NullReference exception when running a Console App (with WebJob SDK 2.0.0-alpha-10349 bits) locally and disconnected from Azure Storage

Hi Guys,

I get a NullReferenceException when running a Console App (with WebJob SDK 2.0.0-alpha-10349 bits) locally and disconnected from Azure Storage.

        JobHostConfiguration config = new JobHostConfiguration();
        config.HostId = Guid.NewGuid().ToString("n");
        config.DashboardConnectionString = null;
        config.StorageConnectionString = null;
        config.UseTimers();

        JobHost host = new JobHost(config);
        host.RunAndBlock();

I have a Functions.cs with the following:

    public static void TimerJob([TimerTrigger("00:00:05")] TimerInfo timerInfo,
                                TextWriter log)
    {
        log.WriteLine("Scheduled job fired!");
    }

From some of the comments I read in this pull request (PR) here (https://github.com/Azure/azure-webjobs-sdk/pull/639), I guess you saw it comming.

Any thoughts on how to deal with this? I have some ideas and I would be glad to contribute with a proper fix + PR.

Cheers, Carlos Sardo

[mathewc]

TimerTrigger requires a storage account to operate (it stores schedule information and singleton lock info in blob storage). While the null ref exception you mention might not be proper error handling (we can fix), I don’t expect what you’re trying to do above to work.

While certain triggers like timer could be made to work conditionally in some cases w/o storage, there would be caveats (e.g. the singleton lock is how we ensure only a single instance of the schedule is running across scaled out instances) it hasn’t been a priority because most users of the SDK are using storage.

[mathewc]

This is a really old issue against the 2.0.0-alpha prerelease. SDK is now at 2.3.0. I don’t think this is an issue anymore for 2.3.0. For v3 (3.0.3) the startup model has completely changed (.NET Core) and again, I don’t think we have this issue anymore.

If it turns out this is still an issue we can reopen.