Why use Circular Logging??
At your point it really becomes a point of how important is this information
to the business. When I started here, we were a small org of about 300-500
users, and they had everything on one server, (PDC,mail,F&P,AV,RAS and no
BDC), when I started going over the problems with this, and what they
expected from this server and its setup they were amazed. The bad part was
that I was sitting in a meeting with the money folks going over why this was
a bad setup and what the cost would be if we didn't correct it when the
server started melting away. I was able to get a BDC built and on-line
before melt down, after the 2 days it took to get it and all its data back
to the way it was, they were more than happy to correct the problem. What a
lot of people don't take into consideration when they are putting everything
on the same server is the amount of time it takes to get the server backup,
not only do you have to do a exchange disaster recovery, now you have to
install all the apps, restore all the other share data, recreate all the
network printers, etc.... At some point the cost of the SCSI controller and
a couple of hard drives starts to have a cost benefit. Today we have a
server that supports about a thousand users, and I would not even blink an
eye to adding another thousand tomorrow. I have been given the green light
to treat this as a mission critical server and run it as such, for that I
sleep and vacation much better.
The bottom line is that no matter what the size of your organization the
server (no matter what it is doing) has to be built to meet the business
needs it supports, and yes that even counts for your e-mail server. It is
possible to use one server with good hardware to function as a one server
shop that is less prone to major problems, will it cost more? Sure it will,
but what is the cost if the system is down? Every job I have done as a
private consultant where they did not want to spend the extra cash up font,
they either decided that the sever had a big enough business impact that
they did not want to risk the long term outage, or they sign a
non-compliance form stating they did not want to follow my recommendations
and they understand the risk of the choices they are making. There have
even been a few jobs that I have walked away from because I just didn't feel
comfortable doing what they requested.
Just my .02 worth.
Jeffrey R. Waters
Senior Systems Engineer
Information Technology, Hanover County
-----Original Message-----
From: Benjamin Scott [mailto:[EMAIL PROTECTED]]
Sent: Friday, August 31, 2001 1:24 PM
To: Exchange Discussions
Subject: Very small servers (was: A good space for the Exchange ...)
On Fri, 31 Aug 2001, Daniel Chenault wrote:
>> I thought that if you lost the logs, there was a procedure to at least
>> restore the data in the Stores? Yeah, you lose anything that wasn't
>> committed, of course, but can't you recover something?
>
> Yes, there is. In my world of support, though, lost data ==
> unacceptable.
Everything is relative. The company I work for supports very small
businesses (no more than 50 people, usually less than 20) and SOHO
enviornments. One of our most common "new customer" scenarios is when the
office has finally grown to the point where they need to buy their first
dedicated network server.
In that sort of situation, you cannot justify stand-by servers for
restores and recovery. Hell, we usually have to argue just to get them to
buy a tape drive.
Now, consider this configuration: NT 4.0, Exchange 5.5, with patches kept
current. Single server for everything -- file, print, mail, backups, etc.
Either a single non-redundent disk, or two disks as a single mirrored
volume. All files, databases, database logs, etc., on the same disk.
Circular logging turned on. It gets backed up every night to tape -- often
a full backup every night. Multiple tapes are used in rotation (father/son,
etc.).
The customer is fully aware that hardware failure will likely result in
permenently losing the changes made since last night's backup. It is
considered an acceptable risk.
My belief has been that this scenario should work. While performance is
far from the theoretial ideal, there is nothing that would cause data
corruption or system failure. If a catestrophic failure occurs (e.g., dead
disk), once the failure is corrected, we should be able to recover to the
point of the previous night's backup.
Is my belief correct, or are there additional factors that I am not aware
of?
--
Ben Scott <[EMAIL PROTECTED]>
| The opinions expressed in this message are those of the author and do not
|
| necessarily represent the views or policy of any other person, entity or
|
| organization. All information is provided without warranty of any kind.
|
_________________________________________________________________
List posting FAQ: http://www.swinc.com/resource/exch_faq.htm
Archives: http://www.swynk.com/sitesearch/search.asp
To unsubscribe: mailto:[EMAIL PROTECTED]
Exchange List admin: [EMAIL PROTECTED]
_________________________________________________________________
List posting FAQ: http://www.swinc.com/resource/exch_faq.htm
Archives: http://www.swynk.com/sitesearch/search.asp
To unsubscribe: mailto:[EMAIL PROTECTED]
Exchange List admin: [EMAIL PROTECTED]
