Closed StevenCTimm closed 1 week ago
We've figured out that this big blurp of files heading to tape simultaneously came this morning when I put a bunch of pre-beam files into the declad dropbox, resulting in DUNE_CERN_EOS rules being made for some pretty big pre-beam datasets that didn't have them before. at one point the peak was 80K now it is closer to 60K so we are catching up.
https://cern.service-now.com/service-portal?id=ticket&is_new_order=true&table=incident&sys_id=796c1f10978b8a5081ef33f71153af3a (INC 3943610) has been filed with CERN service desk to ask for investigating the timeouts.
About half the backlog was cleared overnight, helped by the fact that we didn't take much data overnight. The timeouts have stopped. We still need to get a "bringonline" timeout into the FTS jobs we are sending to CTA.
All backlog is now cleared.
From fts3-public.cern.ch