Repository navigation
Conversation
|
Thanks Igor. I see the intent of this. However, I am not a big fan of configuration in code, and in fact there is the danger that the approach is not foolproof (as concluded from a Claude Fable review): the correction does not only depend on the production tag, but also on the TF length and the anchor time of each run. Since MC-prod-2026-v12, the anchoring shifts the first orbit by the ITS ramp-up duration from
The last row exists because the O2DPG fix rounds the shift with the data's OrbitsPerTF (32) before In addition, there is a unit bug from #15897: We could consider attacking these corrections on the CCDB side, as we do usually. For instance, the MCProdInfo objects in |
It turned out that the MC productions done with the MC-prod-2026-v13 tag are still affected by the TF shift issue described in O2-7220.
This is because the v13 tag used O2PDPSuite::daily-20260929-0000-1, which predates the fix introduced on October 4 to address the issue with the ITS ramp-up offset.
The affected productions were added based on the list extracted from:
https://alimonitor.cern.ch/job_details.jsp?packages=MC-prod-2026-v13