replication/hot-copy and auto-fallback

What is the best way to accomplish a fully redudant database so that the
secondary will automatically be used if the the primary fails?
I see these options for the secondary (and other-ary) servers:
(1) Shift the secondary responsibility to the storage device only - use RAID
1 for example. This is good if a drive fails, but not if the server box
fails. I could use NAS RAID drives and have a secondary machine waiting in
the wings to elect itself as the DB machine talking to the same NAS drives,
but I'm not sure how to make this transition. (I could have the RAID with a
CPU box that has hot-swappable drives, fans, power supplies, etc., but I'd
rather have two cheaper, but totally isolated enclosures.)
(2) Use SqlServer snapshot or transaction replication to keep the secondary
updated.
(3) Update both machines myself in app code.
try { ...update primary... } except { ...doh! - set secondary active...}
try { ..update secondary...} except { ...if I'm active, we're hosed,
must be an EMP bomb... }
In either case, I'm not sure how to do the "auto fallover" to the secondary
server on access. I see these options here:
(A) Handle in application code:
try { ...access primary... }
except {
try { ..access secondary...}
except { ...give up...}
}
If I do this, I can always update both machines as well, as in (3) above.
(B) Have Sql server handle automatically?
Of course, I need the primary to be updated again once it comes back
on-line, so I'd have to probably set for merge replication or a seperate
reverse snapshot/transaction log. (Database will be small-ish -- a few tens
of MB.) Other scenerios could take advantage of both servers, by
interleaving reads between the two servers while writing to both, like a
RAID 1 of SQL servers.
Am I missing included services on the DB client-side that handle multiple
servers out of the box?
thanks for any help,
mikeOn Fri, 18 Mar 2005 10:33:18 -0800, Mike wrote:

>What is the best way to accomplish a fully redudant database so that the
>secondary will automatically be used if the the primary fails?
(snip)
Hi Mike,
I don't have any personal experience with it, but based on a quick scan
of the description in Books Online, it looks like failover clustering
will do the job for you.
If you use the Books Online index and search for failover clustering,
you should find all the information.
Good luck!
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)|||Wow, if it was really just an RTFM issue, I expected a lot more responses!
Thanks...
mike
"Hugo Kornelis" <hugo@.pe_NO_rFact.in_SPAM_fo> wrote in message
news:t2pm315o1togf8pr3f03fmppaferbodoor@.
4ax.com...
> On Fri, 18 Mar 2005 10:33:18 -0800, Mike wrote:
>
> (snip)
> Hi Mike,
> I don't have any personal experience with it, but based on a quick scan
> of the description in Books Online, it looks like failover clustering
> will do the job for you.
> If you use the Books Online index and search for failover clustering,
> you should find all the information.
> Good luck!
> Best, Hugo
> --
> (Remove _NO_ and _SPAM_ to get my e-mail address)|||http://www.microsoft.com/windowsser...
ering_4ctw.asp
http://support.microsoft.com/defaul...kb;en-us;260758
http://www.microsoft.com/technet/pr...n/failclus.mspx
--
Mike Epprecht, Microsoft SQL Server MVP
Zurich, Switzerland
IM: mike@.epprecht.net
MVP Program: http://www.microsoft.com/mvp
Blog: http://www.msmvps.com/epprecht/
"Mike" <vimakefile@.yahoo.com> wrote in message
news:ujpgMOELFHA.3404@.TK2MSFTNGP15.phx.gbl...
> Wow, if it was really just an RTFM issue, I expected a lot more responses!
> Thanks...
> mike
>
> "Hugo Kornelis" <hugo@.pe_NO_rFact.in_SPAM_fo> wrote in message
> news:t2pm315o1togf8pr3f03fmppaferbodoor@.
4ax.com...
>