We have just migrated all our jobs from Windows Scheduled/Service over to VC. Several of our jobs kick off using a file trigger. However, when the files are posted to our FTP, they were not executed. When we looked at the system log for VC, there were several errors: Object reference not set to an instance of an object when it was watching the directory. This is a shared directory for all our files, so there are many files that get posted to this directory. In most cases, when we manually started the jobs in VC, they pulled the file across and processed file.

We are using 5.1.6.

Any thoughts on what maybe going on? I'm attaching the log file that I pulled down from VC.

It seems like you are using the wrong username/password/domain on that shared folder. Are you using a UNC path? (\\servername\foldername)?

---

We are using UNC paths to our FTP server but this does not happen all the time. The connectivity with the file triggers is very inconsistent. On the same day, some jobs will run successfully and other won't. All of our jobs use the same set of credentials, so if one doesn't work, none of them should work.

Also, when we are manually restarting the jobs, we are getting could not start job as another user. The username, password or domain was incorrect. Please check your values. However, all we do is edit the task that executes the job, change it's credentials to another set, hit OK, then edit the task again and put it back to the correct credentials and it works fine. It seems that VC is having difficulties using the correct set of credentials.

Thanks,

Richard

That sounds strange and we should have major problems then.

1. Is the Credential selection changed or do you feel that the wrong Credential is used?
2. Are you more than one working with the Server? In this case you can see who update the Job/Task and when.

---

The correct creditials are selected when we go in and only by changing it, saving the change and then reverting it back to the correct credentials, does the jobs run manually.

I've been the primary user of it for this effort. We had been using Windows scheduler and a home grown file watcher serive to kick of those jobs. Also, I had noticed just a bit ago, when I have the client open, some of the triggers were being disabled by VC, with an error that a parameter is incorrect. Errors in the log leading up to that is when it is trying to reconnect to the directory to poll it, we are seeing logon user unknown or password is incorrect.

I don't know if the number of jobs or the number of file triggers makes a difference or not. All the file triggers are polling the same directory on the FTP server. We have 48 active jobs for this one client and of those, 25 have file triggers, 20 of which poll the same directory every 60 seconds.

It sounds like the system, either VisualCron server or remote server, is overloaded in some way.

What is the load on these servers?

Is it possible to install VisualCron and run it from another Server?

---

We don't keep active metrics on the application servers like that. When I was on there manually running the jobs, I noticied minimal CPU activity on the server (its a 4 processor server with 4 GB of RAM).

We do have VC installed on our lower environments (QA, UAT) where I was testing and was also pointing to the same UNC path, except for a different directory. Is there any known limits on how many File Trigger tasks or how many active jobs can be configured in VC? What would be some next steps in troubleshooting this issue, as it is starting to cause us issues on client deliverables. When we were using Windows Scheduler and the home-grown File Watcher service, we didn't encounter these types of issues and it was on the same hardware where we are running VC on.

There are no known limits.

I would try to:

1. Change the user that runs the VisualCron service to a user that has access to remote drives
2. Remove Credentials in file triggers

This way we can rule out the Credential problem.

---

Okay, we are making those changed. The files for our production environment, won't come in until around 2:00 am our time, so we will post our results tomorrow when we return to the office.

Okay, we've made the change to the Visual Cron service to run under the user id/pwd that has access to the FTP site and for a few days, it worked fine. Then we had to reboot the app server and since then several of our file triggers were disabled and or was receiving the Impersonation error while watching the FTP site. However, there were other jobs in that time frame that executed fine, the only different was that they were posting files to the FTP and not monitoring for files.

In addition, we are also getting the Impersonation error when we try to rerun several of the jobs. We are at the point of having to run the EXE's individually.

It seems that we are still having issues with VC in polling the FTP server, even running under the user that has full access to both the app server and the FTP server. We've had to spend a lot of developer time rerunning these jobs and we need to get this issue resolved as quickly as we can. What logs or other information do you need from us to help in getting this resolved?

About impersonation issue. If you get errors after switching the user that runs the Service there are probably three reasons for this:

1. you are trying to impersonate a user that is the same user that you run the service at (solution: deselect Credential)
2. the new user that runs the service may not have correct permissions to impersonate
3. The username or password is wrong.

About reboot problem. Maybe your settings has reverted to original ones. Is that the case?

---