periodically cleanup /store/user/crabint1

[belforte]

It looks like we never (!?) cleanup the EOS directory where our pipeline jobs write output, i.e. /cms/eos/user/store/crabint1 on CERN EOS

There are files as old as May 2024

belforte@lxplus800/0000> pwd
/eos/cms/store/user/crabint1/GenericTTbar/autotest-1714770940/240503_211540/0000
belforte@lxplus800/0000> ls -l|head -2
total 50593
-rw-r--r--. 1 cmsprod zh 5417952 May  3  2024 output_1.root
belforte@lxplus800/0000> 

We need a periodic cleanup. Deleting files in the pipeline may be annoying since it will have to be done using X509 authentication and gfal-* commands, maybe setting up an acrontab script in crabint1 account is better. But we need to be sure we do not forget to check that it keeps working ! (a panel of used quota for this and Rucio output in Grafana Overview ?)

[nausikt]

Roger that, will give this fair priority.

[belforte]

do not forget, but also note that we are only 50% full after 2 years ... Maybe a rm * by hand every 3 months is enough !