Eiger testing notes

[jacob720]

Notes from eiger testing with @niamhdougan @DominicOram @EmsArnold

• Ensure stream mode is enabled. Sometimes it gets set to disabled, can't tell why
• Set number of triggers and images in TriggerInfo, don't set the PV manually in daq.
• Default timeout is 10s. If the collection takes longer than this you'll get a timeout error. Need to set a timeout to match the expected time of collection or increase the default massively. (Although 10s good for testing).

Acceptance Criteria

• Specific criteria that will be used to judge if the issue is fixed

[jacob720]

At 12:12 and at 12:39 there was an issue where odin was stuck in writing and hitting "Stop Writing" did nothing, had to restart the IOC

• Should be able to restart IOC without deleting app in argocd

BlueAPI seems to push RMQ messages to /queue/public.worker.event.nexus. We will need to push them to a topic for workflows to pick them up.

Logs of nexus writer in k8s are horrible to read, e.g.:

{"timestamp":"2026-05-13T11:48:43.781","level":"INFO","thread":"pool-1-thread-4","logger":"uk.ac.diamond.daq.domain.Scan","message":"Adding all cached documents to file","context":"default","beamline":"i15-1","application_name":"Document to Nexus Converter"}
{"timestamp":"2026-05-13T11:48:43.781","level":"INFO","thread":"pool-1-thread-4","logger":"uk.ac.diamond.daq.messaging.export.RabbitMQExporter","message":"Sending file output notification: {\"status\":\"UPDATED\",\"filePath\":\"/tmp/40814198-cc73-496a-988d-007c0c54f27c.nxs\",\"visitDirectory\":\"/tmp\",\"swmrStatus\":\"ENABLED\",\"scanNumber\":405535474,\"scanDimensions\":[-1],\"scannables\":[],\"detectors\":[],\"percentageComplete\":-1.0,\"processingRequest\":{}} on topic gda.messages.scan","context":"default","beamline":"i15-1","application_name":"Document to Nexus Converter"}
{"timestamp":"2026-05-13T11:48:43.781","level":"INFO","thread":"pool-1-thread-4","logger":"uk.ac.diamond.daq.messaging.export.RabbitMQExporter","message":"Sending file output notification: {\"status\":\"UPDATED\",\"filePath\":\"/tmp/40814198-cc73-496a-988d-007c0c54f27c.nxs\",\"visitDirectory\":\"/tmp\",\"swmrStatus\":\"ENABLED\",\"scanNumber\":405535474,\"scanDimensions\":[-1],\"scannables\":[],\"detectors\":[],\"percentageComplete\":-1.0,\"processingRequest\":{}} on topic gda.messages.scan","context":"default","beamline":"i15-1","application_name":"Document to Nexus Converter"}
{"timestamp":"2026-05-13T11:48:43.781","level":"INFO","thread":"pool-1-thread-4","logger":"uk.ac.diamond.daq.domain.Scan","message":"Received StopDocument with ID 83237eab-e336-4fc3-8c7b-634f6efb8359 for scan 40814198-cc73-496a-988d-007c0c54f27c","context":"default","beamline":"i15-1","application_name":"Document to Nexus Converter"}

[niamhdougan]

For stream mode we can do a quick test to see if initialising the detector resets this.

[jacob720]

• Eiger IOC logs seem to stop after a while. @shihab-dls aware of this and looking into it
• Ophyd async is not triggering the correct start_writing PV. It is triggering BL15J-EA-EIGER-01:OD:FP:StartWriting but I think it should be BL15J-EA-EIGER-01:StartWriting. We currently can't do this as we want ophyd-async to maintain parity with the JF so instead we can put create_interleave_vds in Odin. Action on @jacob720 to do this. Move VDS generation into fastcs-odin fastcs-eiger#97
• Virtual datasets not being created, even when the PV is set to True and the correct start_writing PV is triggered. Two issues - div by 0 error caused by the point below. Also not receiving correct path. Action on @jacob720 to look at.
• Should not assume or set block size in ophyd async device
• VDS doesn't work if both frames_per_block and frame_count are 0. In the long term frame_count needs to be able to be 0, for an infinite number of frames (@DominicOram to write up an issue) but in the mean time, we can change ophyd-async to specifically set odin frame number to be ntriggers*nimages. Handle unknown number of frames when generating VDS fastcs-odin#111
• frames_per_block needs to default to 1, not 0 (should never be 0). Action on @shihab-dls to fix default.

At 12:12 and at 12:39 there was an issue where odin was stuck in writing and hitting "Stop Writing" did nothing, had to restart the IOC

@shihab-dls thinks that this is due to the scan loop in the Eiger dying. Once logs are fixed we can find out why

• We were not seeing the eiger IOC logs in graylog at all
• To write the VDS we need the IOC to be able to write to /dls/ and run as ixxdetector - this wasn't working, we got errors when setting the IOC to run as i15-1detector and mounting /dls/i15-1/data. @shihab-dls to investigate

[OCopping]

• Should be able to restart IOC without deleting app in argocd

This is due to be fixed in the next ec release. My fix for this has been merged in to main already.
You will be able to do ec restart <eiger-service> like all other IOCs

UPDATE: Turns out ec v5.2.0 was released yesterday, so this should now be fixed.
https://github.com/epics-containers/edge-containers-cli/releases/tag/5.2.0

[jacob720]

VDS issues captured here bluesky/ophyd-async#1274