Continuously running vs. repeated deployments
I have a self-service system on my domain that I would like to run every 30s-1min to check for new requests. Are there performance issues with:
- having the script run continuously on a single deployment,
vs.
- setting up a schedule to run the script repeatedly at short intervals?
0
Comments
I would recommend having your script call PDQ once it receives a request. PDQ has a command line interface:
The script does call PDQ to perform the actual software deployment - the requests are logged on a different server, and the script pulls those requests and fulfills them. I was thinking of putting the script into a package so that I could manage it (start, stop, check for errors, etc.) from within PDQ as opposed to being a Windows background task.
So you want to build a package that contains a script that checks a server for requests and attach that package to a schedule that runs every 30 seconds? Do I understand your layout correctly?
Yes. The question is whether that is a better idea than having the script run itself every 30s on a manually deployed package.
I think it would be best to write the script with a built in timer and use Task Scheduler to launch it at boot.
While ( $true ) {
Do-Stuff
Start-Sleep -Seconds 30
}
Running it from a single deployment wouldn't work well because they time out after a specified time period. Triggering it from a schedule would work, but you are going to clog up PDQ with 2,880 extra deployments per day.
That's the info that I was looking for. Thanks!