The process AeXSvc.exe has begun to consume large amounts of memory, even bringing the system to 99% memory usage. If the Altiris Services are restarted the memory consumption goes down, but begins to trend upward, sometimes over the course of a few days, and sometimes quickly over the couse of a few minutes or hours.
1. One item that can cause high memory usage in AeXSvc.exe (Altiris Service) is when trace logging is left on after troubleshooting. When trace logging is enabled, memory cleanup is delayed for troubleshooting purposes. Trace logging is only meant to be enabled for up to a few hours while troubleshooting. Enabling trace logging also will roll logs more often and cause extra performance overhead on production server. Running the SMP with trace logging always enabled will not always result in high memory but has a tendency to occur more often when processing a large number of NSE files (inventory times) and logging all of the trace level string information.
2. There have been cases where a data source used by a Connector Rule has been configured to export data to a csv or text file to be used for reimporting. Within the data source object is a box labeled "Allow overwriting of existing data". This box, if unchecked, will cause the import file to become larger every time the rule is run. At some point the file will reach a size that cannot be processed efficiently (or at all) when the import rule runs. As a result the memory requirements to attempt running the rule double and tripple a it gets larger.
1. Disable trace logging.
2. Look for any Connector rules that may be importing data but not allowing the the previous data in the import file to be overwritten, and evaluate if it really needs to be configured this way. The data sources can be found in the console in the following path:
"Settings > All Settings > Notification Server > Connector > Data Sources"
Symantec Management Platform 7.1 SP2 MP1
Imported Document ID: TECH203994
Subscribing will provide email updates when this Article is updated. Login is required.