Rucio Meeting

Thursday Jun 30, 2022, 3:00 PM → 4:00 PM Europe/Zurich

4/S-030 (CERN)

Martin Barisits (CERN)

3:00 PM → 3:05 PM News

• July & August meeting schedule
  • Bi-weekly meetings over the summer again!
  • July 7
  • July 21
  • August 4
  • August 18
  • September 1 (Back to weekly schedule)
• Rucio workshop 2022
  • Will be co-located with the WLCG Workshop
    • Lancaster, UK
    • Nov 7-11
      • Most likely first 3 days WLCG workshop, last 2 days Rucio workshop
  • Program committee still needs to be established, will start work soon
  • Official announcement eMail to come soon
• Rucio 1.29 "Into the Donkeyverse"
  • Code-freeze since this morning!
  • 1.29.0rc1 tomorrow or Monday
    • Depending on Merge progress

3:05 PM → 3:20 PM Community News & DevOps roundtable

• ATLAS
  • RPG development
    • RPG is a system in ATLAS managing some data flows
    • Idea is to get RPG functionality into Rucio Rules & Subscriptions
    • #4508
• CMS
• Fermilab/Ikarus/DUNE/Rubin
  • Getting DUNE policy package (lfn2pfn method) to work
  • FTS3 testing for Rubin
    • Tuning transfer parameters
  • Setting up cluster deployment for Rubin
• DUNE
  • Testing policy package for lfn2pfn method
    • Ahead of data challenge in July
    • Some issues, working on resolving these
• Belle II
  • Testing 1.28.5 release (Running 1.26 in production)
  • Migration to 1.29 still needs DIRAC python3 deployment
• ESCAPE
  • Issues with replication, some transfers take too long, needs investigation
    • Might be an issue with large conveyor-submitter bulk_size
• SKAO
  • Transition from user_token to admin_token
    • Identified quite a few problems 
      • Some Rucio related, some otherwise (FTS, ...) 
      • One bug in reaper, conveyor-poller needs to be updated too
    • No group claim ('wlcg.groups') existent in token
      • Some storages might require the claim -> doesn't work
• RAL/MultiVO
  • K8s deployment
    • Daemons deployed
    • Server & Auth server - x509 auth does not work yet
    • proxy renewal cron needs a bit of work to renew multiple proxies for multi-vo

3:20 PM → 3:25 PM Hot topics

3:25 PM → 3:55 PM Developers roundtable

• Rucio 1.29 "Into the donkeyverse" progress tracking
  • In Progress​​
    • Further ingetragtion of CMS-style consistency checking #5438 [Igor]
      • Waiting for PR to be merged - then this is done (until further steps are identified)
    • Remove old readthedocs documentation once the new one is fully operational #5446 [Joel, Martin]
      •  
  • In Review
    • Hermes 2 evolution #4610 [Cedric]
      • Comments to be included
    • Calculation of container length and events #5422 [Radu]
      • Difficult to calculate proper statistics for containers without performance degradation
      • Open PR, but unclear if it fully closes the issue
    • Further OIDC and SSO optimization and documentation #5437 [Rizart]
      • Ticket done, now looking on related ticket which will be submitted as one
    • Removal of Python 2.7 dependencies from the code #5435 [Martin, Joel]
    • Further integration of temporary tables to other major workflows #5441 [Radu]
  • Done
    • re-evaluate the greedy = 1TB constant #5358 [Cedric]
    • Clean dead code in the probes (ATLAS and common) #83 [Eric, Dimitrios]
    • boto vs boto3 #5346 [James, Mario]
    • Fix client API documentation page #5444 [Joel]
    • Ticket Cleanup Marathon [Martin, All]
    • mypy static code analysis #4455 [Mayank, Ben, Joel]
      • Discussion for type checking and documentation for developers guide coming
    • Analyze feedback on UI Survey #w/13 [Mayank]
    • Setup WebUI repository for external collaboration #w/12 [Mayank]
    • Flag transfer errors coming from tape sources #5410 [Radu]
    • Re-factor/cleanup deletion workflow #5442 [Radu]
    • Add session object to API layer #5439 [Radu]
    • Remove author list from headers [Martin, Joel]
  • Todo
    • Create a contributor guide for probes #84 [Dimitrios]
    • Adherence to rule grouping dependent on order of ATTACH #5251 [Igor]
    • Document and cleanup discrepancies between surl and lfn2pfn algorithms #5129 [Martin]
    • Unable to Delete File DID via Undertaker #5154 [Martin]
    • Exchange of function-based indices with normal indices #5440 [Radu, Mario, Martin]
    • Create documentation page for S3-type storage #5450 [Rizart]
      • Exists as GGDoc (#5284)
    • Add support for S3 virtual hosted buckets #5451 [Rizart]
      • Test on-premise
      • Ongoing support issue with AWS/Fresno X.509 shenanigans
    • Merge list_dids and list_dids_extended methods #5448 [Rob]
  • Delayed
    • Update/Re-design core.meta module #5224 [Rob]
    • Reduce rule tickets to les than 10 [Martin]
    • Reduce core component tickets to less than 10 [Mario, Martin, Gabriele]
    • Create a server/daemon installation howto #5445 [Joel]
    • Add symlinks for the new daemon names #5131 [Joel]
      • Prepare a clear list of changes which need to be made to use new daemon names #5132 [Joel]
    • Migration of SQLAlchemy queries to new query syntax (1.4 -> 2.0) #5247 [Yuyi, ALL]
      • ongoing (sessions.py)
      • not yet too scary :-)
      • Radu: also work on transfer part? how to enable warnings
        • SetEnv PYTHONWARNINGS always::DeprecationWarning
          SetEnv SQLALCHEMY_WARN_20 1
          https://docs.sqlalchemy.org/en/14/changelog/migration_20.html
      • Now trying to build a test to identify missing queries
    • Create developers testing guide in the documentation #5452 [Mayank, Joel]
      • Ben working on sub-task to test locally
    • Remove direct DB dependence from the WebUI #4319 [Mayank]
    • Reduce logging of server & daemons #5449 [Mayank]
      • Now access to ES
      • Started to populate dashboards about frequency distribution of log patterns
      • --> DELAYED for codefreeze
    • Authentication workflows in WebUI 2.0 #5135 [Mayank]
      • stuck on setting up base repo for webui
      • --> together with GSoC student, DELAYED for codefreeze
    • Add python 3.10 tests to CI framework. #5145 [Mayank]
      • Fedora images are ready for 3.10
      • Stopped to work on CC7 - not worth it
      • Currently looking into centos stream 8 and stream 9
      • --> On track
    • Remove full_mode switch from conveyor-receiver #5453 [Radu]
    • Re-work/cleanup distance table #5454 [Radu]
    • Introduce heartbeat handler to all daemons #5443 [Radu, ALL]
      • Started to move some daemons to the new daemon_handler
• Other topics
  •  

3:55 PM → 4:00 PM AOB