I'm studying for my MCDBA and I am getting confused about optimal placement
of data within filegroups. BOL says (in the discussion on OLTP design
considerations):
"I/O bottlenecks are a big concern for OLTP systems due to the number of
users modifying data all over the database. Determine the likely access
patterns of the data and place frequently accessed data together. Use
filegroups and RAID (redundant array of independent disks) systems to assist
in this."
Question 1] what is meant by frequently accessed "data"? Do they really mean
"frequently accessed tables"?
If so, this seems to be in conflict with a question in sample test 70-229
that I downloaded from http://www.certificationking.net where they say that
you should create tables that are frequently joined together in different
filegroups, to allow separate threads to access the tables.
In practice, I have just used one default filegroup and a RAID array.> Question 1] what is meant by frequently accessed "data"? Do they really
mean
> "frequently accessed tables"?
Sounds like it to me.
> If so, this seems to be in conflict with a question in sample test 70-229
> that I downloaded from http://www.certificationking.net where they say
that
> you should create tables that are frequently joined together in different
> filegroups, to allow separate threads to access the tables.
>
In my experience, extra filegroups by itself does no good whatsoever. Others
may argue with that, but thats what Ive seen. Now taking those filegroups
and placing them on other drives will help out quite a bit.
"Curly" <XeveryidiwantiskenX@.yahoo.com> wrote in message
news:iSuCb.850$0s2.241@.newsread2.news.pas.earthlink.net...
> I'm studying for my MCDBA and I am getting confused about optimal
placement
> of data within filegroups. BOL says (in the discussion on OLTP design
> considerations):
> "I/O bottlenecks are a big concern for OLTP systems due to the number of
> users modifying data all over the database. Determine the likely access
> patterns of the data and place frequently accessed data together. Use
> filegroups and RAID (redundant array of independent disks) systems to
assist
> in this."
> Question 1] what is meant by frequently accessed "data"? Do they really
mean
> "frequently accessed tables"?
> If so, this seems to be in conflict with a question in sample test 70-229
> that I downloaded from http://www.certificationking.net where they say
that
> you should create tables that are frequently joined together in different
> filegroups, to allow separate threads to access the tables.
> In practice, I have just used one default filegroup and a RAID array.
>
>
>
>|||Chris & Curly:
BOL make some subtle distinctions. Distinguish between
multiple files vs. multiple disks
Read and compare the following 2 articles in BOL:
"Placing Tables on Filegroups"
Excerpt: "Eventually, there is a saturation point when
there are too many files and therefore too many parallel
threads causing bottlenecks in the disk I/O
subsystem. ... If the Disk Queue Length counter is
greater than three, consider reducing the number of files."
and "Files and Filegroups":
" ...A table can then be created specifically on the
filegroup fgroup1. Queries for data from the table will be
spread across the three disks, thereby improving
performance. The same performance improvement can be
accomplished with a single file created on a RAID
(redundant array of independent disks) stripe set."
The first article deals with the distinction.
The quesiton you refer to deals specifically with RAID.
Brian K
>--Original Message--
>> Question 1] what is meant by frequently
accessed "data"? Do they really
>mean
>> "frequently accessed tables"?
>Sounds like it to me.
>> If so, this seems to be in conflict with a question in
sample test 70-229
>> that I downloaded from http://www.certificationking.net
where they say
>that
>> you should create tables that are frequently joined
together in different
>> filegroups, to allow separate threads to access the
tables.
>In my experience, extra filegroups by itself does no good
whatsoever. Others
>may argue with that, but thats what Ive seen. Now taking
those filegroups
>and placing them on other drives will help out quite a
bit.
>
>"Curly" <XeveryidiwantiskenX@.yahoo.com> wrote in message
>news:iSuCb.850$0s2.241@.newsread2.news.pas.earthlink.net...
>> I'm studying for my MCDBA and I am getting confused
about optimal
>placement
>> of data within filegroups. BOL says (in the discussion
on OLTP design
>> considerations):
>> "I/O bottlenecks are a big concern for OLTP systems due
to the number of
>> users modifying data all over the database. Determine
the likely access
>> patterns of the data and place frequently accessed data
together. Use
>> filegroups and RAID (redundant array of independent
disks) systems to
>assist
>> in this."
>> Question 1] what is meant by frequently
accessed "data"? Do they really
>mean
>> "frequently accessed tables"?
>> If so, this seems to be in conflict with a question in
sample test 70-229
>> that I downloaded from http://www.certificationking.net
where they say
>that
>> you should create tables that are frequently joined
together in different
>> filegroups, to allow separate threads to access the
tables.
>> In practice, I have just used one default filegroup and
a RAID array.
>>
>>
>>
>>
>
>.
>
