Comment 1 for bug 1862315

dhcp agent's sync_state() function is supposed to get a write lock and proceed, and seems it waits to get the lock for a long time because of many regular update events coming to the agent's queue (due to many concurrent port create/update/delete operations).
These update events are taking read locks.
Write lock is supposed to be of higher priority than read lock, but for some reason it doesn't work.