Hi,

I am getting the error "FATAL:  sorry, too many clients already" in in my local Galaxy instance in the POSTGRES database log.

I have set max_connections = 1000 in already in "postgresql.conf" and STILL get this error...

I was running a Flowcell with about 96 samples and a workflow with about 50 steps each, but I never had a problem before when max_connections was set to 100 and did not change the database...

I do have a cluster that is much faster so i guess there may be several connections going to the database, but maybe there is some leakage in the galaxy code that is not closing the connections?

It really screws up the galaxy instance and jobs are just in limbo and never finsihed (in the database)...

Any ideas?

I adjusted 

already but maybe I set those too high? 

What is a normal setting for max_connections in galaxy production servers?

Thanks,

Thon