Anyone deployed stretch clustering ?

Ive come across this article
http://www.microsoft.com/technet/pro...y/hasog05.mspx
and was interested in stretch clustering .. and have one node of my cluster
in a geographic dispersed location ..
Has anyone done that and can elaborate more on the storage section of this
doc.. We have EMC SAN in place and curious on the technology of how to make
those 2 SANs appear to be one SAN to the MSCS cluster.. Basically if i setup
a 3 node cluster with 2 passive.. each of the passive nodes being on
different geographical sites and my priority of failover would be is Prodn
Server A fails, then it would failover to the local passive node and if
thats not available ( due to site failure) then move to passive node on the
other site.. Is this possible ? Love to hear more
Hassan,
You would need to first implement a remote mirroring solution with your SAN.
EMC's solution, SRDF (Symmetrix Remote Data Facility), is one of the only
solutions that has been qualified for geographically dispersed Microsoft
clusters. If you implement SRDF, you should also look into EMC's
SRDF/Cluster Enabler for MSCS product (formerly known as GeoSpan) to
automate the failover to the remote site. It's a nice solution and it works
quite well.
If you're considering a geographically dispersed clustering solution, you
might want to take a look at the following KB article for some of the
restrictions/limitations of such a solution:
http://support.microsoft.com/?id=280743
Regards,
John
"Hassan" <fatima_ja@.hotmail.com> wrote in message
news:ezMQejIkEHA.1996@.TK2MSFTNGP09.phx.gbl...
> Ive come across this article
> http://www.microsoft.com/technet/pro...y/hasog05.mspx
> and was interested in stretch clustering .. and have one node of my
cluster
> in a geographic dispersed location ..
> Has anyone done that and can elaborate more on the storage section of this
> doc.. We have EMC SAN in place and curious on the technology of how to
make
> those 2 SANs appear to be one SAN to the MSCS cluster.. Basically if i
setup
> a 3 node cluster with 2 passive.. each of the passive nodes being on
> different geographical sites and my priority of failover would be is Prodn
> Server A fails, then it would failover to the local passive node and if
> thats not available ( due to site failure) then move to passive node on
the
> other site.. Is this possible ? Love to hear more
>
>
|||One challenge u may face with stretch is the fact that the IP Heartbeat has
to be on same subnet and have a guaranteed roundtrip of <500ms as per below
http://www.microsoft.com/technet/pro.../geoclust.mspx
Also with 3 node you will need either Datacenter or Windows 2003
Andy.
"Hassan" <fatima_ja@.hotmail.com> wrote in message
news:ezMQejIkEHA.1996@.TK2MSFTNGP09.phx.gbl...
> Ive come across this article
> http://www.microsoft.com/technet/pro...y/hasog05.mspx
> and was interested in stretch clustering .. and have one node of my
cluster
> in a geographic dispersed location ..
> Has anyone done that and can elaborate more on the storage section of this
> doc.. We have EMC SAN in place and curious on the technology of how to
make
> those 2 SANs appear to be one SAN to the MSCS cluster.. Basically if i
setup
> a 3 node cluster with 2 passive.. each of the passive nodes being on
> different geographical sites and my priority of failover would be is Prodn
> Server A fails, then it would failover to the local passive node and if
> thats not available ( due to site failure) then move to passive node on
the
> other site.. Is this possible ? Love to hear more
>
>
sql

Any issues with SP2 and 2005 x64?

I came across someone who mentioned in passing that there are lots of issues
with SP2 and x64, is that true? Are there any significant differences to be
aware of between SP2 on x32 vs x64?
TIA
Michael MacGregor
Database ArchitectMost of my clients run SP2 on X64 with no problem. I would rather run on SP2
than SP1.
--
Andrew J. Kelly SQL MVP
Solid Quality Mentors
"Michael MacGregor" <macnoknifespam@.noemailspam.com> wrote in message
news:Ot3z3zaFIHA.3548@.TK2MSFTNGP06.phx.gbl...
>I came across someone who mentioned in passing that there are lots of
>issues with SP2 and x64, is that true? Are there any significant
>differences to be aware of between SP2 on x32 vs x64?
> TIA
> Michael MacGregor
> Database Architect
>|||Thanks Andrew.
MTM

ANSI-Style Joins & Old-Style Joins!

I came across an article which states the differences between
ANSI-style JOINs & old-style JOINs. This was one among them:
---
The ANSI-style JOIN supports query constructions which the old-style
JOIN syntax does not support.
---
What does the above mean especially the term "query constructions"?
Thanks,
ArpanOne example is that the old-style syntax only directly supports joins
based on equality. In the ANSI syntax we can write:
SELECT *
FROM A
LEFT OUTER JOIN B
ON A.col1 BETWEEN B.col2 AND B.col3
To do that without the OUTER JOIN operator would require an inner join
and a UNION.
David Portas
SQL Server MVP
--|||The *= syntax was equality only and a table could appear only once as
eitehr preserved or unpreserved. Google aroudn for one of my olds
postings on how the outer works|||On 30 Jul 2005 01:38:32 -0700, Arpan wrote:

>I came across an article which states the differences between
>ANSI-style JOINs & old-style JOINs. This was one among them:
>---
>The ANSI-style JOIN supports query constructions which the old-style
>JOIN syntax does not support.
>---
>What does the above mean especially the term "query constructions"?
>Thanks,
>Arpan
Hi Arpan,
In addition to the answers provided by David and Celko, here are two
more examples:
1. The ANSI-style permits you to write
SELECT something
FROM A
LEFT OUTER JOIN B
ON B.col1 = A.col1
AND B.col2 IS NULL
or
SELECT something
FROM A
LEFT OUTER JOIN B
ON B.col1 = A.col1
WHERE B.col2 IS NULL
And the results of these two are quite different. With the old-style
joins, they'd be the same:
SELECT something
FROM A, B
WHERE B.col1 =* A.col1
AND B.col2 IS NULL
And which of the two possible result sets you'd get depends on how the
programmers decided to implement this. From what I've heard, Oracle's
result set would be different from SQL Server's.
2. FULL OUTER JOIN. This is not available in old-style.
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)