Deployment Solution

I have a simple job that does: 1) partition 2) install OS. This is for unknown computers

So the unknown computer goes into PXE and the job starts; I can see the status of the job on the console. I see that the PECTAgent runs and checks for driver (once without DB and then with the DB). I see the Partition task complete successfully on the console. Then the PECTAgent is done and the Window Setup starts, Expanding windows files, etc.. This is when I see another PECTagent pop up and do the EXACT same thing as before. I check the console and see that now the partition task is showing as FAILED!! Looks like the same job just started in the middle of running :( The PECTAgent keeps running doing the driver precheck forever!! Sometimes, the PECTAgent will start while another one is ongoing too. I have attached a screenshot.

I am using a custom Unattended file, but I doubt that has any affect on the above behavior. Has anyone else seen this? Any pointers appreciated!

Update: The windows installation does go through, although right after the first installation completes, it tries to install again, but the answer file has been deleted and it fails. After this, the reboot brings up the windows properly. Attaching screenshot of the logs folder that shows 2 nearly back to back jobs were ongoing.

Definitely not a normal behavior.

But the problem is purely that second pectagent is starting.

Some search needs to be done to find out what is calling pectagent for second time.

Once I remember I saw for one customer they had a call twice in bat file for pectagent thus pectagent was starting two times for them and exactly same behavior that job/task was running twice.

I'm attaching the runagent.bat file as a text file. Since I have seen the pectagent get started while another one is ongoing, it seems to me that the runagent.bat is getting called multiple times. Any ideas on what might be causing that or how to debug it? Thanks.

Attachment(s)

runagent.txt   2 KB 1 version

Also attaching how the console of the job status changes. Any ideas anyone?

Indeed runagent seems to be clean, no double-call

startnet.cmd might have also a double-call.

overall I fear only manual tracing will reveal where the double-call is hidden.

How can I go about doing such a trace?