Crash in app under load

Hello Zoom,

We developed an app based on Windows SDK that launches Zoom on several PCs in a studio used for teaching. Last week, during meetings with approx 350 users, the app crashed. The app has never crashed in production before, although these were much larger meetings than we usually host. The app was relaunched and crashed again a few minutes later, relaunched again and it stayed up through meeting. The app crashed on both PCs at exactly the same times, after processing the same events from Zoom (we handle and log various events such as leave/join/activespeaker). The machines were not configured to create mini-dumps so I don’t know where the crash occurred.

Are there known vulnerabilities to large meetings in the Windows SDK? We are currently using SDK release from 2019-11-04 and I see there are two releases since then.

Do you have any suggestions for testing under load? I’ve thought about automating the web client to simulate large numbers of participants.

Thanks for any advice,

Bill

Hi Bill_Gardner,

Thanks for the post. Our SDK will create dump if a crash happens, you may find the dump files at %appData/zoomsdk/logs/. Please send the dump or any log files to us to further investigate the issue you are facing.

Our SDK has the same capacities as the Zoom client, so it should support large meetings. If you would like to have a large meeting and would like to have a better performance, please consider using a Webinar instead.

Hope this helps. Thanks!

Thanks Carson, that’s helpful. I did not know about the local crash dumps. On PC1, the log directory was empty. On PC2 there were a number of crash dumps, none from the date of the crashes I mentioned above, but one from 3/19 which apparently coincides with a crash that happened with 120 users. PC1 runs SDK in classic mode, PC2 runs SDK in custom mode to display custom galleries. Below is link to dump, not sure if you can tell anything. Meanwhile I am configuring mini-dumps on all machines to track this down. Thanks - Bill

Hi Bill,

Thanks for the reply and the dump file. I will forward this to the engineering team for investigation. Thanks!