0

I have SQL Server 2012 Web Edition. I want to configure High Availability on this Server. Do I need to have another SQL Server to configure High availability?

How is SQL Server Native Client 11.0 is part of this?

I know this is a wide area question, but I am looking for a start on this.

HaBo
  • 13,999
  • 36
  • 114
  • 206

1 Answers1

2

Yes, for all high availability scenarios you need more than one server. As for high availability in the web edition, all you get is log shipping, so you will need to manually fail over to another server if your primary dies.

The native client is a combined ODBC driver and OLE DB Provider for SQL server. It doesn't in itself provide high availability, but it can be used, for example if you were using database mirroring, to provide the name of the mirror server for automatic failover.

steoleary
  • 8,968
  • 2
  • 33
  • 47
  • Is this the same for enterprise edition? I need two servers to handle fail overs? Using Native client if i use mirroring for automatic fail-over, can i stick to one physical server? – HaBo Mar 18 '13 at 13:21
  • Yes, for true high availability you need separate servers, this protects against failures at the OS or hardware level, a single box is always a single point of failure. – steoleary Mar 18 '13 at 18:34
  • Plus, mirroring wont let you mirror to the same instance, you need at least another instance on the same box, but that doesn't give you high availability as if the OS dies or the power gives out to that box, you have no database. – steoleary Mar 18 '13 at 18:36
  • Thank you for all the information you are sharing. All clouds are clearing, one last question. What is clustered(installation) fail-over configuration with Native Client on Enterprise edition? Is this same as we discussed so far? – HaBo Mar 18 '13 at 21:12
  • Clustering is a bit different, you can make a 2 node cluster with standard but if you want more you need enterprise. Clustering is where you have 2 or more machines with sql server installed that share resources such as disks, when one server fails, the resources move across to the other node and the sql services start up there, taking over the workload. – steoleary Mar 18 '13 at 21:23
  • In a cluster, you don't need to do anything special in the native client as the server name stays the same (the name and IP is a resource that can be transferred between nodes) the connection gets dropped briefly as the services restart, but then as the clients reconnect, service is resumed automatically. – steoleary Mar 18 '13 at 21:26
  • Thank you very much this is a very helpful information for me. – HaBo Mar 19 '13 at 01:29