Hi You said : “We have configured 25% RAM as shared buffers,50% as effective cache.we have configured 50% RAM for Huge page settings” So I can say that you are wasting about 25% of your RAM! Apart from shared buffers (and dynamic shared memory), no other memory areas in PostgreSQL use huge pages. Regards Michel SALAIS De : Ramakrishna Reddy Nandyala <rkreddy.n@xxxxxxxxx> Hi Team System details We are using pg-16.8 on rhel 9.x RAM-96GB CPU-16 We have configured 25% RAM as shared buffers,50% as effective cache.we have configured 50% RAM for Huge page settings Max connection set to 2000 Work-mem -4 mb Maintenance work mem set to 1 GB Issue details 1) we have an application that selects the data and will process in application and try to insert the data into an other table During this process we are seeing so many temporary files been generated by the system,after some time we are seeing the application Connection is accumulating the memory at OS Level after reaching ~50%,we are getting Out Of Memory error and Kernel is killing the top memory consumer process ,here it is killing Postgres process that is doing the entire process.while automatic Postgres Service restart ,system status showing as activating and it is coming to Active state only after clearing of all the temporary files crated during the query running.The clearing of temp files are very slow such that it is causing unplanned outages to the system how to fasten the Temp files removal? How to get rid of this situation? Any suggestions on the configuration to avoid the temp file generation Thanks in advance Ramakrishna |
