Re: Review of new design - Please



Al,
Thank you very much for your response. This does answer the questions I was
having regarding the physical disk requirements.
Thanks again,
Ron


"Al Mulnick" wrote:

> Easiest first:
>
> "Do I need separate physical disks for each of the above sets?" For some
> yes, for some no. There's a trade-off for each. Read on..
>
> "Is an array necessary for the SMTP que?" - No, but it will provide some
> performance boosts in some situations. RAID 1 would be fine in many
> situations. Your situation may vary, but it's likely it would be fine.
>
> The reason to separate them as a best practice is because of I/O types. I'm
> sure you're familiar with that concept already, but you can generally get
> better perforamance and space with RAID 1+0 than you could with just RAID 1.
> RAID 5 typically incurs a write penalty to update parity across the drives,
> so some disk calculations will have you looking at it like this:
> RAID 0 = write penalty (WP) = 1
> RAID 1 = WP = 1
> RAID 1+0 = WP = 1
> RAID 5 = WP = 4
>
> SMTP traffic is typically write heavy in nature with 7:1 w/r not uncommon.
> Could be higher. That's going to be determined by message size, frequency,
> and volume in most cases. The larger the message the higher the write
> incurred because Exchange has made the decision to follow the RFC by
> accepting the message before processing it. You accept every message before
> doing anything with it in a default configuration. Bottom line? If you
> don't have that much SMTP traffic you may be able to put those queues
> elsewhere (not on the log file drives though).
>
> "Why would you mirror the page file?" Why would you separate it?
> Performance mostly. I think you could leave it on the OS drive and likely
> be fine according to the chart you have listed.
>
> "If I have more than one storage group do my trans logs have to be on
> different physcial drives that are mirrored?"
> It's all about trade-offs. If you have more than one SG, you'll have more
> than one set of logs. One set for each SG to be exact. Since log activity
> is serial write in nature, you may want to separate these to avoid
> performance issues related to disk latencies. Then again, some SAN devices
> can mask this along with buffers in the application etc. Typically, it is
> recommended to separate these to their own disks (phys) to alleviate
> bottlenecks related to disk latencies and to allow for better recovery if
> you lose the store or the binary drives.
>
>
> Was it me and I wasn't going to consider the cluster considerations I might
> lay it out like this:
>
> > C:\OS, Exchange - Raid 1 - direct attached storage >> Keep
> > D:\Page File - Raid 1 - direct attached storage >> Move to OS
> > E:\SMTP and MTA queues - Raid 0+1 - SAN >>D:\RAID1 DAS (note the cluster
> > config may preclude this but I would still leave it RAID 1 if on the san)
> > F:\log files for Group1 - Raid 1 - SAN >> Keep
> > G:\DB for Group1 - Raid 0+1 - SAN >> Keep
> > H:\log files for Group2 - Raid 1 - SAN >> Keep
> > I:\DB for Group2 - Raid 0+1 - SAN >> Keep
>
>
> Does that help?
>
> "rdw24" <rdw24@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
> news:13CFDFC3-60F8-4ED9-BB3A-1E4B41047988@xxxxxxxxxxxxxxxx
> > Have been pulling my hair out trying to design a new Exchange enviroment
> > based on best practices from Microsoft. This is enough to drive someone
> > nuts!
> > We currently utilize pop3 mail thu a Linux box with about 500 users. Will
> > setup and install an Exchange enviroment with the mailbox count for users
> > moving to approximately 1400.
> > Overall plan is to setup an FE/BE config utilizing ISA 2004 out front w/
> > FE
> > residing in internal network. Have already purchased ISA and FE hardware
> > and
> > am now grappling with best setup for BE. My current idea is to set up an
> > Active/Passive Cluster with a small SAN using the HP packaged cluster
> > which
> > is (2) DL380 with a MSA1000 Fiber storage controller.
> > Here is where I get confused. Microsoft best practice recommends: ( Raid
> > 1
> > for OS)-(Raid 1 for page file????)-(Raid 1+0 for SMTP queues)-(Raid 1 for
> > log
> > files)-(Raid 1+0 for DB). Questions are as follows:
> >
> > Why would you mirror the page file?
> > Do I need separate physical disks for each of the above sets?
> > Is an array neccesary for the SMTP que?
> > If I have more than one storage group do my trans logs have to be on
> > different physcial drives that are mirrored?
> >
> > According to MS Optimizing Storage doc my layout would be as follows:
> > C:\OS, Exchange - Raid 1 - direct attached storage
> > D:\Page File - Raid 1 - direct attached storage
> > E:\SMTP and MTA queues - Raid 0+1 - SAN
> > F:\log files for Group1 - Raid 1 - SAN
> > G:\DB for Group1 - Raid 0+1 - SAN
> > H:\log files for Group2 - Raid 1 - SAN
> > I:\DB for Group2 - Raid 0+1 - SAN
> > Thats 4 physical drives in each server and a minimum of 16 drives in the
> > SAN
> > - I only have 14 bay available in the SAN.
> > I would greatly appreciate any help.
> > Thanks,
> > Ron
> >
>
>
>
.



Relevant Pages

  • Re: Building a file server - advice please
    ... > connecting up 5 drives in a RAID5 system does not affect the Mean Time To ... > important reason to use a RAID system. ... Hardware controllers generally can have an additional spare disk configured ... Hardware raid presents each raid array to the host as one disk, ...
    (comp.os.linux.setup)
  • Is it HighPoint, is it Seagate, or is it Windows 2000 Professional
    ... Then I tried to boot my system: Windows reported ... Barracuda disk running off a HighPoint HPT370A RAID ... never new that my drives were set up this way. ...
    (microsoft.public.win2000.setup)
  • Re: Is it HighPoint, is it Seagate, or is it Windows 2000 Professional
    ... As you've discovered Raid 0 has no fail safe redundancy, ... Then I tried to boot my system: Windows reported ... Barracuda disk running off a HighPoint HPT370A RAID ... never new that my drives were set up this way. ...
    (microsoft.public.win2000.setup)
  • 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)
  • Re: Paul and Old Man: Cannot fix RAID5 failure ...
    ... Considering 3 SATA drives spining up after powerdown may cause a 12V ... Now I need to concentrate on raid array or data recovery (if still ... I removed the drive reported in error, no boot or rebuild. ... I was considering a parallel WinXP installation on a 4th disk, ...
    (alt.comp.periphs.mainboard.asus)