Weirds files making the site 320MB more fat

Author Message

Nicolò Zanetti

Monday 14 April 2008 7:47:30 am

Hi ezcommunity,
I was uploading a website that should be really small, just 4 pages, but the file of the directory was 391 MB uncompressed, that was a bit strange.
Having a step backward, this site is a copy paste and edit of a much larger site. Those files, something like 320MB, was in /var/ezwebin_site/storage/original/application, when i deleted the full /application directory the site still worked, and it's working with no problems, so this is almost solved, i just found that strange, have you any tip about what those file was and why they where such huge?

One more thing, in /var/ezwebin_site/storage/image/ there are all the images that i uploaded with the ezinterface, also the one of the copied site, it's a 36MB on 77MB space waste, I'm afraid it's a bit dangerous to delete them also if all the old nodes (the one from the copied site) where deleted and rewritten. If i manually delete them can this have conseguences on the site stability?

Maxime Thomas

Tuesday 15 April 2008 10:38:43 pm

Some information for you :
- eZPublish is using image aliasing : it means that each image you upload will be declined in several alias as small, medium and large pictures. If you have a lot of pictures it can take some space.
- eZPublish is versionning content. It means that each content you upload is stored and each time you edit this content, a bunch of new attributes will be created. It implies that if you use the File content to store big files (eg Windows SP1), and you see that you have made a mitake in the comment of the content and you modify it, the Windows SP1 is duplicated. You can limit that with some Version settings.
- all the data files (images, videos, PDF and so on) are stored in the var/<siteaccess> directory. If you delete them you will lose data files and so, if you try to download a file from the front interface, you will have an object not found error.

Hope it helps

Maxime Thomas
maxime.thomas@wascou.org | www.wascou.org | http://twitter.com/wascou

Company Blog : http://www.wascou.org/eng/Company/Blog
Technical Blog : http://share.ez.no/blogs/maxime-thomas

Nicolò Zanetti

Thursday 17 April 2008 1:13:37 am

Hi Thomas,
First of all thank you for your help.
If the only consequences of deleting those file will be "missing file error" I will delete them now, I was just worried about system error, since I don't need them in the copied site.
Anyway I haven't uploaded anything such huge in the site, and the removal of the var/siteaccess/storage/original/application directory hadn't any conseguences in the site, apart from the loss of some *.swf file, reuploaded, now it's everything fine.
Since I don't have unlimited disk space, any useless file deleted is a good thing.
Thank you again, bye

Nicolò Zanetti

Thursday 17 April 2008 3:32:31 am

Ok that 320 MB was the *.swf pdfs and all the other files I uploaded, like Thomas said, but they was about the main site I copied the data from, it was simple... now i know it.

All the pdf did lost their extension, and their original name (now they are something like "02d261ab6246eec2f8b3b3e8b1fbf06f") so it was not that easy to catch this, anyway if i upload a file by the ez interface, and than i delete it, it will still be there, because the local full site directory (images, swf, pdf, video, docs...) is 220MB, the ezpublish /storage/original/application itself is 320MB, and i deleted all the files that i use no more with ezinterface, so there is a lot of MB that still should be useless, and if i upload a file, i save the http adress , than i delete the file with the ezinterface and i use the saved adress i will still find the item. So I think that deleting them from the ez panel is just deleting the link, but the files are still there in that directory. But this is a waste of space, how can i clean it without a full wash&reupload operation?

Powered by eZ Publish™ CMS Open Source Web Content Management. Copyright © 1999-2014 eZ Systems AS (except where otherwise noted). All rights reserved.