Just a quick question here. Our company is looking to replace the built-in Windows Task Scheduler. At the moment all we have is a long list of jobs that run periodically; typically every minute or every 5 minutes, although some run every day.

There is a great deal of functionality and flexibility in VC that we could leverage, particularly regarding process restarts (if hung) and the sending of email notifications in cases of success/failure. However, our main reason for dumping Task Scheduler may not be addressed in VC.

We have had a problem in the past with Scheduled Tasks just dying. The whole service would get stuck, and as a result, NONE of the many jobs defined would run. We have some external notification processes set up, and so we would catch this fairly quickly, but there have been cases where the jobs did not run for a considerable time, due to the tasks service getting stuck.

The initial plan was to go with a competing product, which installs each task as a separate service. In this way, one service getting stuck or going down would not bring down any other jobs. This is an approach we thought would be duplicated in VC, but it appears not.

Is there any way around this? The price, service, and feature set in VC is unparalleled, and I really feel we could use it to our great advantage. But before making the purchase of several licenses, we have to address this very serious concern. Has anyone had issues with the VC service ever getting stuck? Are there any provisions to restart it automatically? Is the Task Scheduler service failure a known issue with Windows 2000?

Any and all feedback is appreciated; from VC support staff and from the general user community. Thanks!

There is a lot of internal error handling in VisualCron. If there, for any unknown reason, an unhandled error should occur VisualCron is capturing that error as well and restarts itself within 10 seconds after error. What happens before that is that we log the error to the log file log_serverDATE.txt.

If anyone has received such a restart it is important that we get this log information. Then we can make this error handled.

What we plan to introduce in VisualCron is an internal Notification/Event. This internal Notification can warn to an email or other source if anything, against odds, should go wrong.

I have no experience of W2000 task scheduler so I cannot give any comment about that.

All tasks are executed in new threads so there is no risk that a certain Task locks VisualCron unless you use the shared system resources in a bad way. Currently we have no limits on a Task besides that you can set priority in the Execute Task of the started process.

Let us know if you have any further questions.

---

How does VC "restart itself?"

What I'm wondering is what would happen if the main VC service were to crash. I assume the service is set to "Automatic" so that it starts when the server boots up and the OS gets loaded. If the whole service crashed, how could it get restarted again? We could always write a script that polls to see the process is running (and restarts it if necessary), but if it is running but in a strange error state that wouldn't do anything.

I realize this scenario may seem hypothetical and strange, but our main reason for moving from the built-in task scheduler is this weird global outage issue (which is always easily resolved by restarting the service).

There are two types of exceptions that are thrown when an error is not handled. Unhandled and Threadexception. It is possible to subscribe to these events and handle this. In our case, when this occurs, the event is triggered and we restart the service process.

It is also possible to control this at server level as you describe but that may conflict with our solution.

---