Resources not freed when making a large number of request

[pgagne]

While evaluating rlocker for use by the LNST team, we discovered that sometimes there appear to be issues with a large number of requests.

I developed a script that allowed me to do a large number of simualtanious lock/release operations and found when I did 10,20,50, 70 unlock/release operations on a single resource, the resource was always released and locked as expected. However, when I tried to do 100 operations we found that sometimes one of the signoffs would not get releases.

I did some debugging in order to get to the bottom of this, however I wasn't able to get to anything conclusive.

I some scripts that I will share that does 2 things:

1. provide a context manager based wrapper around rlockertools around locking/releasing a resource, albit it is still a work in progress when it comes to error handing.
2. Does the test and collect data able the operations.

You should be able to use these as a base to debug this. At some pojnt I think the context manager based wrapper would be a useful addition to rlockertools once it is more fully fleshed out.