Skip to content

Microsoft Exchange Server 2007 CCR with Standby Continuous Replication Target – an additional Disaster Recovery Option

August 4, 2010

Activating Standby Continuous Replication as a additional DR setup
Single Copy Cluster (SCC) and Cluster continuous Replication (CCR) are built on top of Windows Failover Clusters. Clustered mailbox server is name of Exchange Virtual server which holds all the exchange resources on Active failover cluster node. SCC is same method of clustering used in Exchange 2003, which has single copy of database on shared drive. This shared drive is accessible from all the nodes in Cluster. Node which holds the quorum resource can read and write in the database. CCR keeps the two copies of databases. One on active node and another one on passive nodes and provides fault tolerance in situation where one copy of database is corrupted or lost.      

With Exchange 2007 sp1, Microsoft introduced another method of high availability Standby Continuous Replication.       

To provide an addition disaster tolerance to Single Copy Cluster and Cluster Continuous Replication, we can implement Standby Copy Replication target machine at the remote site (different subnet) or in local site (same subnet). If SCR target is install at the remote site with different subnet it is required to have HUB and CAS servers available in that Subnet.          

In this post, I am demonstrating to setup SCR with CCR as an additional high availability.       

1 Virtual Machines Setup       

Figure 1. Virtual Machine Setup

 DC-01 – Services:Active directory / DNS : 192.168.2.66       

CASHUB-01  Services: Client access and Hub Transport Server:       

                                        CASHUB-01: 192.168.2.70       

                                         Mail.mywaters.com: 192.168.2.74 (receive connector)       

EXCCLUN-01 (192.168.2.67) : Active node in Cluster EXCMBXV-01(192.168.2.71)       

EXCCLUN-02 (192.168.2.68) : Passive node in cluster EXCMBXV-01(192.168.2.71)        

                            EXCMBX-01: Clustered Mailbox server Name resource 192.168.2.72       

EXCSCRT-01((192.168.2.69): Standby continues replication target: Installed Exchange 2007 Sp2 as Passive node in Cluster DREXCMBXV-01 (192.168.2.75)       

WINXP-01 – Outlook 2007 SP 2 Installed       

I have already setup two node windows 2008 SP1 cluster EXCMBXV-01 with EXCCLUN-01 (Active) and EXCCLUN-02(Passive) nodes. Cluster continuous replication cluster mailbox server name is EXCMBX-01.       

 To activate Standby continuous replication, we need to install failover clustering on the node EXCSCRT-01.  I have created cluster resource named DREXCMBXV-01 and run setup to install passive copy of Exchange 2007 server SP2.       

 Assign Full permission on EXCMBX-01 computer account to DREXCMBXV-01 computer account in Active Directory user and computer MMC console. This is required so that DREXCMBXV-01 can reset the account when moving resources from EXCMBXV-01.       

 1. Activate standby replication for all the storage groups by using shell command       

 Get-StorageGroup -Server EXCMBX-01 | Enable-StorageGroupCopy -StandbyMachine EXCSCRT-01       

Figure 1.1 Activate SCR target replication

     1.2 Get-StorageGroupCopyStatus ‘EXCMBX-01\First Storage Group’ -StandbyMachine EXCSCRT-01    

Figure 1.2 SCR Copy Status

Wait for some time so that seeding of databases from EXCMBX-01 to EXCSCRT-01 completed.       

2. Moving Site/Mailbox Server on Standby Continuous Replication database server       

2.1 Dismount database on EXCMBX-01       

 Get-MailboxDatabase –Server EXCMBX-01 | dismount-database       

Figure 2.1 Dismount All Databases

2.2 Make SCR database copy active command below will activate all the databases        

       GetSCRSources | Restore-StorageGroupCopy -StandbyMachine EXCSCRT-01       

Figure 2.2 Activate SCR database copies

2.3 STOP Cluster mailbox server       

 Stop-ClusteredMailboxServer EXCMBX-01 –StopReason “Testing Site Failure” –Confirm:$False       

Figure 2.3 Stop Clustered Mailbox Server

Shutdown both the nodes (Passive node EXCCLUN-02 first then Active node EXCCLUN-01)       

 2.4 Delete DNS record of EXCMBX-01, which will be recreated again when running setup with /recoverCMS       

Figure 2.4 Delete DNS Record

2.5 Disable copying of Storage groups from failed Clustered mailbox server       

GetSCRSources | Disable-StorageGroupCopy –Confirm:$false   

Figure 2.5 Disable Storage Group Copy

2.6 Run Setup program to recover CMS on EXCSCRT-01, enter different CMS IP address on this server.  

 Setup.com /recoverCMS /CMSName:EXCMBX-01 /CMSIPAddress:192.168.2.78  

2.7 When setup completed mount database  

Get-mailboxdatabase –server EXCMBX-01 | mount-database   

Figure 2.7 Mount databases

2.8 Restore host TTL value which is reset to default with /recovercms command  

Cluster.exe res “Network Name (excmbx-01)”  /priv HostRecordTTL=300  

Figure 2.8 Set Host Record TTL

2.9 Login to outlook and check connection with recovered clustered mailbox server on DREXCMBXV-01.  

Figure 2. 9 Outlook Connection

2.10 Check mails are available which were sent and received before failure  

Figure 2.10 Outlook mails

        3. Moving back to old CCR setup – Active / Passive node CCR with SCR as disaster recovery server.  

Old Setup -(EXCCLUN-01 and EXCCLUN-02 CCR active passive cluster and EXCSCRT-01 as standby continuous replication Target)  

To verify success at the end sent/received few mails.  

Bring both the nodes up EXCCLUN-01 and EXCCLUN-02 (Bring primary node up first)  

 3.1 Remove EXCMBX-01 clustered mailbox server configuration from EXCCLUN-01  

 Setup.com /ClearLocalCMS /CMSName:EXCMBX-01  

Figure 3. 1 Clear CMS configuration from CCR

EXCMBXV-01 is now a failover cluster with two Passive Nodes, EXCCLUn-01 and EXCCLUn-02, which each have the passive Mailbox server role installed. At this point, there is no clustered mailbox server on EXCMBXV-01 cluster.  

 3.2 Enable CMS computer account from AD which got disabled during /ClearLocalCMS operations  

 —Verifying using cluster administrator that all the Cluster resources are removed  

3.2 Cluster resources on EXCCLUn-01   

 
 

Figure 3.2 Cluster Resources on EXCCLUN-01

       3.3 Cluster Resources on Excclun-01  

Figure 3.3 Cluster Resources on EXCCLUN-02

 Remove database files from EXCCLUN-01 as it is obsolete.  

> Reseed the database from current EXCMBX01 (on eXCSCRT-01) to EXCCLUN-01 using following command:  

Get-StorageGroup -Server EXCMBX-01 | Enable-StorageGroupCopy -StandbyMachine EXCCLUN-01  

3.4 Check status of standby replication to EXCCLUn-01  

 Get-StorageGroupCopyStatus ‘EXCMBX-01\First Storage Group’ -StandbyMachine EXCCLUN-01  

Figure 3.4 SCR copy status

 If reply queue length is 0( Zero) then it is safe Dismount the database on CMS EXCMBX-01 which is on DREXCMBX-01 cluster.  

 Get-MailboxDatabase –Server EXCMBX-01 | dismount-database  

 3.5 Disable Storage group copy otherwise /recoverCMS setup will fail  

 Disable-StorageGroupCopy -Identity “EXCMBX-01\First Storage Group” -StandbyMachine EXCCLUN-01 -Confirm:$False  

Figure 3.5 Disable Storage group Copy

 3.6 Verify all the log files sequence on EXCclun-01.  

3.7 Verify all the log files sequence on EXCSCRT-01. log Sequence should match with point 3.6.  

3.8 Prepare database to be mounted on EXCCLUN-01 

 GetScrSources | Restore-StorageGroupCopy -standbyMachine EXCCLUN-01 

3.9 Stop clustered mailbox server on EXCSCRT-01 and shutdown EXCSCRT-01 

Stop-ClusteredMailboxServer Excmbx-01 –stopReason “moving to original CCR” –Confirm:$false 

3.10 Again Delete the DNS record of EXCMBX-01 on DNS server 

Figure 3.10 Delete DNS host record

Wait for DNS replication to complete if you are running multiple sites and domain controllers 

 3.11 Login to EXCCLUN-01 to recover cluster in old form, you will get the error mentioned below if standby replication was not disabled 

Figure 3.11a Errors

 If no errors recovered setup should finish 

  Setup.com /recoverCMS /CMSName:EXCMBX-01 /CMSIPAddress:192.168.2.72 

 

Figure 3.11b Recover CMS

 3.12 Set TTL on the clustered mailbox server EXCCLUN-01, DNS record will be set back to its default. 

Cluster.exe res “Network Name (excmbx-01)”  /priv HostRecordTTL=300 

3.13 Mount database 

As ReplyQueueLenght was not Zero when I dismounted and disabled replication in step 3.4 and step 3.5, I encountered errors shown below 

Figure 3.13a Database Mount Errors

Database shutdown state.. on EXCCLUn-01 

Figure 3.13b Database Header of database copy EXCCLUN-01

 

 

Missing log files details in Event ID 455 

Figure 3.13c Event ID 455 E00.log missing

Copy Missing log files from EXCSCRT-01 to EXCCLUn-01 database logs path. In my case only E00.log was missing so copied it and tried mounting database 

Get-mailboxDatabase –server EXCMBX-01 | Mount-database 

Figure 3.13d Database Mount Success

3.14 Restore copy on passive node EXCCLUN-02 

 Get-StorageGroup –server EXCMBX-01 | Resume-StorageGroupCopy 

Figure 3.14 Resume CCR replication

3.15 Open outlook and Check if all the data intact.. 

Figure 3.15 Check Data in outlook

4 Activate EXCSCRT-01 again as Standby continuous replication target 

 4.1 Clear the cluster mailbox server configuration from EXCSCRT-01 to set it back to standby continuous replication target. Run this command on EXCSCRT-01 

Setup.com /ClearlocalCMS /cmsname:excmbx-01 

Figure 4.1 Clear CMS on EXCSCRT-01

4.2 Enable the computer account of EXCMBX-01 which got disabled due the /clearlocalcms command 

Figure 4.2 Enable Computer account of Clustered Mailbox Server

4. 3 Restore SCR target replication 

 Get-StorageGroup –server EXCMBX-01 | Enable-StorageGroupCopy –standbyMachine EXCSCRT-01 –TruncationLagTime 0.00:00:00 –ReplayLagTime 5.00:00:00 

Figure 4.3 Enable Standby Continous replication

Advertisements
2 Comments
  1. scorp permalink

    Nice article.

    I have a single CCR + SCR combination hosting about 45 storage groups, holding overall 2TB data.
    (each database atleast 50 to 200 GB), then its not feasible to do SCR testing since complete reseeding has to be done.

    Since after SCR failover is complete, we will again move back to CCR on Primary datacenter.
    SCR partnership is lost, and complete SCR Re-seeding has to be done again across the WAN for 2TB for all 45 databases!

    This is a limitation in Exchange 2007 SCR.

Trackbacks & Pingbacks

  1. Exchange 2007 – Activation of DR Site and Failback to primary « SMTP Port 25

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: