pertinent documentation for performing live upgrades? Specifically I'm looking for best practice for setting up and attaining High Availability especially during application upgrades which have db changes.
    Thanks in advance,
    Rich

    Hi,
    For internal clients, yes, your requirement can be met by setting the autodiscover site affinity by using the “AutoDiscoverSiteScope” parameter in Set-ClientAccessServer command:
    https://technet.microsoft.com/zh-cn/library/bb125157(v=exchg.150).aspx
    For external clients, I am afraid “No” because the “autodiscover.domain.com” only points to one CAS server.
    Thanks,
    Simon Wu
    TechNet Community Support

  • Office Web Apps farm - how to make it high available

    Hi there,
    I have deployed OWAs for Lync 2013 with two servers in one farm. When I was testing, e.g. the case when other server is down I found out the following error:
    Log Name:      Microsoft Office Web Apps
    Source:        Office Web Apps
    Date:          17/03/2014 22:59:54
    Event ID:      8111
    Level:         Error
    Computer:      OWASrv02.domain.com
    Description:
    A Word or PowerPoint front end failed to communicate with backend machine
    http://OWASrv01:809/pptc/Viewing.svc
    Does that means, that OWAs cannot setup as high avaiable when using standalone OWAs only? Above error appeared when the server OWASrv01 was rebooting.
    Petri

    Hi Petri,
    For your scenario, you have achieve a  high performance for your Office Web Apps farm but not high availability. 
    For performing a high availability Office Web Apps farm, you can refer to the blog:
    http://blogs.technet.com/b/meamcs/archive/2013/03/27/office-web-apps-2013-multi-servers-nlb-installation-and-deployment-for-sharepoint-2013-step-by-step-guide.aspx 
    Thanks,
    Eric
    Forum Support
    Please remember to mark the replies as answers
    if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact [email protected]
    Eric Tao
    TechNet Community Support

  • How can I achieve high available solution for directory server

     

    You can start with deploying multi master replication which will give you 2 servers available for writes (and as many read-only consumers as you want).
    You can also install Directory Server in a Cluster (using Sun Cluster) which will provide more failover capabilities.
    If you combine both, you should be able to have almost no downtime.
    You can also use the Directory Proxy Server (aka iDAR) to provide transparent failover for client applications.
    I hope this help.
    Regards,
    Ludovic.

  • How to configure high availability and disaster recovery? And user authenticate

    We are in the process of rolling out our online help which was created using Robohelp.   In our initial rollout we will provide access to the files via our Client Portal which requires authentication.  We are also planning for our next version where we intend to implement Robohelp server functionality.
    Our IT team is looking at options on how to configure for High Availability and Disaster Recovery.  It seems that Robohelp doesn't have any built-in functionality in this area.  In addition we require that our users authenticate.  The options for the server version seem to be more internally focused and we would need to solution the authentication using a third party.
    Would anyone be willing to share their approach in these areas?  Would you be willing to participate in a conference call with our IT Professionals?

    Hello again
    I see my good friend Peter replied to your LInkedIn post where you cross-posted the same question. For those here that have no clue what Peter stated, here it is:
    What are you seeking to recover? Your projects? Your outputs? This sounds like a question more appropriate to Disaster Recovery consultants and far wider reaching than RoboHelp. To me it seems like a question your IT people should be asking direct to such consultants who would expect a fee for their advice.
    I would agree with Peter's reply.
    I'll also go further and ask what exactly is being done in this realm for the application? Help files generally are there to support an application on the server. So whatever you are doing for the applciation should also be able to be used for the WebHelp, FlashHelp or web based AIR Help files, no?
    Cheers... Rick
    Helpful and Handy Links
    RoboHelp Wish Form/Bug Reporting Form
    Begin learning RoboHelp HTML 7, 8 or 9 within the day!
    Adobe Certified RoboHelp HTML Training
    SorcerStone Blog
    RoboHelp eBooks

  • Highly Available Management Servers in SCVMM 2012

    Dear All,
                I need to build Highly Available Management Servers (2) (SCVMM 2012) for my environment.  I am interested in following things:
    1.  What are my options.  I have done little reading and seek confirmation that without building a windows failover cluster, one is not able to build HA SCVMM Management Servers?  Am I correct in understanding this?  If so, do I need
    a shared disk as well to create a cluster first and then install Management Servers? 
    2.  What service accounts or groups are need during the configuration and what would be their permissions on respective servers such as Hyper-V Hosts and in Active Directory?
    3.  Do we really need to create a container in Active Directory for an HA SCVMM scenario?
    4.  Can HA solution be implemented without the WFC?
    5.  Would appreciate, if you could share some of the links to blogs/articles about step by step instructions to it.
    Thanks in advance.

    easy.  create 2 node  cluster-nested virtual or physical, no matter.  if virtual, set affinity (rather lack of to keep them on separate nodes-also build  a sql cluster as well and also keep them on separate nodes)follow the articles. 
    I found no less than 10 step by steps when searching for when I built the HA SCVMM / SQL environment for my company.
    MS article and third party, as you can see they are pretty straight forward to follow.  Best of luck.
    http://technet.microsoft.com/en-us/library/gg610678.aspx
    http://www.thomasmaurer.ch/2013/08/how-to-install-a-highly-available-scvmm-management-server/
    Brian

  • Highly Available BIP Clusters

    Thread for HA and clustering questions from the blog entry here:
    http://blogs.oracle.com/xmlpublisher/2007/09/07#a504

    easy.  create 2 node  cluster-nested virtual or physical, no matter.  if virtual, set affinity (rather lack of to keep them on separate nodes-also build  a sql cluster as well and also keep them on separate nodes)follow the articles. 
    I found no less than 10 step by steps when searching for when I built the HA SCVMM / SQL environment for my company.
    MS article and third party, as you can see they are pretty straight forward to follow.  Best of luck.
    http://technet.microsoft.com/en-us/library/gg610678.aspx
    http://www.thomasmaurer.ch/2013/08/how-to-install-a-highly-available-scvmm-management-server/
    Brian

  • WLC 5508 - High Availability

    I am upgrading the Wireless Infrastructure with two 5508 WLC.
    I am setting up High Availability, but I think is not quite working.
    Primary Controller = WLC1
    Secondary Controller = WLC2
    LAP = LAP1
    LAP1 has WLC2 as the primary controller for HA
    LAP1 has WLC1 as the secondary controller for HA
    While connected to LAP1, I shutdown WLC2. After ~ 20 seconds, LAP1 move to WLC1.
    I lost connection from LAP1
    Don't LAP1 should move with all its clients to WLC1?
    Am I missing something in my configuration?
    Thanks for all your help in advance

    Hello,
    Let me clarify high availability/failover here. As previously mentioned, you can set the heartbeat times to reduce WLC failure detection time. Once an AP determines it has lost the connection to it's controller, it will perform the discovery/join process to the configured Secondary WLC. During this process, the AP will need to verify configuration from the secondary WLC and the Radios will get reset.
    So, you will not experience a perfectly seamless transition when an AP moves from WLC1 to WLC2. Detection time with defaults is 30 second heartbeat, so I would expect anywhere from 15-45 second outage on a WLC failure scenario. The network will then be operational again and clients should be able to connect.
    Thanks,
    -Patrick Croak
    Wireless TAC

  • High Availability, User Profiles, and UPRE

    I'm setting up High Availability using Metalogix Replicator to mirror the Content DBs. There's a Primary (current production SP2010 environment) and a Secondary environment. Both consist of a WFE, SQL, and APP server.
    User Profiles are set up on both. I created the Secondary environments by copying Profile, Social, and Synch DBs from the Primary SQL server, and creating the UP using the same DB names.
    Since then, we have created new Audiences and new User Profile settings.
    So the two choices are:
    1. Manually update both environments.
    2. Use UPRE to keep them synchronized. I've looked over
    https://technet.microsoft.com/en-us/library/cc663011(v=office.14).aspx and it seems like a good solution.
    I'm not sure whether to use synchronous or asynchronous. I was also wondering if anyone has experience with setting this up, and if there's any chance that I'll mess up my current Primary User Profile service/synch by using UPRE.
    Has anyone been down this road? Is this what I'm looking for?
    Thanks,
    Scott

    > I'm setting up High Availability using Metalogix Replicator to mirror the Content DBs
    Wouldn't recommend it. It makes schema changes to the databases and is completely unsupported by Microsoft.
    UPRE would be appropriate.
    Trevor Seward
    Follow or contact me at...
    &nbsp&nbsp
    This post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

  • IRecruitment High Availability functionality

    Hello,
    Can some one provide a procedure to setup the High Availability functionality in iRecruitment?
    Thanks for your help!
    --Kumar                                                                                                                                                                                                                                                                                       

    Follow these steps to set up high availability:
    Set the “IRC: Availability Mode” profile option. This defines the mode in which the primary and the high availability instances are run. You must set this profile option on the primary and the high availability instances to transition between the two instances to support the candidate job site.
    Setting up the High Availability Instance. This would primarily involve:
    Building the high availability instance. To do this you can either make a full copy of the primary instance or install a clean instance and create all the data required for the iRecruitment Site Visitor functionality applicable to the High Availability pages.
    Setup the other high availability profile options.
    These steps are technical in nature and for detailed list of setup instructions please refer to the Oracle iRecruitment implementation guide.
    Thanks

  • High Availability Test Cases for ECC6.0, PI 7.0, BI 7.0???

    Dear All
    I am in a pretty difficult situation here. I have to write test cases for the components in the High Availability environment. We are running AIX 5L with HACMP/XD clusters ACTIVE/PASSIVE setting. The following components are in the landscape, ECC6.0, PI 7.0 BI 7.0 cFolders 4.0 Portal 7.0 SLD 2004s and Solman 4.0. I have to write test cases for the above components especially ECC, PI BI and cFolders in the HA environment as to
       1.<b>WHAT WILL BE THE EFFECT ON OTHER COMPONENTS IF ECC</b> <b>FAILOVER  and so on for PI and BI</b>,
       2. How do I check the consistency of the system after a failover, since the time for a failover will be anywhere between 3 - 4 mins.
    Can anyone pls give me a 5 or 6  checks like
    Checking RFC connections
    ABAP dumps
    Consistency check (txn: SICK)   and so on
    In the test cases I have to make sure that after a failover the system is up and ready to perform the activities that it was able to before the Failover......
    Will appreciate any suggestions and will definitely reward points for any helpful answers....Thanks in advance.

    Hi Shailesh,
    If you want you can activate SCM_SNC_GEN_1 depending upon business needs or configuration.
    As per my understanding there is no need to activate any business function but if you need additional feature of SNC to be used then refer below help link for more details:
    http://help.sap.com/saphelp_SNC700_ehp01/helpdata/en/a2/6d67d9b60149cb887acf72ae0b90ac/frameset.htm
    In ECC you are going for EPh5 which means no need to activate any additional business function(Normally SNC needs EPh4 only for some the latest feature to be use like in workorder collaboration ManufacturingWorkOrderInformation and ManufacturingWorkOrderProductionProgressNotification XML are used if ECC is with EPh4)
    Regards,
    Nikhil

  • How to set up an automatic upper case at the begining of a sentence

    Can someone help me if or how to set up an automatic upper case at the beggining of each sentence. I am doing huge amount of writing and just got fed up with it.

    What application? Pages? If so, that option (and many others) is not available in the current version. You can use the previous version which should be in your application folder. Or you can do what I finally did and use Open Office or Libre Office.

  • Exchange server 2013 CAS server high availability

    Hi
    I have exchange server 2010 sp3(2 MB, 2Hub/Cas) servers.
    Planning to migrate to exchange server 2013.( 2 cas servers and 2 mbx servers).
    I dont want to go all traffic single so i am keeping the role separate..
    In exchange 2010 i achieved hub/CAS high availability through NLB.
    In exchange 2013 how to acheive this...
    Please share your suggestions with document if possible...

    Here ya go:
    http://technet.microsoft.com/en-us/library/jj898588(v=exchg.150).aspx
    Load balancing
    and
    http://technet.microsoft.com/en-us/office/dn756394
    Even though it says 2010, it applies to 2013 vendors as well.
    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.

  • How to tell if REQ1 is set active high

    Hi I am using the 6533 PCI-DIO-32HS card in a PC with traditional DAQ, looking at an existing design.  I could not find how to tell if REQ1 is set active high or low, guessing it would be set up in the DIO confg vi.

    thanks for the additional information, yes will have to wait for the external hardware.  I did have a general question about the 6533 data and control line outputs using traditional DAC.  I think for this design it is set up as a pattern IO transfer using internal REQ and an internal clock set up as a VI input, (DIO START).  Anyhow I am monitoring the card outputs with a logic analyzer and it looks like an 8 element array (of 8 bit integers) is being used for the data input (to the DIO WRITE VI) with the first 4 elements being written out on DIOA0-A7, DIOB0-B7, DIOC0-C7, and DIOD0-D7 followed by the last 4 elements also written out to DIOA0-A7, DIOB0-B7, DIOC0-C7, and DIOD0-D7.  Anyhow it looks like the clock input to the (DIO Start VI) controls the pulse width of the data.  It looks like REQ1 goes high approximately 78 microseconds before the first 32 bits are written out followed by 5 microseconds of data (with the clock set at 200k), finally followed by the second data group (5 microseconds of data).  It then looks like REQ1 stays high for 2-4 ms after the second group of 32 bits is written out before going low.   Anyhow if you are trying to clock this data into a storage device I would expect to see rising clock edges when both the first and second group of data appear on DIOA0-A7..DIOD0-D7.  I monitored all the other control lines pclk1,2 ack1 and did not see signal levels other than 0, again I do not have the DIO card connected to anything except a logic analyzer.  Would you know if it is typical to use REQ1 as a clock input to a storage device or if this can be done?

Maybe you are looking for