Invalidate event cache on duplicate key error

[neSpecc]

The OOM is caused by infinite recursion in handle(). Here's what happens:

1. getEvent() queries MongoDB, finds no document, and caches null (since CacheController treats null as a valid value — only undefined triggers a cache miss)
2. Since existedEvent is null, the code enters the "first occurrence" branch and calls saveEvent()
3. If another worker instance already inserted this event, MongoDB throws a duplicate key error
4. The catch block recursively calls this.handle(task) to re-process as a repetition
5. But getEvent() returns the stale cached null instead of fetching the now-existing event from the database
6. The code thinks it's still a "first occurrence", tries to insert again, gets another duplicate error, and recurses again — forever

[Copilot]

Pull request overview

Fixes an OOM caused by infinite recursion in GrouperWorker.handle() when a MongoDB duplicate-key insert happens after getEvent() has cached a null (cache hit on retry prevented re-fetching the newly inserted document).

Changes:

• On duplicate-key insert errors, computes the getEvent() cache key and deletes it before retrying handle(task).
• Ensures the retry path re-reads the event from MongoDB instead of using a stale cached null.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

@neSpecc I've opened a new pull request, #537, to work on those changes. Once the pull request is ready, I'll request review from you.