When you have a two-member
DAG, the system can no distinguish between the different types of
failure. It could be a single server failure, a multiple-server failure
or a site failure. To manage a datacenter switchover for a two-member
DAG that is extended across multiple datacenters you must use Windows
failover cluster management tools.
More information on datacenter switchovers can be found here:
http://technet.microsoft.com/en-us/libr ... 51049.aspx
If you can wait until Exchange Server 2010 Service Pack 1 changes has been made so that DAC supports two-member DAGs in separate datacenters. DAC, or Datacenter Activation Coordination mode is a property setting for a database availability group (DAG). It enables control of the activation behavior of the DAG members preventing split-brain syndrom by not allowing databases to mount automatically even though the DAG has a quorum.
More information on DAC can be found here:
http://technet.microsoft.com/en-us/libr ... 79790.aspx
More information on datacenter switchovers can be found here:
http://technet.microsoft.com/en-us/libr ... 51049.aspx
If you can wait until Exchange Server 2010 Service Pack 1 changes has been made so that DAC supports two-member DAGs in separate datacenters. DAC, or Datacenter Activation Coordination mode is a property setting for a database availability group (DAG). It enables control of the activation behavior of the DAG members preventing split-brain syndrom by not allowing databases to mount automatically even though the DAG has a quorum.
More information on DAC can be found here:
http://technet.microsoft.com/en-us/libr ... 79790.aspx
Datacentre Activation Mode is a mode specifically for multsite Data
Availability Groups with 3 or more members. It is there to stop
datacentre DAG split brain syndrome.
For example you have 2 Sites with DAG members in each. Catastrophic errors occur and your WAN linking the two sites goes down. You now have a potential scenario where both sites will make a passive database active. The site that already had the active database, and the second site seeing the loss of the active database and automatically activating a copy.
so DAC operates this using literally a bit that it flips from either 0 or 1. 0 meaning it cannot mount a database and upon talking to other DAG members using DACP and finding another server with 1, will mount the databases as it knows it is allowed to.
For a more detailed explanation see here:
http://technet.microsoft.com/en-us/library/dd979790.aspx
DAC must be enabled manually if you are in a split datacentre scenario with 3 or more members. To see the cmdlet to activate again see:
http://technet.microsoft.com/en-us/library/dd979790.aspx
For example you have 2 Sites with DAG members in each. Catastrophic errors occur and your WAN linking the two sites goes down. You now have a potential scenario where both sites will make a passive database active. The site that already had the active database, and the second site seeing the loss of the active database and automatically activating a copy.
so DAC operates this using literally a bit that it flips from either 0 or 1. 0 meaning it cannot mount a database and upon talking to other DAG members using DACP and finding another server with 1, will mount the databases as it knows it is allowed to.
For a more detailed explanation see here:
http://technet.microsoft.com/en-us/library/dd979790.aspx
DAC must be enabled manually if you are in a split datacentre scenario with 3 or more members. To see the cmdlet to activate again see:
http://technet.microsoft.com/en-us/library/dd979790.aspx
No comments:
Post a Comment