phwissmann
commented
2 months ago Scicat provides:
- a "reset" admin job that tries to copy the datablocks again to the LTS, potentially under a different path in order to not fail if a datablock is already present. Here, the split into two sub flows already provides strong preconditions such that an archive job could just be submitted again and the datablock creation could be skipped and the copying to LTS step will be executed again without any datablock being present in the LTS.
- a "delete" adming job that deletes partially archived (?) data. Not relevant due to stronger post conditions on the archiving jobs
Scicat defines some cleanup jobs in case of e.g. archival failure ("reset"). These need to be understand and an API to trigger these jobs needs to be implemented