Clean up share_fileshare table

Seafile Server
1

Hi,

I try to clean up our seafile database. We have more than 26.000 inactive users in the ccnet_db.LDAPUsers table. Most of them are ex-students who left university.

If I delete an inactive user from seafile, it seems that not all references to this user are deleted from the database tables. I deleted a user who was inactive for more than 10 years. His libraries have been deleted, but I found out, that there are still 131 entries in the table seahub_db.share_fileshare. The table has 61.000 entries, the oldest are from 2013.

How can these (and other) data be cleaned up, when a user is deleted?

Thanks,
Dirk

2

Please note the table ccnet_db.LDAPUsers (and its related table ccnet_db.LDAPImported) is no longer used since Seafile version 11.0 .

In version 11.0, Seafile underwent a major architectural change regarding LDAP integration:

  • Reimplementation: LDAP authentication was moved from the C-based ccnet-server component to the Seahub Python codebase .
  • User Storage: LDAP users are now stored directly in the ccnet_db.EmailUsers table alongside regular users, rather than in separate LDAP-specific tables .
  • Mapping: A new mapping system was introduced using the social_auth_usersocialauth table to link external LDAP identifiers to internal Seafile user IDs .

When upgrading to version 11.0 or later, administrators are required to run the migrate_ldapusers.py script to merge data from the old LDAP tables into ccnet_db.EmailUsers . After this migration, the old tables like LDAPUsers or LDAPImported remain in the database as legacy/obsolete data and can technically be cleaned up once the migration is confirmed successful.

3

Seafile includes a management command specifically designed to clear records for libraries that no longer exist in the system. This command likely covers the share_fileshare table because it is indexed by repo_id.

Run the following command inside the Seafile server directory (or inside the Docker container):

./seahub.sh python-env python3 seahub/manage.py clear_invalid_repo_data

Note: You can use --dry-run=true first to see how much data will be removed without actually deleting it

Manual Cleanup (If necessary)

If the management commands do not clear the specific entries in share_fileshare, it may be because the repo_id still exists in some form or the command doesn’t target that specific table for user-based deletion. You can manually identify and remove these records using SQL:

DELETE FROM share_fileshare WHERE username NOT IN (SELECT email FROM EmailUser);

See more at Clean database - Seafile Admin Manual

4

Hi Daniel,

thanks for the reply. We are in the process of upgrading Seafile from 11 to 13. I thought it was a good idea to clean up the database first. Or should we upgrade first, and delete the old records later?

Dirk

5

There are no known issues regarding database cleanup in version 11.0. Both ways are okay as long as you backup the database first.