Please implement a DatabaseStartFailed event

[Breck_Carter]

Please implement a DatabaseStartFailed event that gets control when the engine is in the process of failing to fully start a database because of some startup error. This event should have access to the same diagnostic information that is written to the -oe text file.

I realize that there may have to be (severe?) limitations on when/if this event gets fired, and what actions it can take, but here's the business case...

When a service fails to launch, or an embedded database fails to start, it would be great to be able to send an email containing the -oe information... otherwise, such a failure to start can easily go unnoticed.

As people (like me) push more and more critical application functionality down into web services running in the database engine, diagnosing problems becomes more and more of a SQL issue... in some cases, there IS no "client" other than a slack-jawed knuckle-dragging web browser, it's up to the server to do all the heavy lifting.

[justin_willey]

Perhaps there needs to be some sort of "engine start timeout" setting to determine when the start is regarded as having failed. The "Start Pending" message has to be one of the most irritating things out there - is it just taking a long time or is there a fatal problem.

[Breck_Carter]

Where do you see "Start Pending"? Maybe it's the lack of coffee but I'm not familiar with that message... although I am very familiar with long startups, sometimes a longgggg recovery after Something Bad happened.

[MCMartin]

Start Pending is displayed in the service control of Windows

[Breck_Carter]

@Martin: Oh that Start Pending... yes indeed that is truly irritating. FWIW I always recommend -o and -oe options, especially for service-started engines... they sometimes help.