#57 Implement psearch query fallback mechanism: parallel normal queries
Closed: Invalid None Opened 10 years ago by pspacek.

For plugin startup and for error handling situations is necessary to have fallback from psearch to normal LDAP query mode (and then back to psearch).

If some problem inside psearch occurs:

  • switch to normal query mode
  • flush and repopulate caches from LDAP DB
  • then switch back to psearch mode

While plugin starts:

  • psearch dumps whole LDAP DB to plugin's cache.
  • In meanwhile use normal queries to answer client requests.

Dumping can be be pretty long operation, if database is large. It's something between several seconds to two minutes or so, it depends.

Current plugin doesn't have negative cache, so described behaviour will be realized by implementing ticket #44.

This ticket should be re-opened after negative cache implementation.

This is no longer valid: With [[BIND9/Design/RBTDB]] there is no concept of 'cache' anymore.

Metadata Update from @pspacek:
- Issue assigned to atkac
- Issue set to the milestone: 0.0 TRIAGE

5 years ago

Login to comment on this ticket.