marcelo Cortez <jmdc_marcelo@xxxxxxxxxxxx> writes: > the pg_statistics show too many lock table ( fron my > app.) and all of them is in waiting state. > the only way to go out from this situation is restart > the cluster. Surely a few query cancels would be sufficient. > where is the way to investigate lock's table tree ? The pg_locks view should tell you who's hogging the locks. regards, tom lane