Deleted items, garbage and filesystems

[simon]

Alfresco,

Yesterday I removed some files in my 1.3.2 Enterprise version. They showed up in the "manage deleted items" as expected. I removed these files here as well.

What's happening at filesystem level at this point? Is the file removed?

I copied my content store at filesystem level later that day and noticed some of the files I removed from the "manage deleted items" passing by. As I had been testing with large ACP packages my disk space went form 25% full to 60% full in one day even though I thought I removed everything twice.

In short… is the contentstore.deleted cleaned when I remove deleted files with the "manage deleted items"? I'm not planning to keep al these removed files for ages.

Thanks!

[derek]

Hi,

I'm putting together a WIKI page to describe the behaviour and how it can be changed.  I might not have the time before 2.0 final, but I'll include a sample xml config that will throw the content away faster.

http://wiki.alfresco.com/wiki/Content_Store_Configuration

Regards

[simon]

Thanks Derek. I suppose you are still working on this page as the most relevant part for me seems missing:

In the default configuration, the contentStoreCleanerTrigger fires the‍

Fires what, how, where to change it, is this something for v2 or can it be done in 1.4.x as well? In other words, how can I clean the deleted items now?

It seems that it's possible to remove the contentstore.deleted files on file system level. I did this and Alfresco is still running, is this correct? I now have a clean contentstore.deleted folder.

I removed a different file afterwards, expecting it would show up in the contentstore.deleted folder but it didn't. Is this normal?

[derek]

Hi,

It is not possible to clean the items immediately.  Realistically, a system should never be on the edge of running out of disk space.  It is, however, possible to reduce the amount of time that content 'lurks' in the system and to have it permanently deleted instead of backed up.

Indeed the WIKI page was WIP, but it should be adequate for your requirements now.

Regards

[simon]

Thanks Derek, I'm trying this approach now to see if it fits our needs.

I count all the files in the contentstore.deleted folder, add a new file and remove it in the web client, clean out the deleted items (again via the web client) and recount all files in the contentstore.deleted.

This shows me exactly the same number of files even though I removed a file that wasn't there before. What am I missing?

[derek]

Hi,

The protectDays should be at least 1, so nothing will happen for at least a day as far as the server is concerned.  If you have configured in the eager contentStoreCleaner as given in the WIKI, the contentstore.deleted will no longer be used.  Describe and post your config and I'll describe the system's behaviour, but you'll have to verify it over days rather than minutes.

Regards

[simon]

Derek,

I may not have been clear enough but I think we are discussing 2 different points here. This is the lifecycle of a document if I'm right:

created > removed > removed in the "deleted items" > removed from the file system

What you are describing in the scheduled-action-services-context.xml is when the document is between step 2 and 4, isn't it? So when the document is removed from Alfresco it takes the protectDays to show up in the contentstore.delete?

What I need is the file on the file system. My contentstore.deleted is full of files even if I cleaned out the "deleted items" in the Alfresco web client. So can I remove files in the contentstore.deleted? What is it used for? Is this the "garbage bin" you get when you open the "deleted items" in the web client or is this the place where documents who haven't anything else to do end up.

In short, my contentstore.deleted is taking 60% of my diskspace, how do I clean this without braking the database?

[derek]

I've updated the Wiki, but the essence is this:  The contentStoreCleaner can either erase or sequester orphaned content.  Everything you see in contentstore.deleted has been orphaned and can be removed at will and with impunity.  The example in the Wiki of eager cleanup will not sequester the orphaned content but will just throw it away.

And I've notified our technical writer who will make sure that all details are made clear in the manuals.

Regards

[simon]

I'm testing this out right now and can clean my production disk if this works.

Thank you for the great help and useful feedback!

[lakshmi]

Hi Simon,

Even am testing this functionality, my requirement is

Deleting files manually in Alfresco UI using trash can does not reduce the disk space. So the automatic cleaning process should also "reduce the disk space".

Where you able to clean your disk space also?

Did you try the eager cleanup?

Can you post your observation pls…..