Thread

eeart - before we address the affects of removing tag caching from your site, can you go into detail about why you chose that method of caching? Are you also using template caching?

can you go into detail about why you chose that method of caching? Are you also using template caching?

Some templates have “Enable Caching?” set to Yes under Template Preferences, but not all. I don’t know why tag caching was used, but I assume to try and speed things up. Do you think using tag caching in templates that are themselves being cached causes problems?

I just found one template that had 11 times ‘cache=“yes”’ on different tags and that template itself also had caching enabled. Is that excessive caching?

The cache folder gets full so quickly. Yesterday when you asked me to empty the cache folder via FTP it took my FTP program over an hour to finish and when it was done there were hundreds of new directories and files already again.

Honestly I’d go through and take out the caching from the tags and only use it in cases where you really need it.  It sounds like someone blindly went through and added features that you honestly don’t need.

There are times to use tag caching, but this isn’t one of them.  I’d rely more on template caching where possible, yes.

There are times to use tag caching, but this isn’t one of them. I’d rely more on template caching where possible, yes.

Okay, we’ll try this.

Can you give an example of when one really needs tag caching as opposed to not?

The help page just says “By caching individual tags you will reduce the amount of scripting and server resources needed” but doesn’t mention the downsides.

The only time I use tag caching is if the template is using something that can’t be cached, such as member conditionals for secure information.

Have a look at this PDF: Performance Guidelines under the Caching section.  It will explain it in detail for you.  It also has some other great hints for optimizing should you be interested.

It all makes sense. Thank you very much for all your help. We have some work to do.  :lol:

eeart, also have a look at Handling Extreme Traffic for more information.

CP remained fast with Tag Caching completely turned off. I removed cache=“yes” in about 75% of cases where the template was also being cached but left it in the rest. In those cases the refresh is set to “30”.

Yesterday I turned Tag Caching back on and today CP is again quite slow when saving things.

I wonder if the server is too slow. I can’t believe it takes a minute for CP to empty the cache folder. Could it be poor server performamce?

I also have a question about the refresh rate. What is the use of setting the refresh rate with tag caching? I understand that the cache is emptied or overwritten when a new entry or comment is posted. So why set a refresh rate? If the cache is always automatically emptied then it can just as well be set to a very high number, right?

eeart,

I wonder if the server is too slow. I can’t believe it takes a minute for CP to empty the cache folder. Could it be poor server performamce?

While template caching caches the entire page, tag (and query) caching is based on the URI which means that every cache file created by EE is tied specifically to a specific page request (ex: /site/comments/42/).

For example, if a page is requested and the template has a tag being cached,then EE will create a cache folder for that page’s URI (ex:/site/comments/42/) and put the cache file for the cached tag in that folder.

So you can see how these files would build up quite fast. Is disabling tag caching permanently an option?

Do you have SQL query caching switched on? This can cause quite a bit of folder build up also. If you have MySQL doing its own caching then this is double work for your server. Try switching it to off to see if it makes a difference

CP Home ›  Admin ›  System Preferences ›  Database Settings

Our website has been getting slower again lately. Today we were informed by our hosting company that the website is using a lot of resources. At one time there were 7 processes of 200 Mb each, using up 1.4Gb of RAM. That can’t be normal, right?

What could be the cause of this? Do we have a memory leak in an extension? Would it be useful to turn off the extensions one by one and see what happens?

There are currenty 25 modules installed plus 8 extensions, including FieldFrame with 8 Fieldtypes. Is that too much or should that be okay?

Is EE 2.0 going to be different in this respect? Will it handle server load and memory usage different?

Thanks.

Would it be useful to turn off the extensions one by one and see what happens?

The first step in testing would be to disable them globally. If that changes anything you need to test them one by one anyway.

That said, server performance issues can be tricky. Every server, every site is different.

I tried Display Template Debugging and the average memory usage seems to be between 8 to 11 Mb. This is nowhere near the 200 Mb that the host is seeing. Does 8 to 11 Mb seem average for EE, or is it a lot?

I’d think that’s about average. This is probably per process, though, so a larger number of processes could multiply that number.

So the memory total that Template Debugging is showing at the bottom is for one process?

The host is talking about 200 Mb per process. At one time they saw 7 processes of 1.4 Gb total. How could this be if I see 8 to 11 Mb here?

A different (runaway?) process, higher server load, different circumstances? They are not claiming these numbers at the same time your’re getting 10 MB, are they?