Reviews:
 ·TekSavvy DSL
| deep drilling access_logs There is a weird issue going on a site I manage.
During peak points of the day, the website is freezing up. I know it is related to the database (mysql) and I have tried to track down the exact cause.
The main issue is tracking down where the issue is occuring. I have apache logs and individual user logs.
Anyone recommend any stat genertors that will let me drill down deeply through the logs. Everything I can find will give me general stats but I can't seem to find one that will let me focus on "3pm on Monday June 22" and see what is happening "around' that time or see what is being driven.
I know I can open the log file and read it directly, but that doesn't seem to be giving the over all map I need.
Thanks.
--
»www.codecipher.com - Marking the way to tomorrow's solutions -- Did you know that Perl is not Dead? »perlisalive.org/ |
|
 nilJava Geek
join:2000-11-27
kudos:1 | Apache logs probably won't help you as much as you think. Do you have any application logs?
--
Life is too short to be boring |
|
|
|
 geekamongusReal Slump QualityPremium,MVM
join:2004-07-27
Asheville, NC | reply to Devanchya
This sort of thing can usually be best figured out by comparing a number of things during times which the web site freezes up.
Apache logs can be helpful for seeing who was on your site at the time, and how many people were on your site at the time.
You would also want to monitor the performance of MySQL during peak hours as well. Here are some tips for doing that.
Checking the MySQL error logs could also be helpful.
It would also be wise to monitor your server's performance to make sure you aren't looking at a memory and/or processor bottleneck.
--
geekamongus tech blog | moi |
|
| reply to Devanchya
Modify the app to log execution times for various steps. Then you'll be able to focus on the issue. I've seen issues where the database seemed to be lagging, yet the cause was the app trying to process too much data. YMMV. |
|
Reviews:
·TekSavvy DSL
| reply to Devanchya
I have a bunch of logging already in. the 'easy read' part of it was never built.
So I did figure out what happened:
3-4 of my users were infected by FunWebProducts. They were creating 2-3 connections a second each to the server. Now the site is busy enough that I didn't see it right away. I happened to catch them just as th PC were obviously turned on in the morning.
I have created a rule against this product and will be doing research on how to help prevent this type of fun.
Luckly the FWP was not able to get into the secure part of the website, though I know anything visible on the screen is visible to any viruses.
--
»www.codecipher.com - Marking the way to tomorrow's solutions -- Did you know that Perl is not Dead? »perlisalive.org/ |
|
 sporkmedrop the crantini and move it, sisterPremium,MVM
join:2000-07-01
Morristown, NJ
 ·Optimum Online
| reply to Devanchya
If you strongly suspect mysql is the problem, turn on the slow query log:
»dev.mysql.com/doc/refman/5.0/en/···log.html
--
with every mistake we must surely be learning |
|
 Jafo232You Can't Spell Democrat Without Rat.Premium
join:2002-10-17
Boonville, NY | reply to Devanchya
Just grep the log for what you are looking for?
Example:
grep 22\/Jun\/2009:03:00 /var/log/httpd/access_log
--
Custom PHP/Perl Development. Vbulletin And Wordpress Mods Too! |
|
