Re: Strategies for optimizing a read-only table

[gerald_clark]

Harald Fuchs wrote:

In article <[EMAIL PROTECTED]>,
gerald_clark <[EMAIL PROTECTED]> writes:
 

Matt W wrote:
   

 

Hi Jeremy,

Sorry, it seems like I'm saying this a lot lately. Is it not true that
if the whole table will fit in [free] RAM, that the OS will cache the
file data and there is no need for a RAM disk. I don't really see how
performance would be any different than using a RAM disk. Either way,
you will still have the overhead of the filesystem calls, even if data
isn't actually read from disk, unlike with a HEAP table.
     

 

Just because a file is in the cache now, it doesn't follow that it
will be in the cache 5 minutes from now.
   

If it gets thrown out of the FS cache, you don't have enough RAM, and
then a ramdisk won't help much.
If you have more data than ram and don't access that file's data often 
enough it will eventually be forced out
of cache.
The question then is: Is having this particular table in ram more 
important than having
that other data in cache?
You are probably right.
If it is being forced out of cache, then it is probably not being 
accessed often enough to belong in a ram disk,
and the cache might make better use of the ram.

 



--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:    http://lists.mysql.com/[EMAIL PROTECTED]