Error running local jobs as if remote agent

We're having jobs refuse to fire with the following error message

%jobname% Failed to execute remote job. Agent LOCAL connection refused

It's as if it thinks local is a remote agent profile? There are none defined and no remote jobs either; it's only happening on one job in particular but I have rebuilt that job from scratch when the issue came up yesterday and the prob came back this morning. I've managed to get it working again this morning by opening the job, changing the host entry to LEAST BUSY and saving the job, and then changing it back to local and saving.

Any ideas? Might come back tomorrow.....

Yes, it sounds like it wants to find and connect to a remote agent called LOCAL. Do you run any scripts that modify job properties dynamically? Are you sure you are fixing the right job? I mean there is no other job with the same name that is responsible for generating the error message.

There's certainly no other job with the same name; the only interaction with other jobs that it does is use Jobrun to run the next job in a chain.

Do other jobs in chain point to LOCAL or any other remote agent?

No, there's no local remote agent profile and the jobs don't point to it.

The child jobs don't error.

And they are all local jobs not remote

Does the job trigger any notification actions, for example, actions that execute scripts or other jobs? Any script library calls resulting in JobRemoteRun calls or similar?

Did you recreate the job from scratch or created it using "copy job" method?

Recreated it from scratch. The only notification the job does is RaiseError in JAL, no jobremoteruns or anything.

If you restart the scheduler and still get this error then it must be coming from some other place.
If it still occurres, please attach screenshot of job properties as they appear in the right side of job explorer screen.

Ok if it reoccurs I will fix it again and get screenshots.

Ok, so this job fails every time it is run with the error

23/07/2007 17:23:51 2 210 0 Run log parser Failed to execute remote job. Agent LOCAL connection refused.

Here are the job properties

I've tried rebuilding it from scratch, and I've tried the trick that worked on the other job which was doing this (setting host to least busy, save, set host to local, save). The other job has not failed again.

I'm running the new JSP web console build. I can run other jobs with it.

I think we are getting somewhere. The property for the host should be blank. If you use the GUI to create or modify a new job, the property will appear as blank or as "Local" which is the same. I wander if it has anything to do with regional settings on your system. When you modify the job using web console the property gets changed to "LOCAL" all upper case, and the scheduler doesn't recognize it anymore as "Local" trying to find a non-existing agent profile with this name. I will do some experimenting with different settings and see if I can reproduce this.

Thanks; FYI the web console does not allow me to remove the host property in favour of a blank and neither does the local console; it's not a text field only a dropdown box.

And regional settings on the server are set to New Zealand English with a US keyboard layout.

I can reproduce it too. It occurs after I edit job definition using JSP based web console. I will try to find a place where it can be changed.