|
|
|
SSC-Dedicated
           
Group: Administrators
Last Login: Today @ 1:14 AM
Points: 31,433,
Visits: 13,746
|
|
tgarland (3/7/2011) Besides the obvious filegrowth setting issues mentioned previously, one thing I found was there was not much use of filegroups. One of the things I like to do is have all of my nonclustered indexes on a separate filegroup that is located on a different I/O path.
To follow on from Brad, we also have cost constraints, as do many other companies. Getting additional drive paths in this hosted environment was outside our budget. We could potentially argue for it, but as Brad mentioned, not sure there is a need for this.
However planning for the future makes some sense. Perhaps Brad, Grant, and I will consider this as a way to plan for future growth and pre-separate things so that a quick alter would allow us to take advantage of a new array.
Follow me on Twitter: @way0utwest
 Forum Etiquette: How to post data/code on a forum to get the best help
|
|
|
|
|
SSChampion
        
Group: General Forum Members
Last Login: Yesterday @ 9:14 PM
Points: 13,381,
Visits: 25,174
|
|
tgarland (3/7/2011) Besides the obvious filegrowth setting issues mentioned previously, one thing I found was there was not much use of filegroups. One of the things I like to do is have all of my nonclustered indexes on a separate filegroup that is located on a different I/O path.
That is a good suggestion. I know we're not talking about performance bottlenecks at this point, but keep an eye out for any indications of I/O loading since something like this could help.
Did you see any issues with any of the other properties?
---------------------------------------------------- "The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood..." Theodore Roosevelt The Scary DBA Author of: SQL Server 2012 Query Performance Tuning SQL Server 2008 Query Performance Tuning Distilled and SQL Server Execution Plans
Product Evangelist for Red Gate Software
|
|
|
|
|
SSCrazy
      
Group: General Forum Members
Last Login: Monday, May 13, 2013 11:21 AM
Points: 2,163,
Visits: 2,148
|
|
I quickly reviewed a bunch of stuff and have some notes, though I won't go into detail as I have limited time:
Review both instances for updating to the most current SQL Server SP, and possibly CU.
Instance 1:
A lot of growth rates are 10% or 1MB (Review and set for each DB/file)
SQLServerCentral_Data file is limited to 10,000.0 MB and it is over 50% of that size, is that limit what you want?
tempDB: Has testing indicated that you actually need 8 data files?
ips_notifications_log size: 0.0MB?
Instance 2:
tempDB: Has testing indicated that you actually need 8 data files?
Simple-Talk DB: Compatibility level: 80 (Is that needed?) Simple-Talk_Log Size: 12,908.0 (4x the size of the data file?)
apps_excuse DB: Compatibility level: 80 (Is that needed?), change page_verify to CHECKSUM
apps_prettifier DB: Compatibility level: 80 (Is that needed?), change page_verify to CHECKSUM
sqlMonitor22 DB: autogrowth: 1GB, log size: 46GB? (Almost 8x the size of the data file?)
|
|
|
|
|
SSCertifiable
       
Group: General Forum Members
Last Login: 2 days ago @ 2:17 AM
Points: 6,862,
Visits: 8,049
|
|
|
|
|
|
SSC-Addicted
      
Group: General Forum Members
Last Login: Tuesday, January 22, 2013 8:49 AM
Points: 490,
Visits: 107
|
|
I noticed the same things that others have mentioned: file growths and sizes, tempdb, filegroups...
I have always been advised to use a set size, rather than percentage, for file growth. I have participated in a Microsoft SQL Server RAP, and they advised us to do things in increments of 1024MB.
So I set all all my file growths and sizes in multiples of 1024M.
I realize you wanted us to focus on settings we can see from the monitoring tool, but I am curious about instant file initialization, ad-hoc optimization, asynchronous update stats, etc.
Is there a way to customize what the tool shows you?
Regarding TempDB and the number of files, I was not able to find how many cores are on the servers.
What about separating TempDB MDF and LDF?
Also regarding separate filegroups for non-clustered indexes. In addition to the possible IO related benefits, I find that there are several administrative benefits, as well. For example, you then have more flexibility in where you locate your files, you can take filegroup backups, you can do filegroup dbcc checkdb commands, etc.
The forums database seems to be using full text filegroups. What about locating them on a separate file system from the regular data files?
|
|
|
|
|
SSC-Dedicated
           
Group: Administrators
Last Login: Today @ 1:14 AM
Points: 31,433,
Visits: 13,746
|
|
|
|
|
|
SSCommitted
      
Group: General Forum Members
Last Login: Yesterday @ 1:32 PM
Points: 1,820,
Visits: 1,899
|
|
| One of the other things I noticed was that the collations are different for some of the databases. SQLServerCentral, as well as others, use Latin1_General_CI_AS as where some of the others use SQL_Latin1_General_CI_AS like CommunityServer and SQLServerCentralForums.
|
|
|
|
|
SSCommitted
      
Group: General Forum Members
Last Login: Yesterday @ 1:32 PM
Points: 1,820,
Visits: 1,899
|
|
| I also noticed that there are a lot of unlimited max file size settings for the databases. I personally prefer to set those limits to something reasonable and monitor the filegrowth and adjust them accordingly.
|
|
|
|
|
Forum Newbie
      
Group: General Forum Members
Last Login: Thursday, March 24, 2011 2:11 PM
Points: 4,
Visits: 24
|
|
| Can you tell us more about the hardware configuration?
|
|
|
|
|
SSC-Enthusiastic
      
Group: General Forum Members
Last Login: Tuesday, May 14, 2013 1:08 PM
Points: 175,
Visits: 719
|
|
Steven O (3/9/2011)
I realize you wanted us to focus on settings we can see from the monitoring tool, but I am curious about instant file initialization, ad-hoc optimization, asynchronous update stats, etc.
Is there a way to customize what the tool shows you?
Regarding TempDB and the number of files, I was not able to find how many cores are on the servers.
What about separating TempDB MDF and LDF?
Also regarding separate filegroups for non-clustered indexes. In addition to the possible IO related benefits, I find that there are several administrative benefits, as well. For example, you then have more flexibility in where you locate your files, you can take filegroup backups, you can do filegroup dbcc checkdb commands, etc.
The forums database seems to be using full text filegroups. What about locating them on a separate file system from the regular data files?
The server has Dual Quad Core Xeon L5520 (2.66GHz) CPUs, for a total of eight cores.
MDFs and LDFs are on separate RAID 10 arrays. Tempdb is also on its own array. The arrays are part of a SAN where we have 20 dedicated disks.
Later in the series, I will go review the rest of the settings you asked about.
Brad M. McGehee Microsoft SQL Server MVP Director of DBA Education, Red Gate Software www.bradmcgehee.com
|
|
|
|