Knock down another shared vlan bug; this one was caused by the cached
DB objects we use in the libraries. We hang onto those objects forever in the expire daemon, but they are stale in the case of shared vlans which are being messed with elsewhere. Rather then track down all the caching bugs, I changed the expire daemon to call the existing cleanupslice script, avoiding all cache problems.
Showing with 5 additions and 8 deletions