Introduction:
The engine log shows for the collect job how long it takes to process an asset before moving on to the next. Sometimes this can take up to 10 minutes while other assets only take few seconds to process.
Question:
What reasons exist for this behavior?
Environment:
Engine Collect task processing inventory uploaded to the scalability server’s sector by agents.
Answer:
There can be several reasons why processing inventory for one machine takes longer than for others. To name a few:
· · agents by default send up delta inventory, which is what changed since they collected information the previous time. This delta inventory can be almost nothing for one machine while for another machine many things changed or it is a whole new agent reporting all of its inventory for the first time.
· · there may be differences in agent configurations: one agent is configured to run e.g. file scans and/or collect many wbem classes which typically causes many rows to update in sql while another agent may be configured to collect much less.
· · the domain manager may be busy with other things like building job containers or processing job output while at other times there is little workload. If the system is busy in general, time to process inventory takes longer than during quieter times.
· · the same goes for sqlserver: when it is processing requests from other ITCM components it takes longer to process update information it receives from the engine collect task than at quieter times.
In general, if the engine is capable of processing all information in the sector in due time, there is little reason for concern. You can verify by checking the number of files in the sector on the scalability server: Program Files (x86)\CA\DSM\ServerDB\Sector\Collect\00000001. If the number of files in there remains stable over time and does not increase, the engine does not seem to have troubles processing everything the agents uploaded