Automatic cleaning of input data Rucio rules#12427
Automatic cleaning of input data Rucio rules#12427nhduongvn wants to merge 13 commits intodmwm:masterfrom
Conversation
|
Jenkins results:
|
57a8335 to
78443bc
Compare
|
Jenkins results:
|
|
These are failed unit tests which not related to the new codes: |
|
please test |
|
Jenkins results:
|
|
Jenkins results:
|
|
Jenkins results:
|
|
retest |
|
rebuild |
|
All 4 errors are pre-existing issues below unrelated to changes in this PR: |
|
test this please |
Fixes #11409
Status
not-tested
Description
To automatically deleting the Rucio rules of input data set (block) of workflows, we implement a cherrypy thread that periodically scans the global workqueue for completed workqueue elements (
percentSuccess == 100 and percentComplete == 100) and deletes the corresponding Rucio rules of input data blockThere are also similar implementations in MSRuleCleaner (a new Pline for cleaning). However, Alan pointed out that MSRuleCleaner only runs on workflows at "later" stages ("announced"?). Therefore, the cleaning will not be done as soon as the processing of input data is complete. The code is still kept here (maybe for proof-of-concept for a new MS?).
Is it backward compatible (if not, which system it affects?)
YES
External dependencies / deployment changes
Yes, it requires the settings/configurations for a cherryPy thread that point to the Rucio account of input data set rules and the global queue