The Selenium Grid Hub distributes tests running between registered RCs, but if there are no RCs available, they will be queued.
Does anyone know the maximum size of this queue?
I ask this because I suspect that if you conduct a large number of tests and you have remote RC-remote nodes, the hub will not be able to fully manage the requests, and the RC session will not be correctly released when the testing ends, so when you run the next test, it will fail, because it will use this RC, which will not be released correctly.
In my case, I had a problem with this configuration:
-Debian server with Selenium Grid 1.0.8, starting the hub.
-A Windows virtual machine with Selenium Grid 1.0.8 and IE, working with RC and registering it against a Debian hub.
If I run tests from one Debian machine one after another, the tests pass, but if I run many tests at the same time as the first test pass, the following fail.
Any idea what is going on?
Thanks in advance
- Victor
source
share