How FNMS Reader operation in detail?

[longdinh]

Hi everyone,
I am wondering how FlexNet Reader is actually working to import data into Staging tables. I am experiencing an issue with the customer which is related to the time FlexNet Reader does import data into Staging.

My client is scheduling the FNMS agent inventory every second day. In the day the FNMS agent was scheduled to collect - it takes 4 hours for the step to import data into Staging. However, on the quite day - the FNMS agent is not scheduled - it takes only 1,5 hours. There is 2.5 hours difference between these days.

I know the FlexNet Reader does not take that long to complete the job to import into Staging, but it seems that there is an issue with this client. Most of the time-taker tasks are related to the UnixFileEvidence or UsageFiles.

With my observation, my hypothesis is that when the inventory file is uploaded to the FNMS application server, the files are written into the queue. At the end of the day, the FlexNet Reader will read and import all the inventory files (.ndi) from the FNMS agent coming for that day and clear the queue. The next day, if inventory files are coming, the whole process will be done again. Because my client schedules the inventory collection to run every second day so the next day, it will be less data coming in -> the processing time will be lower.

I don't know how FlexNet Reader operates in detail, but I believe it does import/process all data it receives - not only the delta between these days. Is my understanding correct? Does anyone give me some insights into this issue?

Thank you for all your answer and response. Much appreciated!!!.

Regards,
Long Dinh

[ChrisG]

When the import "reader" process runs on the batch server (and not from a beacon), there are many steps in the process that are optimized to only retrieve and process data for computers that have an inventory date that is newer than what was seen at the last import.

So that would match your observations here: on the days when inventory is updated for most computers you will naturally see the import "reader" process take longer as it has more data to work through.

[longdinh]

Hi Chris,

So it means that it does not only process the delta right? It will process all of the data coming from that day and that is the reason why process is taking longer. Am I understanding that correctly?

Thank you.

[ChrisG]

No, it's not entirely correct to say "it does not only process the delta". It is the opposite of that: many parts of the import "reader" process (when executed on the batch server) will only operate on the delta since the last import. This is done by the database queries that are executed during the process containing conditions to ensure they only operate on data associated with computers where the inventory date has changed since the last import.

This means that the reader process is likely to take longer on days where more computers have uploaded fresh inventory compared to days where fewer computers have uploaded inventory, which (if I understand you correctly) is what you have observed.

[longdinh]

Thank you Chris for clarifying. I understand how FlexNet Reader works now.