| Author |
Message |
Tony
Guest
|
 Posted:
Tue Nov 08, 2005 5:16 pm Post subject:
Trying to find all the pieces of the storage puzzle |
|
|
Hello,
I'm looking to find out if I've thought of everything, and if what I've
thought if is the way to go. Currently our environment is fairly
departmentaized and the servers are under utilized for the most part, even
on mission critical systems. We have half a dozen servers, one being SCO
and the rest Windows 2000 and 2003. We'll be moving the app that runs on
the SCO box to a Linux box and I'd like to consolidate some of our Windows
servers (we have several SQL servers that for whatever reason are all on
different boxes).
My main concern is keeping the employees up and running without interuption
to their database access, second is performance. I'm planning on
implementing an FC SAN solution and using it for data storage for both the
Linux and Windows systems. I haven't landed on a SAN solution, one reseller
is pricing me up a couple systems from HP, and another will be proposing an
IBM solution.
I'm looking at having redundant FC cards in the servers, as well as
redundant disk trays and SAN switches. We'll be starting with about 1TB of
storage space, achieved by having 7x146GB drives in each storage tray and
having the trays mirrored. I want this setup so that if any part of the
storage system goes down it will not affect the end users, the redundant
hardware is part of this, but I assume I'll also need software to help
control this. I also want to be able to boot from the SAN in the hopes that
if a server goes down I can swap it out and bring a new one online with as
little downtime as possible (I might be fooling myself here based on the
change in hardware and MS's licensing).
So, could you let me know if I'm missing a piece of the puzzle, this is what
I have rolling around in my head:
- Server has two FC cards connected to two different switches, which then
each connect to two storage arrays (small mesh).
- I need a software package to keep the data on both arrays identical.
- Servers use an initiator to boot from the SAN (One for the Windows systems
and one for Linux)
- Server needs an MPIO driver for failover (One for the Windows systems and
one for Linux)
- I need a software package to manage the SAN's "partitions" so that I can
allocate different storage space to different servers, I'd also like to be
able to re-allocate on the fly.
To add to the confusion, since I want redundancy, I would need two Linux
servers and two Windows Servers running SQL. The current Windows servers
are under utilized, so I'm thinking that if I had two physical servers I
could install multiple instances using Virtual Machines. The different
virtual machines would all work off the same database, so that if one went
down, or if the hardware on which some of them were running went down, the
other system would continue to service the users that were attached. Which
adds the following puzzle pieces:
- Some sort of load balancing/failover system for Linux
- Load balancing/failover system for Windows (possibly clustering)
- Virtual Machine software
Now add to this, remote access via Citrix. Is it feasible to have 2-3
physical servers running virtual Windows 2003 servers that have Citrix
installed, as well as our Database app and SQL, or would it be better to
have a virtual server running Citrix and another running the app and SQL?
Have I made this more complicated that it needs to be for a redundant system
that minimizes downtime while maximizing performance and the use of hardware
resources, or have I not even scratched the surface of how complex a system
I need to accomplish the above? In other words, am I missing anything in my
lists above to make all this work, third party software/drives/etc.? Is
there a software package that can handle a bulk of what I want to do, for
example I'll currently checking out the documentation for Veritas Storage
Foundation.
Thanks,
Tony |
|
| Back to top |
|
 |
Eric Bursley [MVP]
Guest
|
| Posted:
Thu Nov 10, 2005 1:17 am Post subject:
Re: Trying to find all the pieces of the storage puzzle |
|
|
There are multiple ways to implement this solution.
I think you are on the right track. I'm not so sure about using a boot from
SAN solution, but if that is what you want to do, it is doable. There are
some precautions to using boot from SAN, such as troubleshooting boot
problems. Especially if your storage vendor requires a special HBA driver
that is not a default OS driver. While you can get around this in Windows
using <f6> during the boot process, it is still just one more step in the
troubleshooting process.
With a cluster, you can install multiple instances of SQL server, so you
would just need to migrate your SQL databases to the new instances in the
cluster.
Also VMWare ESX may be an option for you if you want to maximize server
usage. You can have an ESX server farm to host as many virtual machines as
necessary. Plus provide a nice development machine for any developers in
your organization.
Just as a plug, have you considered using Dell with a Dell EMC SAN solution?
Dell has internal system designers and implementers that can configure a
nice solid solution for you.
Eric Bursley
eric at bursley dot net
Microsoft MVP
RHCE, MCSE, BCFP, EEIE-CS, ESCE-CS
GPG Signature: Key fingerprint = CEAE CF3A 3876 7ECE 9DA7 946F DA9F DDCA
C392 6DCB
"Tony" <tud333@hotmail.com> wrote in message
news:OjCEIUI5FHA.3544@TK2MSFTNGP09.phx.gbl...
| Quote: | Hello,
I'm looking to find out if I've thought of everything, and if what I've
thought if is the way to go. Currently our environment is fairly
departmentaized and the servers are under utilized for the most part, even
on mission critical systems. We have half a dozen servers, one being SCO
and the rest Windows 2000 and 2003. We'll be moving the app that runs on
the SCO box to a Linux box and I'd like to consolidate some of our Windows
servers (we have several SQL servers that for whatever reason are all on
different boxes).
My main concern is keeping the employees up and running without
interuption to their database access, second is performance. I'm planning
on implementing an FC SAN solution and using it for data storage for both
the Linux and Windows systems. I haven't landed on a SAN solution, one
reseller is pricing me up a couple systems from HP, and another will be
proposing an IBM solution.
I'm looking at having redundant FC cards in the servers, as well as
redundant disk trays and SAN switches. We'll be starting with about 1TB
of storage space, achieved by having 7x146GB drives in each storage tray
and having the trays mirrored. I want this setup so that if any part of
the storage system goes down it will not affect the end users, the
redundant hardware is part of this, but I assume I'll also need software
to help control this. I also want to be able to boot from the SAN in the
hopes that if a server goes down I can swap it out and bring a new one
online with as little downtime as possible (I might be fooling myself here
based on the change in hardware and MS's licensing).
So, could you let me know if I'm missing a piece of the puzzle, this is
what I have rolling around in my head:
- Server has two FC cards connected to two different switches, which then
each connect to two storage arrays (small mesh).
- I need a software package to keep the data on both arrays identical.
- Servers use an initiator to boot from the SAN (One for the Windows
systems and one for Linux)
- Server needs an MPIO driver for failover (One for the Windows systems
and one for Linux)
- I need a software package to manage the SAN's "partitions" so that I can
allocate different storage space to different servers, I'd also like to be
able to re-allocate on the fly.
To add to the confusion, since I want redundancy, I would need two Linux
servers and two Windows Servers running SQL. The current Windows servers
are under utilized, so I'm thinking that if I had two physical servers I
could install multiple instances using Virtual Machines. The different
virtual machines would all work off the same database, so that if one went
down, or if the hardware on which some of them were running went down, the
other system would continue to service the users that were attached.
Which adds the following puzzle pieces:
- Some sort of load balancing/failover system for Linux
- Load balancing/failover system for Windows (possibly clustering)
- Virtual Machine software
Now add to this, remote access via Citrix. Is it feasible to have 2-3
physical servers running virtual Windows 2003 servers that have Citrix
installed, as well as our Database app and SQL, or would it be better to
have a virtual server running Citrix and another running the app and SQL?
Have I made this more complicated that it needs to be for a redundant
system that minimizes downtime while maximizing performance and the use of
hardware resources, or have I not even scratched the surface of how
complex a system I need to accomplish the above? In other words, am I
missing anything in my lists above to make all this work, third party
software/drives/etc.? Is there a software package that can handle a bulk
of what I want to do, for example I'll currently checking out the
documentation for Veritas Storage Foundation.
Thanks,
Tony
|
|
|
| Back to top |
|
|
Tony
Guest
|
| Posted:
Thu Nov 10, 2005 5:16 pm Post subject:
Re: Trying to find all the pieces of the storage puzzle |
|
|
Eric Bursley [MVP] wrote...
I'm planning on having one server run the SAN management software and it
will boot of it's own internal disks, but I believe I'll have the others
booting from the SAN. From the sounds of it the VMWare is the way to go.
Right now I'm thinking I'll have two physical servers and on each I'll have
a Linux server, Citrix/App server and SQL database server. Each instance
will be duplicated (clustered) on the other box for redundancy/failover/load
balancing and if I want to add another physical server to boost performance
I can (I believe) replicate the Virtual servers fairly easily onto the new
hardware.
I haven't looked at Dell or EMC, we're mostly an HP shop, but I haven't
locked myself in to anything yet with regards to the SAN, so I'll check it
out. Thanks for the feedback and suggestion.
Tony
| Quote: | There are multiple ways to implement this solution.
I think you are on the right track. I'm not so sure about using a boot
from SAN solution, but if that is what you want to do, it is doable.
There are some precautions to using boot from SAN, such as troubleshooting
boot problems. Especially if your storage vendor requires a special HBA
driver that is not a default OS driver. While you can get around this in
Windows using <f6> during the boot process, it is still just one more step
in the troubleshooting process.
With a cluster, you can install multiple instances of SQL server, so you
would just need to migrate your SQL databases to the new instances in the
cluster.
Also VMWare ESX may be an option for you if you want to maximize server
usage. You can have an ESX server farm to host as many virtual machines
as necessary. Plus provide a nice development machine for any developers
in your organization.
Just as a plug, have you considered using Dell with a Dell EMC SAN
solution? Dell has internal system designers and implementers that can
configure a nice solid solution for you.
Eric Bursley
eric at bursley dot net
Microsoft MVP
RHCE, MCSE, BCFP, EEIE-CS, ESCE-CS
GPG Signature: Key fingerprint = CEAE CF3A 3876 7ECE 9DA7 946F DA9F DDCA
C392 6DCB
"Tony" <tud333@hotmail.com> wrote in message
news:OjCEIUI5FHA.3544@TK2MSFTNGP09.phx.gbl...
Hello,
I'm looking to find out if I've thought of everything, and if what I've
thought if is the way to go. Currently our environment is fairly
departmentaized and the servers are under utilized for the most part,
even on mission critical systems. We have half a dozen servers, one
being SCO and the rest Windows 2000 and 2003. We'll be moving the app
that runs on the SCO box to a Linux box and I'd like to consolidate some
of our Windows servers (we have several SQL servers that for whatever
reason are all on different boxes).
My main concern is keeping the employees up and running without
interuption to their database access, second is performance. I'm
planning on implementing an FC SAN solution and using it for data storage
for both the Linux and Windows systems. I haven't landed on a SAN
solution, one reseller is pricing me up a couple systems from HP, and
another will be proposing an IBM solution.
I'm looking at having redundant FC cards in the servers, as well as
redundant disk trays and SAN switches. We'll be starting with about 1TB
of storage space, achieved by having 7x146GB drives in each storage tray
and having the trays mirrored. I want this setup so that if any part of
the storage system goes down it will not affect the end users, the
redundant hardware is part of this, but I assume I'll also need software
to help control this. I also want to be able to boot from the SAN in the
hopes that if a server goes down I can swap it out and bring a new one
online with as little downtime as possible (I might be fooling myself
here based on the change in hardware and MS's licensing).
So, could you let me know if I'm missing a piece of the puzzle, this is
what I have rolling around in my head:
- Server has two FC cards connected to two different switches, which then
each connect to two storage arrays (small mesh).
- I need a software package to keep the data on both arrays identical.
- Servers use an initiator to boot from the SAN (One for the Windows
systems and one for Linux)
- Server needs an MPIO driver for failover (One for the Windows systems
and one for Linux)
- I need a software package to manage the SAN's "partitions" so that I
can allocate different storage space to different servers, I'd also like
to be able to re-allocate on the fly.
To add to the confusion, since I want redundancy, I would need two Linux
servers and two Windows Servers running SQL. The current Windows servers
are under utilized, so I'm thinking that if I had two physical servers I
could install multiple instances using Virtual Machines. The different
virtual machines would all work off the same database, so that if one
went down, or if the hardware on which some of them were running went
down, the other system would continue to service the users that were
attached. Which adds the following puzzle pieces:
- Some sort of load balancing/failover system for Linux
- Load balancing/failover system for Windows (possibly clustering)
- Virtual Machine software
Now add to this, remote access via Citrix. Is it feasible to have 2-3
physical servers running virtual Windows 2003 servers that have Citrix
installed, as well as our Database app and SQL, or would it be better to
have a virtual server running Citrix and another running the app and SQL?
Have I made this more complicated that it needs to be for a redundant
system that minimizes downtime while maximizing performance and the use
of hardware resources, or have I not even scratched the surface of how
complex a system I need to accomplish the above? In other words, am I
missing anything in my lists above to make all this work, third party
software/drives/etc.? Is there a software package that can handle a bulk
of what I want to do, for example I'll currently checking out the
documentation for Veritas Storage Foundation.
Thanks,
Tony
|
|
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in