Hot Standby Server w/out Clustering?

I want to have a hot standby SQL server for my clients without clustering. My
idea is to install SQL on two separate machines, SRVA and SRVB, and configure
continuous merge replication between them. Client machines would connect to an
alias name, SRVC, with DNS for SRVC resolving to the IP address of SRVA. Upon
failure, DNS would be changed to point to SRVB. Could something like this
work?
Maybe a shorther question would be, is it possible for clients to connect to a
SQL server alias name (not a named instance) instead of the actual machine
name of the server?
--Eric RobinsonWhat it sounds like you need is log shipping, not merge
replication. Replication is not the best of solutions for
a standby solution, and if you use replication for some
form of availability, it's usually transactional.
Replication is good if you have some specific data to make
available, but you're limited by rowsize and such, but as
for a complete copy of your DB, replication is not the
way. Log shipping is the standard SQL method of making a
warm standby.
You can set up NLB to do what you want to alias things,
and this is all covered in the book SQL Server 2000 HIgh
Availability on how to set it up using it as a "switch"
mechanism - not load balanced - for SQL Server. There are
even configuration instructions on the CD-ROM.
>--Original Message--
>I want to have a hot standby SQL server for my clients
without clustering. My
>idea is to install SQL on two separate machines, SRVA and
SRVB, and configure
>continuous merge replication between them. Client
machines would connect to an
>alias name, SRVC, with DNS for SRVC resolving to the IP
address of SRVA. Upon
>failure, DNS would be changed to point to SRVB. Could
something like this
>work?
>Maybe a shorther question would be, is it possible for
clients to connect to a
>SQL server alias name (not a named instance) instead of
the actual machine
>name of the server?
>--Eric Robinson
>.
>|||To have standby , it is better to use Log shipping. U can
even write scripts to take backup and mover to otherserver
and restore ,
Replication is only for data propagation not for Standby.
In case of Replication schema changes won't be applied. Of
course Merge is not useful u need Transactioal replication.
>--Original Message--
>I want to have a hot standby SQL server for my clients
without clustering. My
>idea is to install SQL on two separate machines, SRVA and
SRVB, and configure
>continuous merge replication between them. Client
machines would connect to an
>alias name, SRVC, with DNS for SRVC resolving to the IP
address of SRVA. Upon
>failure, DNS would be changed to point to SRVB. Could
something like this
>work?
>Maybe a shorther question would be, is it possible for
clients to connect to a
>SQL server alias name (not a named instance) instead of
the actual machine
>name of the server?
>--Eric Robinson
>.
>