Should the SQL Server Service resource be set to "affect the group"?

The SQL resource group has a failover threshold and period, and the SQL
Server resource also has a threshold and period - can someone explain the
different meanings of these settings at 2 different levels.

When initiating failure of the SQL Server resource, what exactly is
happening - just the restarting of the sql server service?

Is the threshold count maintained in the Quorum?

TIA.

Hopefully this'll help: http://support.microsoft.com/default.aspx/kb/197047
Paul Ibison

Paul's link covered the first question nicely.

The group-level counters count actual failovers, not necessarily individual
resource failures.

On failure, if an individual resource cannot be brought back online on the
original node, all resources are offlined and moved to another node. They
are then brought back online in dependency order (lowest first). From a SQL
client application perspective, this appears just like a SQL Service
restart.

The Quorum is a location that maintains persistant cluster state and
prevents split brain conditions. The "cluster" abstraction maintains the
actual counts. The "cluster" is a consensus of the various cluster services
working together on all nodes.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP




"JB" <JB (AT) discussions (DOT) microsoft.com> wrote