We have an NT box (SP6a) ALR Revolution 6X6 with a memory leakage problem. It runs Exchange 5.5 SP4 (as the IMC only) and is the print server for a large office building also. We began getting high counters for Pool Nonpaged Bytes yesterday and found that above a value of 60 no one could print. Normal should be below 20. Regular resource usage is low: processors, memory, disk space are all barely breathing hard. Counters for Pool Paged Bytes are also very high. IMS Traffic counters for Exchangeare also high (Messages Entering MTS-IN and Leaving MTS-OUT). Other Exchange counters very low. Spooler service takes a long time to come up…but system resources are not taking a hit.
Service packs rerun, RAM checked and re-seated. 512K RAM, pagefile 1046 MB (523-768MB on both partitions), six Pentium Pro 200 processors, disk space to burn. Nothing helps. Nothing on TechNet either: two docs touch on the issue but not the exact problem, and provide “fixes” that have not been fully tested…Q272194 and Q293841.
Relevant errors in Event Viewer: event IDs 2019, 2021, and 9316.
Running McAfee NetShield 2000 vdef 4.0.4135, engine 4.1.40. Have ruled out virus attack.
How to deal with this resource problem? We want to free up memory and stop the leakage. This server is critical!
