We decided to replace this with a four-node (virtual) grid with S3 (NetApp) storage. We’re running CV11.24.25 with a two-node grid (physical) with CIFS mount paths from a Nexsan Unity that takes secondary copies from MAs that perform backups (no direct backups other than DDB), with a a partition on each MA. Slow Chat: Visual C++: Yesterday, Today, and Tomorrow.Slow Chat: Developing Multithreaded Applications.Slow Chat: Talk with Microsoft Developer Teams.IoT, IoE, and Maker Forum (on VBForums).Project Planning, Design, and Management.Scripting - Server Side (PHP, Perl, etc.).Windows Presentation Foundation (WPF) & XAML forum.NET Installation and Configuration Issues Quick Navigation C++ (Non Visual C++ Issues) Top The semaphore count indicates how many objects are queued.Ī mutex must also be used in each function of the queue (because semaphores does not solve problem #1).

The correct synchronisation object is a semaphore. It could be solved, making the consumer thread test for queue voidness even if the event is not signaled, and waiting for the event if the queue is empty, but that's ugly.Įvents are not meant to be used when there are more than one shared resource. This problem happens even if there is only a single consumer thread (if two resources are produced while it is working, it will consume one of them, and block after). This leak will remain until the end of the program.

Other consumer threads think that the queue is empty, while it is not! Only a single consumer thread will take data on the queue. The producer thread pushes a second thing on the queue (consumer threads are still working) and sets the event (which is already signaled). The producer thread push something and sets the event (assuming it is an automatic event), while all consumer threads, for now, are working on data. Events are either signaled or not (there is no count):.The producer thread works very fast, and push something else on the queue, and set the event.Ī second consumer thread is activated, but the first thread is still in the pop operation, and here is a thread-safety issue.Īdding a mutex for accesses on the queue solves this first problem. Producer thread push something on the queue and set the event.Ī first consumer thread is activated and start poping an element on the queue It does not protect the internal data structure of the queue (so a mutex is needed in addition):.

I don't think that events can really help here.

Queues vs threads crystal disk mark