Open
Description
FTP: 2021-11-24-Dasboard-issue/Chirp testing weird dashboard.pxp
The above experiment does not show any sweeps in the dasboard.
Digging into the labnotebook shows in rows (394-415)
SweepNum TimeStamp TimeStampUTC EntrySourceType AcquisitionState
394 4 3720516340.39 3720545140.39 5
395 4 3720516340.394 3720545140.394 5
396 4 3720516340.411 3720545140.411 5
397 4 3720516340.424 3720545140.424 5
398 5 3720516371.616 3720545171.616 2
399 5 3720516371.65 3720545171.65 2
400 5 3720516371.655 3720545171.655 2
401 5 3720516371.665 3720545171.665 2
402 5 3720516371.672 3720545171.672 2
403 5 3720516371.845 3720545171.845 8
404 5 3720516394.065 3720545194.065 2
405 5 3720516394.103 3720545194.103 2
406 5 3720516394.107 3720545194.107 2
407 5 3720516394.112 3720545194.112 2
408 5 3720516394.113 3720545194.113 2
409 5 3720516394.275 3720545194.275 8
410 4 3720516439.641 3720545239.641 1 0
411 4 3720516439.644 3720545239.644 1 0
412 5 3720516439.675 3720545239.675 2
413 5 3720516439.707 3720545239.707 2
414 5 3720516439.712 3720545239.712 2
415 5 3720516439.777 3720545239.777 2
And here is what happened:
- Acquired 4 sweeps with PSQ_DAScale and DAQ was finished normally (DAQ stop reason: 4)
- Tried to run PSQ_Chirp but configuration failed (DAQ stop reason: 2). In this step already some labnotebook entries for the future sweep 5 were added.
- And then we ran the TP. As we only acquired 4 sweeps, the TP entries are added with sweep 4.
- When the dashboard now searches for "DA Scale Set QC" labnotebook entries it uses "UNKNOWN_MODE" as the source type. This mode is used for user labnotebook entries when adding them. For searching it means that you don't care about the entry source type. And due to that the LBN search finds row 410-411 but those don't have a "DA Scale Set QC" entry and therefore the dasboard does not find anything.
Now the dashboard is fixed in #1184 but the underlying problem remains.
Possible solutions:
- Buffer labnotebook entries added during PRE_SET_EVENT/PRE_SWEEP_CONFIG_EVENT and only write them once we really start DAQ. We do something similiar with the labnotebook entries stored in SweepSettingsWave and written after DAQ. The disadvantage is that this makes post mortem debugging more difficult.
- One reason for the problem is that the TP entries use sweep 4. If the TP would have used sweep 5 the problem would not exist. This sounds tempting to do, but the implications are not yet clear.
- Another idea would be to disambiguate querying the labnotebook with UNKOWN_MODE. It currently means "I don't care" about the mode, but could also mean I want exactly that mode. The problem here is that we still have to support the existing labnotebook entries with UNKNOWN_MODE.
- Don't do anything as this is a rare issue.