Help! How to clustering SQL Server 2000 without Shared Disk?

hi all,

our company is going to setup ms sql server cluster for fault tolerance,
budget has to be kept minimum.

actaully, i was told that sql clustering (includes shared DB data sync)
can be done over a dedicated tcp/ip connection instead of using shared disk
array connected to the same scsi bus of 2 machines, however i could not even
configure cluster service on win2k server as there is no shared disk.

futhermore, i heard windows 2003 server can do it with no shared disk,
butr it can't be used for this purpose because of license issue.

i would to know if there is a way to cluster sql server nodes without a
shared disk array on windows 2000 adv/dc server + ms sql 2000 enterprise.

all advices are welcomed to reach chun_wail@.hotmail.com

thank you very much,

Will Law (hong kong)"wailaw" <chun_wail@.hotmail.com> wrote in message
news:bi7mdl$lpa1@.imsp212.netvigator.com...
> hi all,
> our company is going to setup ms sql server cluster for fault
tolerance,
> budget has to be kept minimum.
> actaully, i was told that sql clustering (includes shared DB data
sync)
> can be done over a dedicated tcp/ip connection instead of using shared
disk
> array connected to the same scsi bus of 2 machines, however i could not
even
> configure cluster service on win2k server as there is no shared disk.
> futhermore, i heard windows 2003 server can do it with no shared disk,
> butr it can't be used for this purpose because of license issue.
> i would to know if there is a way to cluster sql server nodes without
a
> shared disk array on windows 2000 adv/dc server + ms sql 2000 enterprise.
> all advices are welcomed to reach chun_wail@.hotmail.com
>
> thank you very much,
> Will Law (hong kong)

(Removed irrelevant newsgroups from the reply)

SQL Server clustering depends on the Windows clustering service, so you have
no way to avoid the shared disk requirement - clustering is managed by the
OS, not by SQL Server. However, from what you say, I suspect you may be
thinking of replication, which is a totally different concept. Replication
allows you to keep data in sync between databases on different servers, but
it doesn't provide the failover support that clustering has.

I suggest you read the Books Online overviews on both clustering and
replication and see which suits your needs best. Log shipping might be
another option if a warm standby server is good enough for your purposes.
But if you really need high availability, you can't do it without a decent
budget and resources, so make sure you know exactly what the business
requirements are before you choose a solution.

Simon