#3635 Don't let back end code use the sdap_id_op operations directly
Opened a year ago by jhrozek. Modified 6 months ago

This is a follow up to issue #1507. Currently the sdap_id_op code has several issues (such as touching the online/offline status directly). I don't think each LDAP search issues from the back ends should need to deal with the sdap_id_op requests and retries, but there should be a wrapper around those which evaluates the error codes.

We also need to get rid of the ignore_mark_offline flag, instead the id_op should return an error and let the wrapper deal with it, usually by setting the domain offline or inactive.

The idea is to put all the _retry logic into one request. Given the connection is cached, I believe it can be put into ldap_search where you either connect or reuse the connection if available and try to reconnect if it fails.

Metadata Update from @jhrozek:
- Issue set to the milestone: SSSD 2.0

a year ago

Metadata Update from @jhrozek:
- Issue priority set to: major

11 months ago

Metadata Update from @jhrozek:
- Issue set to the milestone: SSSD 2.1 (was: SSSD 2.0)

6 months ago

Login to comment on this ticket.