Reached the PLC, or appeared in your application's send log?
Restart offers our FR5 housing check a second trip
LouisBrown0920 · 27 Oct 2025, 23:39 UTC
16 replies
Worth answering that precisely. On a request problem I asked about here, seeing the bit change still left the receiving side unaccounted for. What evidence have you got from the PLC itself?
11 pointsPLC trace shows receipt, Mia and Ravi. No saved completion. Housing is being held for review.
18 pointsGive the held housing a reference someone on the next shift can find. I've been arguing for this on another startup discussion: pending on a tag tells stores very little about why a part must stay put.
7 pointsAnd name who can release that hold. Not everyone reading the list.
13 pointsI wouldn't freeze every job under one pending bucket either. Work definitely not submitted is different from this housing. But your software has to distinguish them reliably, not let a person guess from the order on screen.
16 pointsDoes the entry identify the physical housing as well as the requested check? If someone removes it for inspection, the software reviewer needs to know which item that decision concerns. A pocket number can become misleading after a tray change.
20 pointsIsaac, a link to the incident would help. Not another separate handwritten explanation to keep matching.
11 pointsThe crash can occur after the PLC acts and before the application saves an answer. Saving a sent flag a little sooner does not cover every gap. The recovery design still needs a way to reconcile a durable attempt identity, or hold an uncertain attempt for review.
6 pointsRosa's physical identity question matters. We once had two sample pieces laid beside a tool and one job label between them. A very orderly bench, completely useless for deciding which one had been checked. The holder position was not a lasting identity.
10 pointsLouis, does a resend currently keep the original attempt identifier or invent a new one?
15 pointsNew one, Victor. Rosa, housing identity is retained. Automatic resend is disabled in the revised offline startup flow.
19 pointsThat stops one tempting button. It still needs to survive another restart without turning the uncertain entry into fresh work. Show the unresolved reason again, not only a disabled control.
4 pointsLin, agreed on separate states. But absence of a receipt is not proof nothing was sent.
10 pointsThanks Isaac. Held housing now links to the incident, and the software draft keeps the uncertainty after restart.
15 pointsRavi, yes, definitely not submitted needs evidence from the design, not an empty receipt field. Louis, include crashes at the save boundaries in the tests. A normal restart with a neatly saved record is the easy case.
14 pointsDiscussion closed
This discussion is closed to new replies after six months without activity. Last activity: .