Testing v9.5.2

[DominicOram]

• confluent-kafka not installing on the control machine. Fixed by just installing this from the workstation (so it's the wrong binaries) then doing the rest of the install from the control machine - will be fixed by Containerize hyperion #250
• Was no dodal release so just checked out main? - no fix required
• GDA still doing xrc when robot load enabled
• Panda config hadn't set seq1 prescale and prescale units, so pulses were being sent too quickly and signals were being dropped - fixed in Hotfixes from 9.5.2 hyperion#1515
• Had to wire the Panda back up from the eyeball experiment - we plugged TTLOUT1 from the panda into TTLIN1 to the Zebra 1. We unplugged the connection from zebra 1's TTLIN1 to zebra 2 for this

Created issue DiamondLightSource/hyperion#1517 to cover all the panda-related issues

[DominicOram]

• PCAP was trying to write to panda/panda - fixed by changing in PandaHDFWriter: - covered by Panda gridscan issues discovered during testing of 9.5.2 hyperion#1517

        # Set the initial values
        await asyncio.gather(
            self.panda_device.data.hdf_directory.set(
                str(info.root)
            ),

• Need to give .h5 a unique name - covered by Panda gridscan issues discovered during testing of 9.5.2 hyperion#1517
• Issue with PCAP writing in that we had to stage the panda after we run setup - Fixed in Hotfixes from 9.5.2 hyperion#1515
• Need to get the aperture positions for moving before snapshot from the aperture scatterguard itself - need to probably have an enum with the positions in that we use that as an external interface - Fixed in Hotfixes from 9.5.2 hyperion#1515, Clean up aperture scatterguard interface dodal#740 to follow up
• Issue with gda not knowing the sample on restart and so sending an empty sample id to hyperion
  [31m[E 240808 14:33:41.263 __main__:254] ['Traceback (most recent call last):\n', ' File "/dls_sw/i03/software/bluesky/hyperion_v9.5.2/hyperion/src/hyperion/__main__.py", line 228, in compose_start_args\n parameters = experiment_internal_param_type(**json.loads(request.data))\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n', ' File "pydantic/main.py", line 341, in pydantic.main.BaseModel.__init__\n', 'pydantic.error_wrappers.ValidationError: 1 validation error for PinTipCentreThenXrayCentre\nsample_id\n none is not an allowed value (type=type_error.none.not_allowed)\n', '\nThe above exception was the direct cause of the following exception:\n\n', 'Traceback (most recent call last):\n', ' File "/dls_sw/i03/software/bluesky/hyperion_v9.5.2/hyperion/src/hyperion/__main__.py", line 246, in put\n plan, params, plan_name, callback_type = compose_start_args(\n ^^^^^^^^^^^^^^^^^^^\n', ' File "/dls_sw/i03/software/bluesky/hyperion_v9.5.2/hyperion/src/hyperion/__main__.py", line 230, in compose_start_args\n raise ValueError(\n', "ValueError: Supplied parameters don't match the plan for this endpoint\n"]�[0m
• Snapshots work with the "gda.mx.hyperion.rotation.snapshots.enabled" flag in GDA but we still have some repeated stuff that hyperion and GDA are doing, needs further investigation - See Remove correct things from GDA when rotation snapshots done in hyperion #208
• Robot load needed the sample id passed correctly
• Hyperion waiting ~20s around https://graylog.diamond.ac.uk/search?rangetype=absolute&q=gl2_source_input%3A%2266264d79952b2458ee4b4bc4%22&highlightMessage=d0729838-558f-11ef-8486-ee44a4a7a9ef&streams=66264f5519ccca6d1c9e4e03&from=2024-08-08T14%3A09%3A13.763Z&to=2024-08-08T14%3A09%3A43.000Z

[DominicOram]

• https://ispyb.diamond.ac.uk/samples/sid/5682018 is the full hyperion doing everything
• Pulse size on PCAP needs to be the same as the exposure time - fixed in Panda gridscan issues discovered during testing of 9.5.2 hyperion#1517
• Box size isn't just changed by changing BOX_WIDTH_UM because it isn't passed through to grid_detection_plan. Trying to add 10 in that call resulted in https://ispyb.diamond.ac.uk/dc/visit/cm37235-3/dcg/12681792 , which failed to kickoff the FGS - covered in XRC: Ensure easy way to change box sizes and exposure time #231

[DominicOram]

• Need to wait for smargon enabled before moving the lower gonio back into place - see Do not move lower gonio until the robot is safely away hyperion#1526
• The extra waiting seems to be happening at some point between after _set_triggered_tip has been awaited and coming back to the plan. e.g. inside the RE? Possible we have some other asyncio task running here? - discussed with @dperl-dls and @olliesilvester, we think the issue is that trigger(x, wait=True) will actually wait for the None group, which we probably have something left over in. To do:
  • Write up an issue on the bluesky repo to fix this with a uuid in the same way that the set does - at least have the docstring tell you about it. - Provide clearer behaviour on waiting without groups bluesky/bluesky#1795
  • We found that the leftover status was from thawing so we should make sure this doesn't end up in the None group. - Make sure we don't wait on thawing accidentally hyperion#1527

[DominicOram]

I believe everything in this issue is now covered by linked issues