Grouper Users - Open Discussion List

[Andrew Morgan <>]

On Fri, 25 Sep 2015, Andrew Morgan wrote:

> On Fri, 25 Sep 2015, Waldbieser, Carl wrote:
>
> > At Lafayette College, we have been running the Grouper Loader as a system 
> > service from /etc/init (according to the host's man pages, "init - Upstart 
> > process management daemon"). This has typically worked well for us, though 
> > I was a bit surprised today when my Grouper status URL reported that a few 
> > new SIMPLE_SQL jobs I recently added had not run in a couple days. The logs 
> > showed no errors, and if I ran the jobs from GSH, they ran without issues.
> >
> > I finally ran a report from the database and realized the jobs were not 
> > recorded as having been run.
> >
> > I tried changing one of the job times, and I could see in the grouper logs 
> > that a session started and ended at the appropriate time, but there was 
> > still no log entry. Finally, I rescheduled the job *and* stopped and 
> > restarted the Loader process.  Now the job was picked up and recorded in 
> > the loader history table.
> >
> > Based on these observations, I am assuming that an existing Loader session 
> > won't see newly scheduled jobs.
> >
> > I was wondering if I am running the Loader in the expected manner, and it 
> > is just an operation step to bounce the service to get it to recognize new 
> > jobs, or if it was intended to be run intermittently or perhaps something 
> > else.
>
> Carl,
>
> My understanding is that anytime you modify a loader job, you need to restart 
> the loader daemon.  That's certainly how it works in my installation!  :)

I should have added this reference:

  https://spaces.internet2.edu/display/Grouper/Grouper+Loader+add+new+job

Towards the bottom, it mentions that the loader needs to be bounced.

        Andy