A coupon acknowledgement between two application runs

BrunoAdams0104 · 26 Aug 2025, 08:34 UTC

Closed
BR
BrunoAdams0104
I am reconstructing an FR5 coupon inspection failure. The Python application restarted and reused its job label; the PLC trace and controller events have different clocks. One acknowledgement falls between the apparent attempts after I align the files. No independent attempt identifier in the exported columns. I do not know whether that alignment is valid.

15 replies

AN
AnnaBaker0503
Replying to BrunoAdams0104

Keep the original files and their own event order first. What event did you use to align them? A similar-looking entry may not refer to the same thing on both sides.

12 points
BR
BrunoAdams0104
Replying to AnnaBaker0503

I aligned a Python connected entry with the nearest controller connection event. There are two possible controller events close together, so that choice is not secure.

18 points
AN
AnnaBaker0503
Replying to BrunoAdams0104

Then show both possibilities instead of quietly choosing one. The restart can still split the application file even if you cannot place it exactly against the controller clock.

12 points
LU
LuisBrown0956
Replying to BrunoAdams0104

Can you get the original PLC capture rather than only the exported columns? There may be a retained counter or request reference outside the selection. I would ask controls before trying more elaborate clock corrections.

17 points
JA
JamieChan1078
Replying to BrunoAdams0104

And ask whether the export changed time zones; our spreadsheet once made a perfectly ordinary event look like a trip into the future.

25 points
BR
BrunoAdams0104
Replying to JamieChan1078

Jamie, the export is local time and the application is UTC. That explains a fixed offset, but leaves the two connection events. Luis, controls is retrieving the full capture.

8 points
JA
JamieChan1078
Replying to BrunoAdams0104

Good, one less distraction, though the nearest connection still isn't necessarily yours

9 points
JO
JonasBrooks0807
Replying to BrunoAdams0104

Did two clients connect, or one client reconnect twice?

6 points
AN
AnnaBaker0503
Replying to JonasBrooks0807

Bruno, can IT identify the endpoints for those two connection entries? That would be more useful than deciding which timestamp looks friendlier.

12 points
BR
BrunoAdams0104
Replying to AnnaBaker0503

IT identified one connection as our monitor and the other as a separate service laptop. The full PLC capture also has a request counter omitted from the export. We are comparing that counter with the receiver's retained results.

6 points
JO
JonasBrooks0807
Replying to BrunoAdams0104

Does that counter survive the restart in question?

9 points
LU
LuisBrown0956
Replying to BrunoAdams0104

Check where the acknowledgement carries the counter too. Having a number on the request side will not help if the only recorded acknowledgement is an unlabelled state that could already have been present.

5 points
BR
BrunoAdams0104
Replying to JonasBrooks0807

Jonas, controls confirms it persisted across this application restart. Luis, the receiver's retained result includes that counter. It matches the earlier attempt; the newer attempt has no matched result in the records recovered so far.

14 points
BR
BrunoAdams0104
Replying to LuisBrown0956

I've corrected the report: earlier attempt has a matched completion, newer outcome remains unknown. Thanks for pushing for the full capture, Luis. We were close to arguing about the clocks indefinitely.

12 points
AN
AnnaBaker0503
Replying to BrunoAdams0104

Will that request counter be included in the standard export now? It would save the next investigation from starting with the same missing column.

25 points

Discussion closed

This discussion is closed to new replies after six months without activity. Last activity: .