oracle performance on solaris 8 Post: 31174

Sponsored Content

Special Forums UNIX Desktop Questions & Answers oracle performance on solaris 8 Post 31174 by Perderabo on Sunday 3rd of November 2002 11:08:58 AM

11-03-2002

Administrator Emeritus

First, see this post. Depending on which version of Solaris you are running, you may need to patch and/or tune your kernel to enable the priority paging option. And notice that one source I quoted in that post says that a 300% improvement *is* possible with this option on some versions of Solaris.

Now on to your next question. First, you should have your oracle databases on raw disk space, not in any filesystem. Oracle will tell you the same thing. Putting a database in a filesystem is guaranteed to slow it down. A lot! This may also render it unreliable.

There is a tunefs command that can tune an existing ufs filesystem, but it is limited it what is can tune. Any what tuning is does will not impact existing files unless you unload and reload them. If you going to do that, you may as well rebuild the filesystem from scratch which opens up more tuning options.

The ufs filesystem handles a mix of files very well. It's weakness is very large files. The absolute worst case is one file that consumes the entire filesystem. What makes a file "large" or not is whether or not it can reside entirely in one cylinder group. Increasing the size of cylinder group makes ufs lean more towards handles a few large files very well. Smaller cylinder groups makes it lean towards handling small files distributed in many directories well. If you have a few large files, large cylinder groups may help enough that you notice. On Solaris, the default is 16 cylinders per cylinder group. The best you can do is double that. The other thing is the "-o space" option. If you pick "-o time", Solaris will spend less time picking a new block when the file grows, at the cost of perhaps making a poor choice. You will pay for that each time you read the file. If you pick "-o space", Solaris will spend more time allocating a new block when the file grows. But you get a file that's easier to read. These are the two things that might be most worth persuing. I know I got burned when I said this about kernel tuning, but massive improvments are going to be very rare. Expect only marginal results.

Another option is the veritas filesystem. It is extent based rather than block based. It will handle very large files almost as well as it handles small files. I think it's a seperate product though with Suns.

Perderabo

View Public Profile for Perderabo

Find all posts by Perderabo

2 More Discussions You Might Find Interesting

1. Shell Programming and Scripting

Oracle-performance tuning

Sorry, This is out of scope of this group.But I require the clarification pretty urgently. My Oracle database is parallely enabled. Still,in a particular table queries do not work "parallely" always. How is this?

2. AIX

Oracle performance optimum vs. SGA memory allocation

Dear all experts, I have a p750 Power 7 3.3GHz server with 4 processors and 48GB Memory. This is a DB server which is using Oracle 9i. I have been told that Oracle 9i can only allocate 10GB as SGA Max to get the oracle optimum performance. Anything more will result in overflow of memory and will...

LEARN ABOUT MINIX

tunefs

tunefs(1M)						  System Administration Commands						tunefs(1M)

NAME

       tunefs - tune an existing UFS file system

SYNOPSIS

       tunefs [-a maxcontig] [-d rotdelay] [-e maxbpg] [-m minfree] [-o space | time]  special | filesystem

DESCRIPTION

       tunefs  is  designed  to change the dynamic parameters of a file system that affect the layout policies. When using tunefs with filesystem,
       filesystem must be in /etc/vfstab. The parameters that can be changed are indicated by the options given below.

OPTIONS

       The following options are supported:

       -a maxcontig

	   The maximum number of logical blocks, belonging to one file, that is allocated contiguously. The default is calculated as follows:

	   maxcontig = disk drive maximum transfer size / disk block size

	   If the disk drive's maximum transfer size cannot be determined, the default value for maxcontig is calculated from kernel parameters as
	   follows:

	   If  maxphys is less than ufs_maxmaxphys, which is 1 Mbyte, then maxcontig is set to maxphys. Otherwise, maxcontig is set to ufs_maxmax-
	   phys.

	   You can set maxcontig to any positive integer value.

	   The actual value will be the lesser of what has been specified and what the hardware supports.

       -d rotdelay

	   This parameter is obsolete as of the Solaris 10 release. The value is always set to 0, regardless of the input value.

       -e maxbpg

	   Indicates the maximum number of contiguous logical blocks any single file can allocate from a cylinder group before	it  is	forced	to
	   begin  allocating  blocks from another cylinder group. Typically this value is set to approximately one quarter of the total contiguous
	   logical blocks in a cylinder group. The intent is to prevent any single file from using up all the blocks in a single  cylinder  group,
	   thus degrading access times for all files subsequently allocated in that cylinder group.

	   The effect of this limit is to cause big files to do long seeks more frequently than if they were allowed to allocate all the blocks in
	   a cylinder group before seeking elsewhere. For file systems with exclusively large files, this parameter should be set higher.

       -m minfree

	   Specifies the minimum free space threshold, or the percentage of space held back from normal users. This value can be set  to  0.  How-
	   ever, up to a factor of three in throughput will be lost over the performance obtained at a 10% threshold. Note: If the value is raised
	   above the current usage level, users will be unable to allocate files until enough files have been deleted  to  get	under  the  higher
	   threshold.

       -o space|time

	   The	file system can either be instructed to try to minimize the time spent allocating blocks, or to try to minimize the space fragmen-
	   tation on the disk. The default is time.

	   Generally, you should optimize for time unless the file system is over 90% full.

USAGE

       See largefile(5) for the description of the behavior of tunefs when encountering files greater than or equal to 2 Gbyte ( 2**31 bytes).

ATTRIBUTES

       See attributes(5) for descriptions of the following attributes:

       +-----------------------------+-----------------------------+
       |      ATTRIBUTE TYPE	     |	    ATTRIBUTE VALUE	   |
       +-----------------------------+-----------------------------+
       |Availability		     |SUNWcsu			   |
       +-----------------------------+-----------------------------+

SEE ALSO

       mkfs_ufs(1M), newfs(1M), attributes(5), largefile(5)

SunOS 5.10							    5 Dec 2003								tunefs(1M)