I am evaluating this software. This post is concerning and I'm curious if this was ever fixed. Thanks.

Jason

I am evaluating this software. This post is concerning and I'm curious if this was ever fixed. Thanks.

Jason

Previously, we believed this was caused by VMWare environment. Now the customer says he is having the same problem in normal environment. We have not been able to reproduce this - if there was a simple way to reproduce this we would be happy.

What we know is that a bluescreen occurs. As VisualCron is written in .NET it is almost impossible to get a bluescreen as methods are not called on that level.

The problem Task in this case is the Stop Job/Kill Task. It seems like it happens when a system criticial process is killed which explains the bluescreen. So, there seems to be something specific about that process and it's child processes - but we are not sure. This problem is only related to these Tasks.

---

vcdkk,

could you try to call the Windows command "taskkill" in an Execute Task instead of using the Stop Job? We would like to rule out the relation between the Stop Job and the crash itself.

If it crashes again using taskkill we would like to see the dumpfile again.

thanks

---

Hi again.

I have setup my enviroment using Task Scheduler in Windows 2008. I'm using taskkill to stop the running processes. This has run for a week now without any problems.

Hi again.

I have setup my enviroment using Task Scheduler in Windows 2008. I'm using taskkill to stop the running processes. This has run for a week now without any problems.

Thanks for the feedback. But what about running "taskkill" from VisualCron instead of "stop job"? (please note that during such a test you cannot use any other method in VisualCron that stops a Job including timeout functionality).

If we know that taskkill works together with VisualCron then the problem is pretty isolated for further digging.

---

I cannot use more time on testing. You probably could trigger the problem in a test enviorment youself.

Stig

I am attaching a new version. What is different in this version is that we do not try to kill a whole process tree when killing the process. Maybe that is the culprit.

---

I cannot use more time on testing. You probably could trigger the problem in a test enviorment youself.

Stig

I understand. We have really appreciated the patience so far. This has been a very frustrating issue for both as we have not been able to reproduce the problem here.

What is the current situation - are you not using VisualCron at all or are you using the Task Scheduler to just stop processes?

Perhaps the reason why we have not been able to reproduce it is because of the nature of your process with the combination of how we killed child processes (not in last uploaded version).

---

The stable situation is:

Only Task Scheduler

1. Task calls .cmd file which starts 10 processes
2. Task calls .cmd file using killtask to kill the 10 processes.

We found a way to reproduce the problem. However, in our VirtualBox environment it just locks up but we guess this is the same as the bluescreen you saw.

We did a reinstall of our Windows 2008R2 environment (which we could not reproduce the problem in before) and installed all updates. It was in this reinstall we could reproduce the problem.

We first tried the latest official version and it hanged within 10-20 minutes - sometimes earlier. Then we tried the version 5.7.6-2 which we uploaded as attachment here. This version has now ran for about 1 hour without any hanging.

We will update the result later but it seems like we have found the issue.

If this is the issue the history is that we had a way to kill processes once. Then a customer found that, if a process (for example .bat-file) is spawning child processes, the child processes are were not killed. What we did then was to extract the child processes and issued a kill on those as well. This version has been running for about a year and we do not know why it causes an issue in this case. In 5.7.6-2, killing of child processes is turned off. If this is the cause, we will make this "killing of childprocesses" optional.

In your case, after examining logs, we never found that any child processes were spawned. Also, what was strange that the crash could come up anytime and did not occur "within" VisualCron as we could not capture any unhandled exception. It is almost impossible for a .NET application to cause a blue screen. But, if we examine exactly what we do (and what causes things to go wrong) is that we query the WMI. A Windows "database" for information. We make a query for child processes for a specific process. While .NET has an abstract layer for this we assume that something goes terribly wrong in the core of WMI. Perhaps it gets overloaded, perhaps it is broken - we don't know yet.

As we said we will update with result later but we really would appreciate if you tested this one more time.

[EDIT] It has now been running without problems for 8.5 hours and we are closing the virtual server for today.

---