* removing db from config and reload; should work - kill connections
* removing user should work - kill connections
- * suspend_timeout - drop stalled conns
+ * suspend_timeout - drop stalled conns, long logins
* when dest db does not exists, stalled logins can halt suspend
fix with suspend_timeout?
* some preliminary notification that fd limit is full
* SHOW TOTALS that shows stats summary (as goes to log) plus mem usage.
+ * new states for clients: idle and in-query. That allows to apply
+ client_idle_timeout and query_timeout without walking all clients
+ on maintenance time.
* ..hopefully.. make -R less scary
* ..maybe.. make plan cache work
== Minor features ==
* drop_on_error/keep_on_error - if released conn is in error state,
- then issue rollback and keep it
+ then issue rollback and keep it
* show stalled; command - like show sockets; but filter only active ones
* keep stats about error counts