Details
-
Type:
Technical task
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: None
-
Labels:None
-
Sprint:AP S22-6 (May)
-
Team:Alert Production
Description
During DM-34828, we found 13 visit+detector datasets failed to process through imageDifferenceTask. They all had errors akin to
Cannot compute CoaddPsf at point (1258.13, 2258.72); no input images at that point. {0} |
lsst::pex::exceptions::InvalidParameterError: 'Cannot compute CoaddPsf at point (1258.13, 2258.72); no input images at that point.' |
The list of datasets is attached to this ticket. Figure out why they don't have full coverage template coadd overlap when we explicitly wrote a selector script to ensure that they did.
Attachments
Issue Links
- is triggered by
-
DM-34828
Process large-scale processing run and persist timing metrics
- Done
Activity
The problem was that I errantly grabbed templates from patch 15 instead of 17 when creating the TAGGED coadd collection. I fixed it by using butler.registry.dissasociate for the patch 15 goodSeeingCoadd dataRefs and associating the patch 17 goodSeeingCoadd dataRefs, and now the problem is fixed.
The template collection is still u/mrawls/DM-34827/coadd/4patch_4431.
When you don't include data, it doesn't get processed!
It appears either patch 16 (or possibly 9? geometry is hard) failed to process at all. See attached plot.