Re: Upgrading physical disks on an Exchange 2003 Clsuter

Tech Tip: Click here to run a free scan for Windows Errors and optimize PC performance



Are you sure you want to do that? Swap RAID 1+0 for RAID 5 that is. The
performance and the user experience will suffer.

Assuming 10K RPM drives and a 2:1 read:write ratio:

For RAID 0+1

Write performance = 85*10/2 or 425 IOPS
Read performace = 85*10 or 850 IOPS
mixed performace (apply the read/write ratio) = 708.33 IOPS

For Raid 5

Write performance = 85* (10-1)/4 = 191.25 IOPS
Read Performance = 85*(10-1) = 765 IOPS
Mixed performance = 573.75 IOPS

I would be very concerned about decreasing overall performance by 25% and
write performance by more than 50% when I'm going to either add more users
or increase mailbox sizes. Take a look at "Optimiing storage for Exchange
Server 2003" before going down this path
http://technet.microsoft.com/en-us/library/bb125019.aspx




"mclay911" <martin.clay@xxxxxxxxxxxxxxxxx> wrote in message
news:1171024199.812550.282580@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Hi

I have a question regarding swapping out (upgrading) physical disks on
an HP SAN which holds Exchange 2003 data.

This system is running MS Exchange 2003 in a 2 node clustered
configuration.

The Exchange database is held on shared storage (HP SAN MSA 500). The
logical drive that holds the Exchange data consists of 10 x 72 GB
Ultra320 SCSI drives which are in a RAID 1+0 configuration. This drive
letter is assigned the letter T: and is a presented as a Data Disk
Resource within the MS Cluster.

I need to upgrade these 72 Gig drives to 10 x 300 Gig drives in a RAID
5 Configuration. There is also an extra disk that is used as an online
spare. This makes 11 drives in total.

I've looked around to try and find an official upgrade path but can't
seem to find anything.
I plan to perform the upgrade as follows but I have concerns regarding
the Data Disk Resource within the Cluster administrator.

Take a full system backup of all Exchange nodes and shared storage to
tape.
Put the cluster service on the 2nd (passive) node into a disabled
state.
Stop the Exchange services using Cluster Administrator on the active
node 1 but make sure that the Data Disk Resource is still running.
Take a flat file copy of the Exchange database folder from the shared
storage (MSA500) and temporarily store it onto the local physical
drives on node 1.
Set all the cluster services on node 1 to disabled so that they won't
try and start on a server reboot.
Shutdown both nodes.
Remove the 11 x 72 Gig drives on the MSA500
Insert the 11 x 300 Gig drives into the MSA500
Power on node 1.
Go into the HP ACU util and configure 10 of the 300 Gig drives into a
RAID 5 config and the 1 additional 300 Gig drive as an online spare
for that array.
Go into MS Disk Administrator and assign that logical drive the letter
T:
Start the cluster service and only start the Data Disk Resource.
Hopefully the service will start and see the new logical drive as it's
been assigned the same drive letter as before.
Copy the flat files from the local disk on node 1 to the T: drive on
the shared storage.
Start all the other cluster services including Exchange.
Once node 1 is online turn on node 2 and configure the cluster service
to start.
Test failover.

As I mentioned above my concern is the original Data Disk Resource
within the Cluster Administrator recognising the new logical drive
even though the drive letter is the same.

Could you confirm whether the above approach will work or do I need to
perform some additional tasks?

Many thanks in advance.

Martin



.



Relevant Pages

  • Re: raid controller
    ... In his article he states four major reasons to discourage RAID; ... Without RAID you could do a traditional backup at 8:00 AM and at 8:30 AM ... But It's asking for 'RAID Controller' ... hard drives with drive/s C on both drives. ...
    (microsoft.public.windowsxp.basics)
  • Re: Disk partitioning question
    ... A hardware RAID Volume may exist without any RAID Drives on it. ... > This single array will show up as Disk 0 in Windows Disk Manager. ...
    (microsoft.public.windows.server.sbs)
  • Re: Need help installing OS6 on Dell PowerEdge 2600 with PERC4/Di RAID
    ... > like a fresh install and porting their application and data to the new system. ... PERC4/Di RAID on the motherboard, Dell DLT tape drive, Dell slimline ... PATA connector usable for two PATA drives. ... After downloading drivers from both SCO and Dell to create the ...
    (comp.unix.sco.misc)
  • How I built a 2.8TB RAID storage array
    ... My 2.8TB RAID 5 array is finally up and running. ... Nine 400GB PATA drives; eight for use, ... Two Highpoint RocketRAID 454 cards. ...
    (comp.os.linux.hardware)
  • Re: Paul and Old Man: Cannot fix RAID5 failure ...
    ... all of the sudden the Intel Matrix Storage ROM showed Rebuild status. ... When booting form RAID, there is some activity while the screen is black, ... Considering 3 SATA drives spining up after powerdown may cause a 12V ... I was considering a parallel WinXP installation on a 4th disk, ...
    (alt.comp.periphs.mainboard.asus)