Matthias Schmitt <freak002(at)mmp(dot)lu> writes:
> I am in trouble with more and more deadlock problems. We are
> programming a web application with multiple users editing content at
> the same time. Multiple times a day PostgreSQL runs into a deadlock,
> which can only be resolved by killing some of the clients. Here is an
> example from the pg_locks table:
All of the rows you showed us have granted=t. No blockage is evident,
let alone any deadlock.
regards, tom lane