Re: Clustering concepts...
From: Rick Collins (rcollins_at_NOSPAMmed.unc.edu)
Date: 09/11/04
- Previous message: Luca Bianchi: "Re: SQL Mail on a Virtual Cluster"
- In reply to: Geoff N. Hiten: "Re: Clustering concepts..."
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 11 Sep 2004 17:30:30 -0400
> Clustering involves a single data store on a 'shared' storage system (SCSI
> Array, SAN, etc. NOT NAS).
Actually, you can cluster to a shared resource on a NAS. We have a NetApp
880 Filer and have been running SQL clusters quite nicely using Snapdrive
over Gig fiber for a couple of years using the procedures outlined at this
NetApp link:
http://www.netapp.com/tech_library/3248.html#2.
"Geoff N. Hiten" <SRDBA@Careerbuilder.com> wrote in message
news:u%23O18vplEHA.3712@TK2MSFTNGP15.phx.gbl...
> Active/Active and Active/Passive are obsolete concepts from SQL 7.0 when
> each SQL instance had a primary owner host node and a partner owner node.
> Under SQL 2000, all nodes are interchangable in a cluster. Some people
> mistakenly use teh terms to refer to single and multiple instance SQL
> clustering.
>
> Clustering involves a single data store on a 'shared' storage system (SCSI
> Array, SAN, etc. NOT NAS). Shared is not exactly correct since the
> cluster
> software arbitrates ownership so that only one host node owns the
> resource(s) at a time. Therefore, you have to have a working MSCS cluster
> in order to create a SQL cluster.
>
> Replication is not a good technique for creating a warm standby server.
> Replication has VERY limited ability to handle schema changes without a
> total resynch. It also does not replicate stored procedures or views.
>
>
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
>
> I support the Professional Association for SQL Server
> www.sqlpass.org
>
> "Karl Albrecht" <karlman@pacbell.net> wrote in message
> news:O8QGJUplEHA.2340@TK2MSFTNGP11.phx.gbl...
>> We have just setup two servers running a copy of SQL 2000 on Windows 2000
>> server. Each of these servers is running vendor software that interrracts
>> with the databases. One server is primary and one is backup. Clients only
>> interract with the vendor process on the servers and not the SQL
>> directly.
>>
>>
>> Some general questions on SQL 2000 high availability options...
>>
>> Does Active/Passive and Active/Active refer to clustering options that
>> require that the servers be clustered on the OS level? Is this the only
>> option that keeps the databases syncronized in real time?
>>
>> Is replication a good method if the failover will be done manually?
>>
>>
>> Thanks!
>>
>> Karl Albrecht
>>
>>
>
>
- Previous message: Luca Bianchi: "Re: SQL Mail on a Virtual Cluster"
- In reply to: Geoff N. Hiten: "Re: Clustering concepts..."
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|