Add a timeout to tape server global lock on the object store

As explained on the operations issue https://gitlab.cern.ch/cta/operations/-/issues/972, the reason for the tape drives becoming STALLED is that they are waiting for the object store global lock.

This causes them to become unresponsive.

In order to avoid this unresponsiveness, we need to include a timeout (30 seconds, parameterised?) in the acquisition of the global lock.