Database Availability Group is a group of up to 16 servers that holds the set of replicated Mailbox databases and DAG is the boundary of replication and the failoverswitchover process among those mailbox databases. DAG member servers can be placed in different sites and Database replication can occur over multiple sites.
We will discuss the Database Availability Group concepts on the next 5 chapters, this will be shot introduction on the high availability options that are available in the previous version of exchange 2010.
High Availability in Exchange Server 2003
Exchange Server 2003 depends on Windows Clustering for high availability and that too, the High Availability solution or the redundancy will be available only at the hardware level and not at the database level.
Exchange 2003 cluster nodes shared the same storage system. If the active cluster node suddenly became unavailable, the Exchange Virtual Server (EVS) and any relevant cluster resources would fail over to the passive node and the end users could then continue to access their mailbox.
With Windows Server 2003 Enterprise Edition and Exchange Server 2003 Enterprise Edition, We can set up 8 node cluster for redundancy. With Exchange 2003, we have the high availability option at the server level and not at the StoreDatabase level.
High Availability in Exchange Server 2007
Exchange Server 2007 introduced with the new high availability option both at the server level and the database level. It has 4 High Availability Solution, which uses the continuous replication technology (Asynchronous log file shipping) to have duplicate copy of database at the same server and also at different server. High Availability Solutions in Exchange Server 2007 are
Local Continuous Replication
LCR high availability option can be achieved only at the storage level, Active Copy of the database is replicated to the passive copy of the database using log file shipping, replication of log files will takes inside the server, but to a different disk.
Disadvantage: If the disk goes down, data will be safe at other disk.
Cluster Continuous Replication
Using CCR, High Availability can be achieved at the Server level as well as the storage level, In CCR active copy of the mailbox database will be replicated to passive copy on a Passive node. Using the Windows failover clustering, redundancy will be achieved both at hardware and database level. Using CCR, there is no single point of failure
Disadvantage: No site level redundancy and there is no option to configure a lagged database copy on passive node
Single Copy Cluster
SCC is just like the Exchange 2003 high availability model, where the Single Copy cluster nodes on the cluster shared the same storage device
Disadvantage: We have the high availability option for server and not the database level.
Standby Continuous Replication – SCR is the new high availability model introduced in Exchange 2007 SP1. Data can be replicated to remote site and Site Resiliency can be achieved by shipping log file to another mailbox server in a remote site and there is no need of Windows failover cluster.
Using Standby Continuous Replication, we can have a high available database copy on the remote site for disaster recovery and we have the option to specify lagged log reply time.
High Availability in Exchange Server 2010
Database Availability Group is a group of up to 16 servers that holds the set of replicated Mailbox databases and DAG is the boundary of replication and the failoverswitchover process among those mailbox databases.
High Availability concept of Exchange 2007 like LCR, CCR, SCC and SCR are not available in Exchange Server 2010. Exchange 2010 has a new high availability model name Database Availability Group. In actual LCR and SCC concepts has been completely removed but the concept CCR and SCR are combined and evolved into more unified High Availability Solution knows as Database Availability Group.
Using DAG, We can achieve redundancy at the server level, database level and the data resiliency. Below are the new component in Database Availability Group that manages the failovers
We will discuss the Database Availability Group concepts on the next 5 chapters, this will be shot introduction on the high availability options that are available in the previous version of exchange 2010.
High Availability in Exchange Server 2003
Exchange Server 2003 depends on Windows Clustering for high availability and that too, the High Availability solution or the redundancy will be available only at the hardware level and not at the database level.
Exchange 2003 cluster nodes shared the same storage system. If the active cluster node suddenly became unavailable, the Exchange Virtual Server (EVS) and any relevant cluster resources would fail over to the passive node and the end users could then continue to access their mailbox.
With Windows Server 2003 Enterprise Edition and Exchange Server 2003 Enterprise Edition, We can set up 8 node cluster for redundancy. With Exchange 2003, we have the high availability option at the server level and not at the StoreDatabase level.
High Availability in Exchange Server 2007
Exchange Server 2007 introduced with the new high availability option both at the server level and the database level. It has 4 High Availability Solution, which uses the continuous replication technology (Asynchronous log file shipping) to have duplicate copy of database at the same server and also at different server. High Availability Solutions in Exchange Server 2007 are
Local Continuous Replication
LCR high availability option can be achieved only at the storage level, Active Copy of the database is replicated to the passive copy of the database using log file shipping, replication of log files will takes inside the server, but to a different disk.
Disadvantage: If the disk goes down, data will be safe at other disk.
Cluster Continuous Replication
Using CCR, High Availability can be achieved at the Server level as well as the storage level, In CCR active copy of the mailbox database will be replicated to passive copy on a Passive node. Using the Windows failover clustering, redundancy will be achieved both at hardware and database level. Using CCR, there is no single point of failure
Disadvantage: No site level redundancy and there is no option to configure a lagged database copy on passive node
Single Copy Cluster
SCC is just like the Exchange 2003 high availability model, where the Single Copy cluster nodes on the cluster shared the same storage device
Disadvantage: We have the high availability option for server and not the database level.
Standby Continuous Replication – SCR is the new high availability model introduced in Exchange 2007 SP1. Data can be replicated to remote site and Site Resiliency can be achieved by shipping log file to another mailbox server in a remote site and there is no need of Windows failover cluster.
Using Standby Continuous Replication, we can have a high available database copy on the remote site for disaster recovery and we have the option to specify lagged log reply time.
High Availability in Exchange Server 2010
Database Availability Group is a group of up to 16 servers that holds the set of replicated Mailbox databases and DAG is the boundary of replication and the failoverswitchover process among those mailbox databases.
High Availability concept of Exchange 2007 like LCR, CCR, SCC and SCR are not available in Exchange Server 2010. Exchange 2010 has a new high availability model name Database Availability Group. In actual LCR and SCC concepts has been completely removed but the concept CCR and SCR are combined and evolved into more unified High Availability Solution knows as Database Availability Group.
Using DAG, We can achieve redundancy at the server level, database level and the data resiliency. Below are the new component in Database Availability Group that manages the failovers
- Active manager is the primary component of Database Availability Group
- Exchange 2010 use AM to manage switch over and failovers, instead of the exchange cluster resource DLL and the associated cluster services used in the previous version of exchange
- Active Manager runs on all mailbox servers that are part of DAG member
No comments:
Post a Comment