Exchange Server TechCenter > Exchange Server Forums > Deploy > Exchange Site to Site High Availability .
Ask a questionAsk a question
 

QuestionExchange Site to Site High Availability .

  • Friday, June 19, 2009 5:25 AMSyed Arsalan Users MedalsUsers MedalsUsers MedalsUsers MedalsUsers Medals
     

    Planning for Huge Exhange Organization .
    Expected Mailboxes : 12000-15000
    Planning for CCR WITH SCR
    Sites :  Two Sites connected with High speed bandwidth .
    Requirement :  High Availability and smart solution .

    SITE A :

    Single Machine ------>        Two Virtual Machines : Edge + Isa Server 2006 (DMZ )
    Single Machine ------>        Two Virtual Machines : Hub Server + CAS(NLB AMONG CAS )
    Single Machine ------>        Two Virtual Machines : Hub Server + CAS
    Multiple Machines ------>     Mailbox Servers 
    SAN ------>                        EMC HIGHILY EFFICIENT AND FAULT TOLERANT .

    Maximum their are 10-15 databases but divided into four mailbox servers located into two different sites .
     
    Mailbox 1 : 

    3 storage groups and single databse in each database :SG1-SG3
    databses  :   DB1 - DB3   :    120 GB EACH 
    LOGS       :    150 GB
    RSG         :    150 GB


    Mailbox 2 : 

    3 storage groups and single databse in each database :SG4-SG6
    databses  :   DB4 - DB6   :    120 GB EACH 
    LOGS       :    150 GB
    RSG         :    150 GB

     DR Site for Site A :            SCR

    Single Machine ------>        Two Virtual Machines : Hub Server + CAS
    Multiple machines for Mailbox Servers : having same configuration as source mailbox server have + San is also on DR Site A .


    SITE B:

    Single Machine ------>        Two Virtual Machines : Edge(Backup MX )  + Isa Server 2006 (DMZ )
    Single Machine ------>        Two Virtual Machines : Hub Server + CAS(NLB AMONG CAS )
    Single Machine ------>        Two Virtual Machines : Hub Server + CAS
    Multiple Machines ------>     Mailbox Servers 
    SAN ------>                        EMC HIGHILY EFFICIENT AND FAULT TOLERANT .

    Mailbox 1 : 

    3 storage groups and single databse in each database :SG7-SG9
    databses  :   DB7 - DB9   :    120 GB EACH 
    LOGS       :    150 GB
    RSG         :    150 GB


    Mailbox 2 : 

    3 storage groups and single databse in each database :SG10-SG12
    databses  :   DB10 - DB12   :    120 GB EACH 
    LOGS       :    150 GB
    RSG         :    150 GB

     DR Site for Site B :            SCR

    Single Machine ------>        Two Virtual Machines : Hub Server + CAS
    Multiple machines for Mailbox Servers : having same configuration as source mailbox server have + No SAN here.

    Question :

    my question is related to my design whether it is fine or it need some changes please advice ,second is that we have no SAN at site B so what is the recommended approach to site to site high availability,whether i have to placed four target mailbox servers there in DR SITE A  as i have mentioned that i have no san in my site B ,please advice in this regards .

    waiting for good reply .

    Thanks
    Syed Arsalan


    syed.arsalan

All Replies

  • Monday, November 02, 2009 9:08 PMMike Crowley Users MedalsUsers MedalsUsers MedalsUsers MedalsUsers Medals
     
    It's been a while.  Is this still an open question?  Have you completed the buildout?

    This type of post would be better if we could discuss the design one element at a time.  Giving the whole thing a thumbs up or down is too broad a request.  Also your details here are tough to follow.


    Here are some bullets I have, but by no means a comprehensive review:

    -how is the backup ISA server selected (in the event of a failure)?  Is there a DNS based site selector or something else?
    -you said CCR but then also a SAN.  these are typically opposite concepts as the whole point of CCR is to remove SAN dependency
    -you have 4 ht and cas, but only two with NLB?
    i cant really follow this db/sg design
    -you said scr.  where is this?

    please be more specific with your questions so that we can supply better answers.