Comments on: MySQL Server Upgrade

By: MySQL Performance Blog » Even minor upgrades are not always safe

MySQL Performance Blog » Even minor upgrades are not always safe — Mon, 04 Sep 2006 14:55:50 +0000

[...] I already wrote couple of weeks ago I keep most of my systems on MySQL 4.1 still as they will not benefit from MySQL 5.0 features anyway while I do not want to likely loose a bit of performance and possibly deal with new bugs and changes introduced in MySQL 5.0 (You never know where to expect problems, ie some people reported JOIN breaking by upgrading to 5.0). Yes this stuff is in upgrade notes but if you run third party software it does not really help. [...]

By: peter

peter — Mon, 14 Aug 2006 21:12:55 +0000

Could be file cache or APC. Any would work much better.

The problem is actually finding a proper way to invalidate the cache - when I want it to expire ? I want articles to appear at once so it should be invalidator plus I have calendar so date change also should invalidate it.

It also can be mixed with e-tag to avoid page transfer if it was not changed.

I will also check first if there are any wordpress plugings which do what I want to do.

Also may be hiding much of the post under [Read More] will be helpful to save traffic.

By: Apachez

Apachez — Mon, 14 Aug 2006 21:06:15 +0000

Perhaps a file cache as in http://www.mysqlperformanceblog.com/2006/08/09/cache-performance-comparison/ ?

By: peter

peter — Mon, 14 Aug 2006 20:53:19 +0000

Yes. I guess this is because I added few large articles with many SQL queries and highlighting module seems to be slow. Server buffers full page before sending and there is gzip compression enabled for clients which can do it.

I should find different module and I should have this page cached rather than generated each time. It just requires some time.

By: Apachez

Apachez — Mon, 14 Aug 2006 19:59:16 +0000

I still takes 1 or 2 seconds before the servers starts to send data but I think it was even worse earlier… or maybe its just a psychological effect

By: peter

peter — Mon, 14 Aug 2006 19:45:35 +0000

I’ll tell you sad story.

This host also runs nagios which monitors bunch of servers in different data centers with many services on each. Just as you was browsing the site there was large packet loss to one of the data center which made nagios services randomly going down and recovering. This generated email storm which was processed by spamasassin which loads server quite seriously. I suspect this was the reason.

I should probably setup nagios forwarders on each of data centers and learn spamasassin to skip local system mail but this time it was as it was

I hope it is better now ?

By: Apachez

Apachez — Mon, 14 Aug 2006 19:30:28 +0000

Hmm I see now that I was a bit unclear…

I meant that 1-2 weeks ago this site had much faster response specially for the first page…

By: peter

peter — Mon, 14 Aug 2006 18:37:55 +0000

Apachez,

It is just pretty old server which does a lot of things besides this web site.
There is eaccelerator installed so it should be OK. I guess I just have pretty complicated first page, which can be over 150KB in size, especially with code formating module which seems to be rather slow.

Yes something needs to be done with it sometime

By: Apachez

Apachez — Mon, 14 Aug 2006 18:27:34 +0000

Peter, you have probably already noticed this but in case you havent…

I have noticed that this site takes an additional seconds before it replies on my http requests. Specially for the first visit.

Perhaps some e-accelerator (or whatever is being used to boost the php on this site) is broken (it kind of feels the same way :P) ?