I have a classifieds site uses MySql as database server in which thousands of new ads are posted and each ad will be tagged with few tags. Now when I store each tag in a separate row along with the ad reference id, I could see about 3 to 4 thousands of rows are added to the table which is increasing load heavily day by day.

Here is the example of storing tags in a table.
postid tagname
1 real-estate
1 homes
1 best-homes
2 job
2 internet-jobs
2 workfromhome

Is there any way to decrease the load on the server. Can I split table into parts and store in different servers and use them as a single table when the query is executed?

what efforts have you made to optimize your queries with appropriate indexes?

I have only 2 columns(postid tagname) in the table and I created a Boolean index on postid.

please do a SHOW CREATE TABLE on the table, and give an example of a search query that uses this table

I used following query to create table

CREATE TABLE `draggin_classifieds`.`cfd_tags` (
`postid` BIGINT NOT NULL ,
`tagname` VARCHAR( 128 ) NOT NULL ,
INDEX ( `postid` )
) ENGINE = InnoDB;

Now when I wnat to retrieve all the posts with a tagname I am using

SELECT * FROM cfd_tags WHERE tagname='sometag' LIMIT 0, 20;

that cannot possibly be the correct table layout, because if postid is the primary key, then there can be only one row for each postid, however, your example says you have multiple rows per postid like this --

1 real-estate
1 homes
1 best-homes
2 job
2 internet-jobs
2 workfromhome

so something is wrong, you have misrepresented something, and i can't help you until you clear this up

I am sorry. I corrected the create table query above.

Here is that

CREATE TABLE `draggin_classifieds`.`cfd_tags` (
`postid` BIGINT NOT NULL ,
`tagname` VARCHAR( 128 ) NOT NULL ,
INDEX ( `postid` )
) ENGINE = InnoDB;

remove the index on postid, then run the following two statements--then try your query again and see how fast it is