[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] Question about Quill

I am running Quill++ and I am using the setting QUILL_MAINTAIN_DB_CONN = False. My problem is that the Postgres server (not the same as my CM), which has 4 CPUs, runs at 95% CPU load. If I try changing QUILL_POLLING_PERIOD this to 120 sec, it does not seem to make a difference. The other thing I notice is that on the postgres server, the connections do not seem to close and I would think I would be able to see this using such a large polling period.

I have approximately 115 cores (50 machines), so this is not a huge pool. All machines are Windows including the servers. Postgres has approximately 50 connections open at all times each utilizing 0-5% CPU usage and 4MB memory.

Is there something in the in the condor logs to determine whether the connections are actually being closed.

Also, Does each machine create a connection or is it each core?

Does anyone have recommendations on scalability with regard to the number of machines/cores and the number of CPUs required for using Quill?

Thank your for the help,