Possible Issue

Question

Hi,

We had some strange issues today, and I'm not 100% sure what the cause is.

Note: I will say upfront that we were stressing the software\hardware, so this was quite an extreme day in terms of users accessing the service and server-side jobs being run.

We ran a report this morning that should have generated 'account review' approval tickets for every account in AD - 1500 of them. We had some (about 20%) approvals being generated twice, which I had though was due to human error - manaully starting the task at the same time as a scheduled start had begun (we set an attribute to prevent duplicate runs, but I had assumed that the two jobs 'overlapped' so some accounts hadn't had their attribute set by the previous job etc.

What has me confused however is that a subsequent scheduled task - that only runs against individual user objects (see below) when run *once* also produced a duplicate audit trail - as if it somehow saw two discreet instances of a single AD account.

Is it possible that we have somehow corrupted the instance of LDS that Adaxes is using?

Comments

To help us troubleshoot the issue, we need to take a look at the Activity Scope of the Scheduled Task. Can you send us a screenshot of the Activity Scope?

---

Hi - slightly complex answer to this one!

The scope *was* all objects in the domain. However, as we had the (fat finger) issue with the 'review process' executing twice and generating duplicate approval tickets, and not wanting to start firing off emails to reviewing managers that one of their accounts had been rejected because they had approved only one, we modified it.

In essense, we reviewed the execution log, identified any accounts that we had seen being denied, then added only those accounts into the scope, so that we could run the job against accounts we knew had been denied only.

A bit messy (our fault - this is a public forum after all!) but we had to be sure at the time.

To be honest we've not had any issues since, and there is no other evidence of anything going wrong, so I'm happy to put this one down to a freaky conflict due to the duplicate scheduled job run, and some of these manual tweaks we were applying to control the impacts afterwards.

Answer 1

Hello,

Since the Activity Scope of the Scheduled Task was the whole domain, we doubt that it could be triggered twice: it is a simple search in the domain to find all users, where any 'overlapping' is excluded. We think that the issue is somewhere in your Adaxes configuration, where the change of the properties could be triggered by some other Scheduled Task or Business Rule.

As far as we can understand, by tweaking the Scheduled Task and its Activity Scope you managed to remedy the issue. We can suggest that you keep tracking the issue and see if such behavior reappears in the future. If it reappears, we'll dig deeper into your Adaxes configuration to identify why such behavior occurrs.