search

linagora / linshare

LinShare
https://www.linshare.org/
GNU Affero General Public License v3.0
449 stars 84 forks source link

check old files #284

Open matdemb opened 10 months ago

matdemb commented 10 months ago

Hello, In our installation I think there is still some old files that have not been deleted automaticaly and not in the database in document_entry (there are in document). How can I be sure there is not "forgotten files" and clean this files? Best regards

matdemb commented 10 months ago

Maybe this files are in shared spaces/workgroup. How can I know the owner of a file in a sharedspace? I know how to do for normal documents but not those in shared spaces... Thanks

wboudiche commented 9 months ago

Hi, You can find sharespaces info in the mongo database

matdemb commented 9 months ago

Hi, I can find workspace withe the name and there author but not the file names. Doesn I need to use the workspace uuid in the posgresql document table in bucket_uuid column? Or how can I find the content of a workspace? Thanks

matdemb commented 9 months ago

I found the way to request informations in mongo If I do the sum of all documents in work_group_nodes table, I obtain 7Go. I have 7Go of data in DOCUMENT_REVISION type, dispite the revision functionnality is disable. Is it counted in the quota? The quota in the UI is indicated at 18Go for shared space. What can be the difference between the information in the database and the UI?

The document_entry is indicated at 34Go

my total space used on my disk is 47Go

So 6Go of data on the disk doesn't seems to be referenced in databases (47-7-34). Or there is something wrong in my method to calculate it? Or some files are stored on my disk but not referenced in databases (nothing else than filesystemstorage directory in my partition)? Any idea?

wboudiche commented 9 months ago

WorkGroupDocument.documentUuid in mongo, is the good place to seek.

matdemb commented 9 months ago

I didn't find any document of the workGroup collections in my pathfile (using uuid,parent_uuid of document or document_revision). All files are the ones in document table of postgresl (except one) but none from mongo. Is there a link between worgroup documents in mongo and documents in pgsql? uuid doesn't match.

What about quota difference? any idea?

matdemb commented 9 months ago

it was document_uuid column... too many uuid :-) So I have only one file not referenced. Now I just need to understant why the quota is anormaly high.

matdemb commented 7 months ago

If a document is in the filesystem.directory but no trace in workGroups but only in document (no entry with its document_id in document_entry), can we consider that this file is not used by anybody and delete it from disk (and from document table)? If so, there where probably a migration issue and a script to correct this will be great.