topic Re: Check for broken jobs in Administration and Deployment

Check for broken jobs

titan31 — Thu, 19 Sep 2019 10:43:27 GMT

Hi,

I'm running into an issue that after a migration from a Windows to Linux system, some DI jobs have broken (datasets missing, joins being broken, mappings disappearing etc).

So far, I've noticed these as I'm going through jobs that I need to change paths on, but wasn't intending to go through every job, however as the amount I'm noticing has increased, I need to be able to quantify the problem easier.

Is there anyway to programatically check for broken jobs in metadata or anywhere, or is going through jobs one by one that only way?

Thanks,

Re: Check for broken jobs

alexal — Thu, 19 Sep 2019 11:32:10 GMT

@titan31 ,

I assume your DI jobs generating SAS logs, right? You can just do something like "grep -i error Logs/*" to see which one has errors.

Re: Check for broken jobs

titan31 — Thu, 19 Sep 2019 11:42:39 GMT

They are, yup, but need to run them for those logs to exist. Was hoping there might be a way to query the actual metadata or code to see are there errors or warnings.

Re: Check for broken jobs

LinusH — Thu, 19 Sep 2019 12:55:33 GMT

I would really love if there was a tool that could compare complex metadata objects, like jobs.

ASFAIK it doesn't exist, not an official one at least.

What you could do is script a comparison program between deployed jobs, where you need to filter out metadata id's, timestamps etc so you can focus on the real differences.

Re: Check for broken jobs

titan31 — Thu, 19 Sep 2019 13:06:52 GMT

Ya, doesn't seem to be anything there. Quite painful to check each job individually so was hoping there might have been