Re: SQL Server 2000 vs 2005 - Replication, Clustering, Double-Take etc.



It could be. Replication is not normally recommended for HA as it does not
do automatic failover or client redirects - they way clustering does.
Clustering does have a distance limitation.

DoubleTake does work and does work well, but you are correct there is some
failover time.

1) So if your failover server is local use clustering. If its remote I would
look at bi-directional transactional, as this keeps both copies in sync and
failover and failback are merely matters of client redirects.

2) File system drivers do work as part of a HA technology but there is some
time required for failover and failback.

3) With replication you have to redirect the clients to the failover site.
With clustering it automatically fails over from one node to the other and
your clients are automatically redirected.

4) Replication in SQL 2005 is slightly more complicated to set up.

As you mention remote DR sites around North America, clustering won't work
as it has a distance limitation which is normally within the same enclosure
or rack.
--
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html

Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Spam Catcher" <spamhoneypot@xxxxxxxxxx> wrote in message
news:Xns993B9C35E963Cusenethoneypotrogers@xxxxxxxxxxxx
Hi all,

We're in the process of investigating several HA solutions for SQL
Server. We currently run a small database (couple GBs) but needs to be
available for 99.99% of the time. We can tolerate some downtime... but
preferably none.

A coworker suggested we take a look at Double-Take due to the simplicity
of installation. It seems to do what we need (replicate a database and
have failover support) but there are a couple points that worry me:

1. It's done at the file system level
2. There is a delay with failover

So my questions are:

1. Is SQL Server replication the way to go or clustering? Or a 3rd party
product?
2. Are file system mirroring software effective in maintaining a HA
environment - especially for SQL Server?
3. With SQL replciation/database mirroring/clsutering ... if a node
fails, will we have to have handle the failure - or will the transition
to the new node be seemless?
4. I briefly looked into replication in SQL Server 2000 but it looked
like a pain to setup - any better in 2005?

If we go with the built in SQL Server tools any suggestions if we should
use Clustering vs. Replication? We like to maximum resource usage, while
maintaining a reliable environment but would also like to keep costs
down. As I understand... Clustering requires a shared disk - and perhaps
that could be another point of failure?

Eventually we will have remote hot stand by sites around North America
...

Lastly - my boss gets e-mails from companies like Polyserve, etc selling
3rd party replication products. Should we even look into these, or will
the built in tool in SQL Server 2000/2005 be sufficient?


.



Relevant Pages

  • Re: High Availability
    ... also does failover. ... Async replication is real-time, as opposed to the scheduled Log Shipping ... > Note that one of the considerations for the active/passive GEO Cluster is ... > SQL Server 2000 High Availability Series ...
    (microsoft.public.sqlserver.clustering)
  • Re: Replication with failover clustering??
    ... We are having Merge Replication setup with push subscription. ... Microsoft SQL Server clustering is not a good fit for geospatial clustering ... Both clustering and database mirroring do do automatic failover, ...
    (microsoft.public.sqlserver.replication)
  • Re: Reconnecting users after failover
    ... A cluster failover looks exactly ... should write a connection object that encapsulates automatic retry. ... not just preparation for clustering. ... I support the Professional Association for SQL Server ...
    (microsoft.public.sqlserver.clustering)
  • Re: Replication with failover clustering??
    ... Microsoft SQL Server clustering is not a good fit for geospatial clustering ... Both clustering and database mirroring do do automatic failover, ... Replication can have a greater ... Looking for a SQL Server replication book? ...
    (microsoft.public.sqlserver.replication)
  • Re: Load balance
    ... Have a look at peer-to-peer replication. ... My application is based on ASP.NET 2.0 and SQL Server 2005. ... SQL Servers in addition to having failover support in case one SQL Server ...
    (microsoft.public.sqlserver.clustering)