Windows Vista Tips

Windows Vista Tips > Newsgroups > Windows Server > Clustering > Request for information

Reply
Thread Tools Display Modes

Request for information

 
 
Roger
Guest
Posts: n/a

 
      01-19-2009
I have been asked to provide to my manager a document outlining what is
required to create failover clusters for our application servers using
Windows 2008.

In particular:

A web-server running multiple sites under IIS using dotnet (32 bit
currently)
A mail server
An SQL server

The purpose would be to provide fail-over redundancy to our systems so that
if the server in question failed, it's partner would take over. In our wish
list, this would all happen automatically and applications would just work.
Of course web DNS is done by IP, so whether that is actually possible or not
I do not know...

Can you point me to documents that outline the basics of the type of
hardware needed.

For example, I know that under win 2008 we require identical servers using
hardware components that are Certified for Windows server 2008. I know that
we need some type of SAN device. Beyond that I don't know how it all goes
together.

For example, since we are looking at 3 clusters of 2 nodes each, can they
share a SAN device between the 3 clusters (if it meets certain criteria...
and what would that be?) Is there a performance hit if the SAN uses Gigabit
Eth. or ISCSI (never used ISCSI not familiar with it).

I am not even sure what questions to ask at this point... so if there are
documents that explain this... enought that I can outline what type of
hardware (not brands but capabilities) I'd appreciate it.



 
Reply With Quote
 
 
 
 
Meinolf Weber [MVP-DS]
Guest
Posts: n/a

 
      01-19-2009
Hello Roger,

Do nor forget the versions from IIS, SQL and Exchange.

Best regards

Meinolf Weber
Disclaimer: This posting is provided "AS IS" with no warranties, and confers
no rights.
** Please do NOT email, only reply to Newsgroups
** HELP us help YOU!!! http://www.blakjak.demon.co.uk/mul_crss.htm


> I have been asked to provide to my manager a document outlining what
> is required to create failover clusters for our application servers
> using Windows 2008.
>
> In particular:
>
> A web-server running multiple sites under IIS using dotnet (32 bit
> currently)
> A mail server
> An SQL server
> The purpose would be to provide fail-over redundancy to our systems so
> that if the server in question failed, it's partner would take over.
> In our wish list, this would all happen automatically and applications
> would just work. Of course web DNS is done by IP, so whether that is
> actually possible or not I do not know...
>
> Can you point me to documents that outline the basics of the type of
> hardware needed.
>
> For example, I know that under win 2008 we require identical servers
> using hardware components that are Certified for Windows server 2008.
> I know that we need some type of SAN device. Beyond that I don't know
> how it all goes together.
>
> For example, since we are looking at 3 clusters of 2 nodes each, can
> they share a SAN device between the 3 clusters (if it meets certain
> criteria... and what would that be?) Is there a performance hit if the
> SAN uses Gigabit Eth. or ISCSI (never used ISCSI not familiar with
> it).
>
> I am not even sure what questions to ask at this point... so if there
> are documents that explain this... enought that I can outline what
> type of hardware (not brands but capabilities) I'd appreciate it.
>



 
Reply With Quote
 
Ryan Sokolowski [MVP]
Guest
Posts: n/a

 
      01-19-2009
I'd strongly suggest that you consider working with a local partner,
preferably one closely associated with Microsoft. It sounds like you need
and want expert input and advice and this is the surest way to get exactly
what you need.

Hope this helps,
--Ryan

--
Ryan Sokolowski
MVP - Clustering
MCT, MCITP x3, MCTS x8, MCSE x2, CCNA, CCDA, BCFP


"Roger" <> wrote in message
news:...
>I have been asked to provide to my manager a document outlining what is
>required to create failover clusters for our application servers using
>Windows 2008.
>
> In particular:
>
> A web-server running multiple sites under IIS using dotnet (32 bit
> currently)
> A mail server
> An SQL server
>
> The purpose would be to provide fail-over redundancy to our systems so
> that if the server in question failed, it's partner would take over. In
> our wish list, this would all happen automatically and applications would
> just work. Of course web DNS is done by IP, so whether that is actually
> possible or not I do not know...
>
> Can you point me to documents that outline the basics of the type of
> hardware needed.
>
> For example, I know that under win 2008 we require identical servers using
> hardware components that are Certified for Windows server 2008. I know
> that we need some type of SAN device. Beyond that I don't know how it all
> goes together.
>
> For example, since we are looking at 3 clusters of 2 nodes each, can they
> share a SAN device between the 3 clusters (if it meets certain criteria...
> and what would that be?) Is there a performance hit if the SAN uses
> Gigabit Eth. or ISCSI (never used ISCSI not familiar with it).
>
> I am not even sure what questions to ask at this point... so if there are
> documents that explain this... enought that I can outline what type of
> hardware (not brands but capabilities) I'd appreciate it.
>
>
>


 
Reply With Quote
 
Roger
Guest
Posts: n/a

 
      01-19-2009
I do not know which version of IIS runs on Windows 2008 (we don't have a
copy yet).

SQL server Assume 2005.

We do not use Exchange we use a product called MailEnable which has few of
the advanced features of Exchange, but we do not need them for our company.

"Meinolf Weber [MVP-DS]" <meiweb(nospam)@gmx.de> wrote in message
news: .com...
> Hello Roger,
>
> Do nor forget the versions from IIS, SQL and Exchange.
>
> Best regards
>
> Meinolf Weber
> Disclaimer: This posting is provided "AS IS" with no warranties, and
> confers no rights.
> ** Please do NOT email, only reply to Newsgroups
> ** HELP us help YOU!!! http://www.blakjak.demon.co.uk/mul_crss.htm
>
>> I have been asked to provide to my manager a document outlining what
>> is required to create failover clusters for our application servers
>> using Windows 2008.
>>
>> In particular:
>>
>> A web-server running multiple sites under IIS using dotnet (32 bit
>> currently)
>> A mail server
>> An SQL server
>> The purpose would be to provide fail-over redundancy to our systems so
>> that if the server in question failed, it's partner would take over.
>> In our wish list, this would all happen automatically and applications
>> would just work. Of course web DNS is done by IP, so whether that is
>> actually possible or not I do not know...
>>
>> Can you point me to documents that outline the basics of the type of
>> hardware needed.
>>
>> For example, I know that under win 2008 we require identical servers
>> using hardware components that are Certified for Windows server 2008.
>> I know that we need some type of SAN device. Beyond that I don't know
>> how it all goes together.
>>
>> For example, since we are looking at 3 clusters of 2 nodes each, can
>> they share a SAN device between the 3 clusters (if it meets certain
>> criteria... and what would that be?) Is there a performance hit if the
>> SAN uses Gigabit Eth. or ISCSI (never used ISCSI not familiar with
>> it).
>>
>> I am not even sure what questions to ask at this point... so if there
>> are documents that explain this... enought that I can outline what
>> type of hardware (not brands but capabilities) I'd appreciate it.
>>

>
>



 
Reply With Quote
 
Robert Smit
Guest
Posts: n/a

 
      01-20-2009
Hi Roger,

Working with a partner is the best solution. but the basics are here.

Remember What is the Load on the applications, Developers are always saying
it runs on my laptop so no issues here, until it is production.

The IIS section I would do this in a web farm there are good papers on this.
But do this in a team with your web site developers not all sites are best
designed for NLB.
Two servers with 2 nic's ( LAN /NLB ) for the IIS , with a DFS for the site
so you need only to update the website in one location.
Two node Cluster for your mail and SQL. IF this is supported with
mailenable.
I think ISCSI is the most flexible solution, make a storage lan, public and
private lan all in different IP subnets !! use separate storage switches and
lan switches.

But check the sizing what is the primary load on this site the CPU/ Memory
is not the main issue here. The under dog is DISK IOPS !!
you can have a 64 core CPU 256 Gb memory but if you have a disk IO from 5000
and you have only a disk set from 10 disk the performance is really bad but
with 5000/130=40 disks it rocks.
So It is not only the hardware and software a cluster design /
infrastructure design is more the a pile of boxes but use x64 where you can
!

In case from the Email product --- not Exchange ;-(

On the mailenable site are nice items about clustering.

http://www.mailenable.com/kb/Content...sp?ID=ME020023

If You must do IIS 6 here is some nice info
http://www.microsoft.com/technet/pro....mspx?mfr=true

Clustering IIS 6.0 using Microsoft? cluster technology on HP BladeSystem
servers and HP StorageWorks EVA5000 SANs
http://h21007.www2.hp.com/portal/sit...10275d6e10RCRD

TechNet Support WebCast: How to cluster Microsoft SQL Server 2005 by using
Microsoft Virtual Server
http://support.microsoft.com/kb/891798


How to: Create a New SQL Server Failover Cluster (Setup)
http://msdn.microsoft.com/en-us/library/ms179530.aspx

How to Cluster SQL Server 2005
http://www.sql-server-performance.co...r_2005_p1.aspx

windows 2008 Info
http://www.microsoft.com/windowsserv...ilability.aspx



So this is in basic my thing , if you need more info you can always ping me.


--
Greetings,

Robert Smit
Blog : http://fiberman.spaces.live.com/


"Roger" <> schreef in bericht
news:...
>I have been asked to provide to my manager a document outlining what is
>required to create failover clusters for our application servers using
>Windows 2008.
>
> In particular:
>
> A web-server running multiple sites under IIS using dotnet (32 bit
> currently)
> A mail server
> An SQL server
>
> The purpose would be to provide fail-over redundancy to our systems so
> that if the server in question failed, it's partner would take over. In
> our wish list, this would all happen automatically and applications would
> just work. Of course web DNS is done by IP, so whether that is actually
> possible or not I do not know...
>
> Can you point me to documents that outline the basics of the type of
> hardware needed.
>
> For example, I know that under win 2008 we require identical servers using
> hardware components that are Certified for Windows server 2008. I know
> that we need some type of SAN device. Beyond that I don't know how it all
> goes together.
>
> For example, since we are looking at 3 clusters of 2 nodes each, can they
> share a SAN device between the 3 clusters (if it meets certain criteria...
> and what would that be?) Is there a performance hit if the SAN uses
> Gigabit Eth. or ISCSI (never used ISCSI not familiar with it).
>
> I am not even sure what questions to ask at this point... so if there are
> documents that explain this... enought that I can outline what type of
> hardware (not brands but capabilities) I'd appreciate it.
>
>
>


 
Reply With Quote
 
Roger
Guest
Posts: n/a

 
      01-20-2009
Thanks for that!

You have given me lots to read and think about!



"Robert Smit" <Robert.smit-nospam-aca-computers.nl> wrote in message
news:...
> Hi Roger,
>
> Working with a partner is the best solution. but the basics are here.
>
> Remember What is the Load on the applications, Developers are always
> saying it runs on my laptop so no issues here, until it is production.
>
> The IIS section I would do this in a web farm there are good papers on
> this. But do this in a team with your web site developers not all sites
> are best designed for NLB.
> Two servers with 2 nic's ( LAN /NLB ) for the IIS , with a DFS for the
> site so you need only to update the website in one location.
> Two node Cluster for your mail and SQL. IF this is supported with
> mailenable.
> I think ISCSI is the most flexible solution, make a storage lan, public
> and private lan all in different IP subnets !! use separate storage
> switches and lan switches.
>
> But check the sizing what is the primary load on this site the CPU/ Memory
> is not the main issue here. The under dog is DISK IOPS !!
> you can have a 64 core CPU 256 Gb memory but if you have a disk IO from
> 5000 and you have only a disk set from 10 disk the performance is really
> bad but with 5000/130=40 disks it rocks.
> So It is not only the hardware and software a cluster design /
> infrastructure design is more the a pile of boxes but use x64 where you
> can !
>
> In case from the Email product --- not Exchange ;-(
>
> On the mailenable site are nice items about clustering.
>
> http://www.mailenable.com/kb/Content...sp?ID=ME020023
>
> If You must do IIS 6 here is some nice info
> http://www.microsoft.com/technet/pro....mspx?mfr=true
>
> Clustering IIS 6.0 using Microsoft? cluster technology on HP BladeSystem
> servers and HP StorageWorks EVA5000 SANs
> http://h21007.www2.hp.com/portal/sit...10275d6e10RCRD
>
> TechNet Support WebCast: How to cluster Microsoft SQL Server 2005 by using
> Microsoft Virtual Server
> http://support.microsoft.com/kb/891798
>
>
> How to: Create a New SQL Server Failover Cluster (Setup)
> http://msdn.microsoft.com/en-us/library/ms179530.aspx
>
> How to Cluster SQL Server 2005
> http://www.sql-server-performance.co...r_2005_p1.aspx
>
> windows 2008 Info
> http://www.microsoft.com/windowsserv...ilability.aspx
>
>
>
> So this is in basic my thing , if you need more info you can always ping
> me.
>
>
> --
> Greetings,
>
> Robert Smit
> Blog : http://fiberman.spaces.live.com/
>
>
> "Roger" <> schreef in bericht
> news:...
>>I have been asked to provide to my manager a document outlining what is
>>required to create failover clusters for our application servers using
>>Windows 2008.
>>
>> In particular:
>>
>> A web-server running multiple sites under IIS using dotnet (32 bit
>> currently)
>> A mail server
>> An SQL server
>>
>> The purpose would be to provide fail-over redundancy to our systems so
>> that if the server in question failed, it's partner would take over. In
>> our wish list, this would all happen automatically and applications would
>> just work. Of course web DNS is done by IP, so whether that is actually
>> possible or not I do not know...
>>
>> Can you point me to documents that outline the basics of the type of
>> hardware needed.
>>
>> For example, I know that under win 2008 we require identical servers
>> using hardware components that are Certified for Windows server 2008. I
>> know that we need some type of SAN device. Beyond that I don't know how
>> it all goes together.
>>
>> For example, since we are looking at 3 clusters of 2 nodes each, can they
>> share a SAN device between the 3 clusters (if it meets certain
>> criteria... and what would that be?) Is there a performance hit if the
>> SAN uses Gigabit Eth. or ISCSI (never used ISCSI not familiar with it).
>>
>> I am not even sure what questions to ask at this point... so if there are
>> documents that explain this... enought that I can outline what type of
>> hardware (not brands but capabilities) I'd appreciate it.
>>
>>
>>

>



 
Reply With Quote
 
Russ Kaufmann
Guest
Posts: n/a

 
      01-21-2009
"Robert Smit" <Robert.smit-nospam-aca-computers.nl> wrote in message
news:...
> The IIS section I would do this in a web farm there are good papers on
> this. But do this in a team with your web site developers not all sites
> are best designed for NLB.
> Two servers with 2 nic's ( LAN /NLB ) for the IIS , with a DFS for the
> site so you need only to update the website in one location.


NLB with Unicast no longer requires two NICs. You can get by fine with a
single NIC. Also, you don't need DFS to make it work. The key is keeping all
NLB'd servers consistent with the same data, etc. You can also use a
clustered file server for the data if you really want to go nuts.

> Two node Cluster for your mail and SQL. IF this is supported with
> mailenable.


It is never a good idea to host two heavy workloads on the same cluster as
in a failure, you must support both on the same physical node and it can
result in extremely poor performance.

> I think ISCSI is the most flexible solution, make a storage lan, public
> and private lan all in different IP subnets !! use separate storage
> switches and lan switches.


iSCSI and Fiber are your real choices when it comes to enterprise level
support. Expect to pay a pretty penny.

Good luck, and have fun.


--
Russ Kaufmann,
MVP, MCSE: Messaging and Security, MCT, MCITP, MCTS and other stuff

ClusterHelp.com, a Microsoft Certified Gold Partner
Web http://www.clusterhelp.com
Blog http://msmvps.com/clusterhelp

 
Reply With Quote
 
 
 
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
information form request form request Windows Media Center 0 05-01-2009 02:44 PM
Certificate Request Processor - The request contains no cert templ BrianS Windows Server 3 05-10-2007 05:42 PM
Encryption information request Eric Logsdon Server Security 2 09-28-2006 01:43 PM
Request for content frustration information Bobbie Harder \(MSFT\) Update Services 6 10-19-2005 12:23 AM
Request for information on Windows Media Author 4.0 Richard Brooks Windows Media Center 0 03-02-2005 06:24 PM



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59