Description of problem:

uss.t testcase sometimes fails due to timeout whenever run with brick-mux enabled.

One of the results of the failure in the regression run can be found here at [1]

Found that, the reason for this is because in the test that leads to the timeout we do the following things.

Create a file (aaa in the testcase)
Take a snapshot (snap6)
Access it via uss
Delete the file (aaa)
Delete the snapshot (snap6)
Again create snapshot (snap6 i.e. same name as the one deleted above)
Again access the file 'aaa' expecting it to fail (as the file was not present
                                                  at the time of snapshot)

But we access snap6 in the last test too soon. The previous instance of snap6 might still be going through its cleanup phase causing hangs etc.

So, after deletion of the snapshot (snap6) wait to ensure that glusterd recognizes it to be not there, before proceeding with creation of the snapshot with the same name.

[1] https://build.gluster.org/job/regression-on-demand-multiplex/613/

Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: