The following changes support the removal of Beanstalk from Teuthology.
In place of Beanstalk, we will now be using Paddles for queue management in Teuthology.
This PR has the corresponding changes for the paddles PR: https://github.com/ceph/paddles/pull/94/files.

The changes include:
1. Removing all beanstalk related code
2. Teuthology scheduler and dispatcher using Paddles queue for scheduling and dispatching jobs
3. Adding support for Paddles queue management
4. Additional functionality of being able to change the priority of Teuthology jobs in the queued state in the teuthology-queue command

Signed-off-by: Aishwarya Mathuria <[email protected]>

Previously the teuthology worker unit test used beanstalk, the test will
be using Paddles now.

Signed-off-by: Aishwarya Mathuria <[email protected]>

1. Add retry loop for the paddles calls.
2. Add run name as a parameter for updating priority of jobs in paddles.
3. Modify the pause queue command to run on server side with an optional pause duration parameter.

Signed-off-by: Aishwarya Mathuria <[email protected]>

Signed-off-by: Aishwarya Mathuria <[email protected]>

…ith Paddles

With the use of the --queue-backend argument the user can specify which backend(paddles/beanstalk) they would like to use for maintaining the teuthology Jobs queue.
In order to avoid overlapping Job IDs, when a job is being scheduled in beanstalk it is also written to paddles which returns a unique ID.
This is the ID teuthology will treat as the Job ID throughout the run of the job.

To differentiate between the 2 queue backends, the teuthology-queue command has been split into teuthology-paddles-queue command and teuthology-beanstalk-queue command.

Signed-off-by: Aishwarya Mathuria <[email protected]>

…alk queue

Signed-off-by: Aishwarya Mathuria <[email protected]>

Makes the same teuthology-queue commands work regardless of the queue backend, Paddles or Beanstalk.

Signed-off-by: Aishwarya Mathuria <[email protected]>