|
|
Sponsored Content
The Lounge
What is on Your Mind?
MySQL Performance Problems
Post 303041393 by Neo on Friday 22nd of November 2019 06:14:47 AM
11-22-2019
Administrator
I just completed (for the most part) a database optimization which was slowing the site down.
In a nutshell, when I created some new code to mitigate against Google soft 404 "errors" (slim content) by adding matching man pages to posts with very little content, I used a "real time" FULLTEXT search of over 350K man pages, and this was causing some performance issues.
Plus, my new code worked, and Google search referrals have just about doubled since I created the solution, but the upside/downside was that the site traffic also doubled (which is a good thing) and the new queries were killing performance (a bad thing).
In the final optimization, I preprocessed all these matches and stored the results in the DB (the ids of all the man pages which match the discussion contents) so, the long MySQL queries doing these FULLTEXT NATURAL LANGUAGE searches on the fly were eliminated (right now about 99.5+ are gone, I have about a half a day of processing to get to 99.999 percent.
The performance improvement is remarkable and the site is much faster, and I have earned my black belt (or at least a brown belt) in MySQL performance tuning, after all these years.
In a nutshell, when I created some new code to mitigate against Google soft 404 "errors" (slim content) by adding matching man pages to posts with very little content, I used a "real time" FULLTEXT search of over 350K man pages, and this was causing some performance issues.
Plus, my new code worked, and Google search referrals have just about doubled since I created the solution, but the upside/downside was that the site traffic also doubled (which is a good thing) and the new queries were killing performance (a bad thing).
In the final optimization, I preprocessed all these matches and stored the results in the DB (the ids of all the man pages which match the discussion contents) so, the long MySQL queries doing these FULLTEXT NATURAL LANGUAGE searches on the fly were eliminated (right now about 99.5+ are gone, I have about a half a day of processing to get to 99.999 percent.
The performance improvement is remarkable and the site is much faster, and I have earned my black belt (or at least a brown belt) in MySQL performance tuning, after all these years.
This User Gave Thanks to Neo For This Post:
|
|
4 More Discussions You Might Find Interesting
1. AIX
... a disk drive to be 100% busy?
hdisk0 100.0 1.3K 342.7 1.3K 22.0 PgspIn 651 % Noncomp 75.5
hdisk1 100.0 1.3K 320.2 1.2K 20.0 PgspOut 6 % Client 75.5
It's really slowing down performance on my system and I would like to know what is causing this.
... (2 Replies)
Discussion started by: bbbngowc
2 Replies
2. SCO
hi
We have migrated SCO 5.0.6 into ESX4, but the VM eats 100% of the virtual CPU.
Here is top print from the SCO VM:
last pid: 16773; load averages: 1.68, 1.25, 0.98 02:08:41
79 processes: 75 sleeping, 2 running, 1 zombie, 1 onproc
CPU states: 0.0% idle, 17.0% user,... (7 Replies)
Discussion started by: ccc
7 Replies
3. AIX
Hello,
I encounter some performance issues on my AIX 5.3 server running in a LPAR on a P520. How do I investigate performance issues in AIX. Is there any kind of procedure that takes me to the steps to investigate my server and find the sub systems that is causing the issues?
The performance... (1 Reply)
Discussion started by: petervg
1 Replies
4. Programming
I need to rotate a pixmap in XLib with some kind of interpolation to reduce the aliasing.
I came up with the following code, which uses bilinear interpolation. It works fine: the rotated image looks perfect, but unfortunately it takes 5 or 6 seconds for each rotation. (in a 300x300, 16 colours... (5 Replies)
Discussion started by: mghis
5 Replies
LEARN ABOUT NETBSD
apropos
APROPOS(1) BSD General Commands Manual APROPOS(1) NAME
apropos -- search the complete content of all man pages SYNOPSIS
apropos [-123456789Ccp] [-n Number of results] [-S machine] [-s section] query DESCRIPTION
The apropos utility performs a full text search over the complete content of all man pages. It uses the FTS engine of Sqlite to perform the search. The database is created with the help of makemandb(8) utility. This implementation of the apropos utility is more sophisticated than the classical version of apropos. Like modern search applications, it uses advanced techniques like stemming and term weighting to rank the matching results in decreasing order of relevance. By default apropos will only display the top 10 matches in the output. Quotes are optional for specifying multiword queries. It supports the following options: -1 Search only within section 1 manual pages. -2 Search only within section 2 manual pages. -3 Search only within section 3 manual pages. -4 Search only within section 4 manual pages. -5 Search only within section 5 manual pages. -6 Search only within section 6 manual pages. -7 Search only within section 7 manual pages. -8 Search only within section 8 manual pages. -9 Search only within section 9 manual pages. -C Do not show the context of the match. -c Do show the context of the match (default). -n Output up to the specified number of search results. The default limit is 10. -p Display all matching results and pipe them through a pager (defaulting to more(1)). -S machine Limit the search to the pages for the specified machine architecture. By default pages for all architectures are shown in the search results. -s section Restrict the search to the specified section of the manual. By default, pages from all section are shown. This option is for back- wards compatibility with the classic version of apropos, using it is equivalent to using the [123456789] options directly. FILES
/var/db/man.db The Sqlite FTS database which contains an index of the manual pages. SEE ALSO
man(1), whatis(1), makemandb(8) HISTORY
The apropos command appeared in 3.0BSD. It was rewritten in NetBSD 6.0 to support full text search using Sqlite. AUTHORS
Abhinav Upadhyay BSD
April 21, 2012 BSD