1 / 5.6 = 17%; pretty average for a SIS factor.
Compression of content will more than make up for it.
Wherever possible, E14 will compress content prior to storing it into the
mailbox database. The cost of doing this is far less than the cost of the extra
I/O.
As Brian notes - JBOD containing multiple copies is the recommended design
criteria. Millions of mailboxes are running in this configuration quite
successfully already. One realtime copy plus a lagged copy.
That doesn't mean you should use the cheapest thing you can find at
NewEgg/BestBuy/whatever; you should still use Enterprise SATA; but it's much
less expensive than Enterprise SCSI.
________________________________________
From: Brian Desmond [br...@briandesmond.com]
Sent: Thursday, May 28, 2009 6:51 PM
To: NT System Admin Issues
Subject: RE: Amusing
They don't cost the same - they're more than the NewEgg variety but they're not
the same cost as a typical SCSI spindle.
You can potentially use the lower end drives for this. There's a few things
with this new setup that help:
--> You can have up to 16 copies of a given database (across 16 servers, 2TB DB
max size)
--> This means you can spend the money on an extra server/replica
potentially as opposed to the cheap disks
--> You can lag these copies
--> This means potentially you can maintain your recovery window for
mail online without going to tape/long term storage
--> The clustering is now natively setup for "stretch" scenarios so you
can cross datacenter boundaries
--> Failover is at a database level now (storage groups are gone, 100 databases
max)
--> If you have one database per disk, and one of your disks fails, you
only failover that database to another server
--> Improvements have been made in the database engine to allow it to detect
more common hardware level storage errors
--> Improvements have been made in the clustering technology such that
the database can repair itself at the page level:
--> Active server detects Page 27 is corrupted
--> Active server requests that the replicas ship Page 27
--> Active server receives Page 27, validates it, and writes it
to the log
--> Active server database is good again without having to
reseed, fail over, restore from backup, etc
Thanks,
Brian Desmond
br...@briandesmond.com
c - 312.731.3132
-----Original Message-----
From: Ben Scott [mailto:mailvor...@gmail.com]
Sent: Thursday, May 28, 2009 5:12 PM
To: NT System Admin Issues
Subject: Re: Amusing
On Thu, May 28, 2009 at 5:41 PM, Brian Desmond <br...@briandesmond.com> wrote:
> ... remember that the SATA drive that comes from HP/Dell in a
> server is (hopefully) not the same drive that you buy for $89 on NewEgg.
The main reason SATA drives are cheaper than SCSI/SAS is that
they're produced in much larger quantities, for the teeming masses of
people who want the most dirt-cheap computer possible. If the SATA
drives you're buying are high-end, server-only products, I would
expect them to cost as much as SCSI/SAS. Which torpedoes the idea
that massively cheaper SATA makes the lack of SIS a design win.
Either we're using crummy drives in our mission-critical servers, or
we're still paying top-dollar for server storage.
-- Ben
~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
