I think we have isolated the issue. It appears to be an Internet backbone routing issue when you are routed through Cogento routers on the backbone. Please see this:

/p937074-site_is_slowing_down_needs_maintenance.html#937074

and post a trace route for Paul. That will help to pinpoint diagnose the problem.

I don't know. I seem to be getting past Cogento and hanging up at pao1.isc.org. I sent Paul 4 runs.

_________________
Bill (AFE7Ret)
Freedom is NOT Free!

Some of today's stats:

58ppm 54.167s (52.237s)
113ppm 18.826s (16.981s)
142ppm 27.348s (20.978s)
116ppm 18.704s (16.856s)

Then, to illustrate the inconsistencies:

165ppm 4.334s (3.662s)
182ppm 0.940s (0.563s)
175ppm 0.355s (0.107s)


The DDoS was more than 4 months ago in February. Are the frustrating problems we are still experiencing still the result of the attack?

I'm fortunate that I have a pretty good understanding of computers, the IT industry, and how to be safe, so I don't come here seeking advice, but rather to offer my help. It makes ME feel good if I can help someone have a safe computing experience. Okay, that's selfish, but that's the way it is.

Don't get me wrong - I appreciate the opportunities we have here, and I have no doubt all those involved in getting the site back up to speed are working hard at it, and are just, if not more, frustrated.

I can't speak for anyone else, but 4 months of nagging delays takes much of the enjoyment out of volunteering time here.

What is left to be done? How can we help?

_________________
Bill (AFE7Ret)
Freedom is NOT Free!

220ppm 0.424s (0.081s)

We're still not out of the tunnel yet, but we have certainly come a long way. There is still more work to do.

I can get MySQL to look at the problem, but the fees are a lot.

Also take note, as of 1 - 2 weeks ago since a couple key updates took place, our traffic has doubled on a per hour basis (on average). Also, we had over 100k PIRT urls submitted last night and our background scripts are running processing them -- but that normally has no significant impact. Right now what is typically happening is some of the php scripts are running harder then they probably should be for certain connections -- and that goes back to optimizations.

158ppm 2.733s (0.238s) -

Thanks for the reply.

Question - this doubling/hour in the last couple weeks, how does that rate compare to pre-attack rates? It appears to me, since the attack, traffic in the forums (at lease those I frequent) has dropped significantly, I suspect out of user frustration causing them to seek answers elsewhere. But as you noted, there have been recent improvements, possibly bringing folks back.

BTW - just before lunch, I marked all forums as read then stepped away from my computer. When I sat down now for this session, I clicked on "New Forum Posts" to see what was new - your two replies were not listed - I only knew about them from email notifications. Only 5 in HJT and 1 in Sunbelt came up on the new list . Upon further investigation, I see there is also one new post in GCP, 2 in Spam and probably some others. It would seem that function is not working correctly either.

Edit add: 192ppm 17.162s (15.938s) for the original submit.

_________________
Bill (AFE7Ret)
Freedom is NOT Free!

I don't trust that new replies thing and instead use the egosearch.

I think part of the traffic increase has been due to the ability for crawlers to access the site quicker, thereby showing better results in search pages.

Changing to SQL?

Hi, I was asked by faith_michele to help if I can with some of your latency issues. I am a fellow Norwich MSIA student with her.

Anyway, I am mostly a .NET developer, but have some experience with PHP and MySql. I don't know how much or how little experience you all have so will just ask some very general questions first:


PHP: There are a lot things that can be done to optimize code but...

What version are you using? 4.x, 5.x?

What OS is the site running on? Linux? Windows?

Are you using a PHP framework, like CakePHP, or scripting natively? Sometimes frameworks, while making development relatively easy, can result in excessive function calls, as you need to work through multiple layers of abstraction.

Is your PHP being run with FastCGI? This essentially enables the runtime to reuse code without having to run the interpreter each time. There is also something called the Zend Optimizer, that supposedly speeds up execution times, but I am not real familiar with it.

MYSQL:

Very basic, but are all your tables fully indexed? Lack of indexes is the biggest cause of db latency. At the very least you should index your primary key, and one other column that would be most relevant.

Hope some of this helps

It is a very busy week (assignments due) this week. Paul, dsf offered to help and I can have him directly email you, if that is okay.