Resolve TMSS-648

Closes TMSS-648 Did a few commissioning observations with the OBS+PL+Ingest strategy template.

• Polished the scheduling of observation + successors
• Pass LOFARENV to ssh and/or docker commands so these run in the same correct environment.
• Feedback for observations is processed fine. Observations end up in finished state.
• QA MS -> hdf5 conversion runs and finishes
• QA plotting runs and finishes
• pipeline starts, runs and submits feedback. I've taken out the pipeline set status calls, and I'm using the feedback finishing/finished status updates only. -> no more race condition.
• processing of pipeline feedback sometimes breaks. To be investigated further.
• If pipeline reaches finished state, the ingest is scheduled
• ingest scheduling properly fails if it needs permission.
• granting ingest permission automatically tries to schedule it, and then succeeds.
• granting ingest permission before the pipeline completes automatically schedules the ingest subtask when pipeline is finished.
• For ingest we have no proper test system... So, auke and I deciced that the easiest solution is to make the ingesttmssadapter listen to the scu199 broker and use the TMSSClient to scu199, but to submit the ingest jobs then to the ingest production broker scu001. That works!
• Ingest transfers fail properly because the project has no resources in the LTA (and we did not set these yet.)

Suggestion: review these changes, merge to master, and continue on a new branch.