eo
MIS
- Apr 3, 2003
- 809
Hi,
I know this is like saying how long is a piece of string, but I need to advise (just as a starting point) for the provision of SQL Server 2008 r2 clustering/ HA solution to the business.
There are three geographical locations. The topology broadly:
• The client side
• Data centre A – Here two production servers reside. Data stored on a SAN.
• Data centre B – Here one Dev/ Test server reside. Data stored on a SAN.
o SAN (Data centre A) and SAN (Data centre B) connected via Dark Fibre
o 10gb LAN between Data Centres A and B
As the locations are geographically dispersed we are obviously looking at “Multi-site failover cluster”, but that’s about where my knowledge stops. Good business sense obviously means limited financial resources and the need to reuse hardware in a clever manner, BUT as the number of applications that will come on-line will increase dramatically the solution needs to be as scalable as possible. Given the brief above, can you please advise the most optimal solution?
In my limited level of experience in this field, would the following topology be suitable?
Data centre A –
1) Production Server (i) for Application (i)
2) Production Server (ii) for Application (ii)
3) Data stored on SAN in the Data centre (SAN A)
Data centre B –
1) Server (iii) for Dev/ Test
2) Server (iv) which will act as a failover cluster node “+1”, which will serve as a failover cluster for Servers (i) to (iii) (the “+1” is therefore left empty in case of failover
3) Data stored on SAN in the Data centre (SAN B)
Considerations:
In the “+1” server, in the case of heterogeneous software configuration on each primary node (i, ii and iii), the “+1” node must be universally capable of assuming any of the roles of the primary nodes it is responsible for.
Storage data replication from SAN A to SAN B for Application (i and ii) in case they failover and node “+1” kicks in) - databases and their logs are synchronously and consistently replicated in order to avoid data loss
I read somewhere that a VLAN is required between the two geographical locations to provide a single subnet across all nodes. Excuse my ignorance, but is this much different from LAN?
In terms of scalability, addition active nodes can be added to either of the two data centres, so too additional failover clusters “+2”, “+3” etc as required.
Can you guys with the expertise give me your thoughts on my ideas above please. I am obviously not looking for a detailed design, but I must come up with a high level design that will give the business what they want, and importantly be scalable.
Regards
EO
Hertfordshire, England
I know this is like saying how long is a piece of string, but I need to advise (just as a starting point) for the provision of SQL Server 2008 r2 clustering/ HA solution to the business.
There are three geographical locations. The topology broadly:
• The client side
• Data centre A – Here two production servers reside. Data stored on a SAN.
• Data centre B – Here one Dev/ Test server reside. Data stored on a SAN.
o SAN (Data centre A) and SAN (Data centre B) connected via Dark Fibre
o 10gb LAN between Data Centres A and B
As the locations are geographically dispersed we are obviously looking at “Multi-site failover cluster”, but that’s about where my knowledge stops. Good business sense obviously means limited financial resources and the need to reuse hardware in a clever manner, BUT as the number of applications that will come on-line will increase dramatically the solution needs to be as scalable as possible. Given the brief above, can you please advise the most optimal solution?
In my limited level of experience in this field, would the following topology be suitable?
Data centre A –
1) Production Server (i) for Application (i)
2) Production Server (ii) for Application (ii)
3) Data stored on SAN in the Data centre (SAN A)
Data centre B –
1) Server (iii) for Dev/ Test
2) Server (iv) which will act as a failover cluster node “+1”, which will serve as a failover cluster for Servers (i) to (iii) (the “+1” is therefore left empty in case of failover
3) Data stored on SAN in the Data centre (SAN B)
Considerations:
In the “+1” server, in the case of heterogeneous software configuration on each primary node (i, ii and iii), the “+1” node must be universally capable of assuming any of the roles of the primary nodes it is responsible for.
Storage data replication from SAN A to SAN B for Application (i and ii) in case they failover and node “+1” kicks in) - databases and their logs are synchronously and consistently replicated in order to avoid data loss
I read somewhere that a VLAN is required between the two geographical locations to provide a single subnet across all nodes. Excuse my ignorance, but is this much different from LAN?
In terms of scalability, addition active nodes can be added to either of the two data centres, so too additional failover clusters “+2”, “+3” etc as required.
Can you guys with the expertise give me your thoughts on my ideas above please. I am obviously not looking for a detailed design, but I must come up with a high level design that will give the business what they want, and importantly be scalable.
Regards
EO
Hertfordshire, England