<Neil.Braebaum@shopdirect.com> wrote:
>> -----Original Message-----
>> From: thin-bounce@freelists.org
>> [mailto:thin-bounce@freelists.org] On Behalf Of Matt Kosht
>> Sent: 19 December 2008 15:59
>> To: thin@freelists.org
>> Subject: [THIN] Re: Citrix - Harddisk Specification
>>
>> My 2 cents...
>> If you size your farm to allow for the loss of 1 or more
>> Citrix servers this isn't a big deal.
>
> It's not just the sizing and dealing with a server being down - it's
> bearing the transient, albeit potentially serious, pain of sessions
> (work and data) dying or being lost, if a server crashes due to an
> essential disk failing.
>
Certainly RAID "might" save you from this. I have personally seen
disks fail and the RAID controller/driver fails to negotiate the
failover to a redundant disk resulting in a system crash. So RAID
isn't foolproof or a guarantee by any means. The performance overhead
of RAID however is felt by the user every day they use the Term
Server vs. the once in blue moon event of a disk failure. These days
with predictive failure features you will most likely know in advance
that the disk is going to die. I still have RAID controllers in the
server I just don't use the RAID features per se.
A session crashing can also be largely mitigated with autosaves
(Office) or the features inherent in most client/server programs
(rollback). Granted I am sure there are scenarios where crashing
sessions are never an option (Emergency dispatchers, Nuclear Power
plant operator, etc) and where RAID protection isn't an option but a
necessity.
>> It's not a big deal these days to lose a disk on a server
>> that's practically identical to other servers in a load
>> balanced farm. The users just get reconnected to another server.
>
> Indeed, users can reconnect to another server. However, users / audience
> / usage may find that losing a session like that, when it can be fairly
> easily avoided (ie employing the use of some hardware RAID controller
> that tends to be either embedded or stock on many servers, and adding a
> disk, having it monitored by management software, and having a bad disk
> simply hot-swapped without incident) may not be that tolerable.
I don't disagree. I use blades that only have 2 drive bays so my
options are limited. I chose performance vs. session uptime as I
think it's the better risk to benefit tradeoff for my users.
>
> Users can and often do lose work and data, if a Citrix server just ups
> and dies (or should that be downs and dies...).
>
And usually because something else happened (network outage, spilled
coffee, app crashed, someone bumped a power cord, etc) in my
experience :) Disks are probably one of the least likely things to
just up and fail when stacked up against all the other foibles of
users and their applications.
Merry Christmas to you too Neil. I have been following you on this
list for quite some time now. I respect your opinions very much.
-Matt
************************************************
For Archives, RSS, to Unsubscribe, Subscribe or
set Digest or Vacation mode use the below link:
http://www.freelists.org/list/thin
Follow ThinList on Twitter
http://twitter.com/thinlist
Thin List discussion is now available in blog format at:
http://thinmaillist.blogspot.com
Thinlist MOBILE Feed
http://thinlist.net/mobile
************************************************
No comments:
Post a Comment