[QUESTION] Object synchronization method was called from an unsynchronized block of code

[jokker23]

I'm getting this error intermittently while running with ConnectionType.Shared and don't with the default.

System.ApplicationException: Object synchronization method was called from an unsynchronized block of code.

   at System.Threading.Mutex.ReleaseMutex()
   at LiteDB.SharedEngine.CloseDatabase()
   at LiteDB.SharedEngine.Upsert(String collection, IEnumerable`1 docs, BsonAutoId autoId)

System.ApplicationException: Object synchronization method was called from an unsynchronized block of code.
   at System.Threading.Mutex.ReleaseMutex()
   at LiteDB.SharedEngine.CloseDatabase()
   at LiteDB.SharedDataReader.Dispose()
   at LiteDB.LiteQueryable`1.ToDocuments()+MoveNext()
   at System.Linq.Enumerable.SelectEnumerableIterator`2.ToList()

I tried researching this but haven't come up with any ideas.
Does anyone have any suggestions on why I'm getting this?

[JensSchadron]

@jokker23 Could you please provide us with more information like, which version of LiteDB you're using, OS, net runtime, et cetera 🙂

[jokker23]

Certainly,

LiteDB: v5.0.2
Win10
.NET Core 3.1

Also using
Newtonsoft.Json 12.0.3
RabbitMQ.Client 5.1.1
Microsoft.Extensions.Caching.Memory 3.1.1
Microsoft.Extensions.Configuration.Abstractions 3.1.1
Microsoft.Extensions.Configuration.Json 3.1.1
Microsoft.Extensions.DependencyInjection 3.1.1
Microsoft.Extensions.Logging 3.1.1
Microsoft.Extensions.Logging.Console 3.1.1
NLog.Extensions.Logging 1.6.1
And a custom SDK dll

Is there other information that is interesting?

[lbnascimento]

@jokker23 Could you provide a sample code that we may use to replicate the issue?

[lbnascimento]

@jokker23 Could you test with the current master? The transaction model was entirely rewritten.

[jokker23]

Yes sorry I will try to accommodate both requests as soon as possible. I'm currently on overload making VPN tunnels for home workers, sorry for the delay.

[alonstar]

I got the same error, too.

System.ApplicationException:  Object synchronization method was called from an unsynchronized block of code
於 System.Threading.Mutex.ReleaseMutex()
於 LiteDB.SharedEngine.CloseDatabase()
於 LiteDB.SharedDataReader.Dispose(Boolean disposing)
於 LiteDB.SharedDataReader.Dispose()
於 LiteDB.LiteQueryable`1.<ToDocuments>d__26.MoveNext()
於 System.Linq.Enumerable.WhereSelectEnumerableIterator`2.MoveNext()
於 System.Collections.Generic.List`1..ctor(IEnumerable`1 collection)
於 System.Linq.Enumerable.ToList[TSource](IEnumerable`1 source)

[lbnascimento]

@alonstar What version are you running?

[alonstar]

@alonstar What version are you running?

5.0.4 and .net framework 4.7.2

[mwlpeeters]

I've the same problem with 5.0.5 on .Net core 3.1 when multiple API calls are trying to connect to the same database.

Update 1:
In my application I use Autofac and configured it to create only a single instance of the LiteDB database (wrapped in a provider, so actually the provider was created once and the internal LiteDB as well). The connectionstring contains 'Connection=shared'.
In this situation I get the same error as described above.

I've changed my Autofac configuration so it will always create a new provider (with internally a new LiteDb instance).
This solved this specific problem, but now I'm running into file in use problems?

Update 2:
Well I now reverted the my Autofac configuration into what I initially had. So a single instance of the provider, with internally a single instance of LiteDB. Changed the connectionstring from 'Connection=shared' to 'Connection=direct' and this seems to be stable after a couple of tests.

[nightroman]

I can reliably reproduce this exception (not necessarily same issue) with the
following test ran using "Debug test" mode (RMB \ Debug). "Run test" seems to
be usually working. If it is still not reproducible, try to change the number
in Sleep(50).

[Fact]
public void Just_DataStream3()
{
    using var mem = new MemoryStream();
    var settings = new EngineSettings()
    {
        DataStream = mem,
        Filename = "Just_DataStream3" //! see https://github.com/mbdavid/LiteDB/issues/1614
    };
    using var engine = new SharedEngine(settings);

    void worker()
    {
        using var db = new LiteDatabase(engine, null, false);
        var col = db.GetCollection("test");
        for (int i = 0; i < 100; ++i)
        {
            col.Insert(new BsonDocument() { ["i"] = i });
            System.Threading.Thread.Sleep(50);
        }
    }

    var thread1 = new System.Threading.Thread(worker);
    var thread2 = new System.Threading.Thread(worker);
    var thread3 = new System.Threading.Thread(worker);
    thread1.Start();
    thread2.Start();
    thread3.Start();
    thread1.Join();
    thread2.Join();
    thread3.Join();

    using var db = new LiteDatabase(engine, null, false);
    var col = db.GetCollection("test");
    Assert.Equal(300, col.Count());
}

[nightroman]

Can you repro the problem with the above code? Please let me know. If not then I will continue the search for repro cases. I think problems might be related to #1614 which is not well repro, too.

[lbnascimento]

@nightroman Yes, I was able to reproduce the problem.

[Sn3b]

Any updates? Having the same issue :(

[temnava]

Is there any progress regarding this issue or at least an idea what is causing the issue?
I'm getting this issue a lot and it is breaking the application. I will need to look for alternatives as soon as possible if this can't be fixed so if you can please tell me what is the situation.
Thank you