BiLD – 27/03/2025

At CERN: Federico, Christophe, Christopher, Alexandre, Ryunosuke, Robin, Theau
On Zoom: Simon, Janusz, Hideki, Henryk, Alexey, Jorge, Vladimir, Xiaomei, Daniela, Luisa, Cedric
Apologies: André

---

Follow-up from previous meetings

• Last BiLD was March 6th
• Last DIRAC certification hackathon on March 20th: https://github.com/orgs/DIRACGrid/projects/24
  • the last for v9?
  • pretty OK, really minor issues found
• Daniela How’s everyone’s CHEP reviews coming along ?

---

DIRAC communities roundtable

LHCb:

Federico+Alexandre+Christophe+Christopher+Ryun+Vladimir+Alexey+Robin+Theau

• On Monday we will start the migration to DIRAC v9
  • will take long time because we are “profiting” for doing many database updates:
    • update MySQL to 8.4
    • update ROW_FORMAT to “Dynamic” (many tables were created when the default was “Compact”)
    • updated character set to utfmb4
    • defrag
    • the many DB changes for DIRAC v9 (this is the only “obligatory” change)
  • we’ll deploy also lhcbdiracx and lhcbdiracx-web

ILC/Calice/FCC

André

• NTR

Belle2

Hideki

• Smooth DIRAC operations, few questions about preparation for v9 migration

EGI

Andrei

• Problems with certificates for some communities, asking for moving to use tokens instead. So, looking for soon migration to DIRAC9.DiracX. Simple workflows, so DiracX simple job management can be enough for this community.
• A young engineer starting on the 1st May for the GreenDIGIT/DIRAC

GridPP:

Daniela, Simon, Janusz

• Nothing to report on production.

Juno

Xiaomei

• NTR

CTAO

Natthan, Luisa

• Fully move to IAM since VOMS server at CC-IN2P3 is now stopped
• Issue with the ProxyManager -> we run it with Tornado till it is not fixed (on the v9 instance)
• DiracX CA, which solution for trusted CA?
  • igtf will not distribute let’s encrypt, maybe the google one will be there by the end of the year

---

Topics from GitHub discussions and bots

• only un-answered DIRAC and DiracX topics with discussion updates:
  • Opensearch question
  • Safely pass secrets within jobs
    • Federico some private communication

---

DIRAC releases

• v8.0.71
  • FIX: (#8083) htcondor x509 unsupported version
  • CHANGE: (#8072) conditionally reset the rlimit for xroot
  • CHANGE: (#8070) Disable Bearer token for HTTPs unless upload/TPC
  • NEW: (#8046) findFileByMetadata method for Rucio
•  

---

DIRAC projects

DIRAC:

Issues by milestone:

• v8.0:
  • Using cgroups to limit job resource usage
    • Federico I was hoping Simon you would complete the job…
• v9.0:
  • Left 2 bug reports in there
• After v9.0:
  • NTR

Other issues:

• dirac-install-component can’t install Tornado service
• stompy errors in matcher
• DMS Error: Server error while serving listDirectory

PRs discussed:

• [8.0] fix: DISET and proxy location
  • hotfixed in LHCb
• [8.0] Add PreferredURLPattern for URL sorting
  • input from Belle2, which likes the idea of global setting
• [9.0] feat: added foreign keys to PilotAgentsDB
• [8.0] Clear any non-UTF encodable environment variables in pilots

WebApp:

• from previous meeting One draft PR

Pilot:

• Robin started work for using new pilot security model (DiracX) note
  • PR feat: Adding JWT support alongside X509 auth

DIRACOS:

• New release https://github.com/DIRACGrid/DIRACOS2/releases/tag/2.52
  • FIX: (#144) Remove setrlimit in XRootD

Documentation:

• NTR

OAuth2:

• NTR

management

• NTR

diraccfg

• NTR

DB12

• from previous meeting https://github.com/DIRACGrid/DIRAC/issues/7760#issuecomment-2482420604
  • Federico proposed to create “alternate” benchmark

Rucio

• NTR, apart from the fact that the Rucio metadata catalog PR was included in the last release.

Tests

• from previous meeting Federico Started adding Rucio to Dirac integration tests
  • –> to Janusz

---

DiracX:

• Road Map : https://github.com/chaen/diracx/blob/roadmap/docs/ROADMAP.MD
• dependabot alert
  • 4 alerts, but last 3 are for the same issue

Issues

• Lack or foreign key use in databases could lead to orphan data
  • Federico created DIRAC PR (would need a porting in DiracX if accepted)
• Open access and require auth not working inside a router

PRs discussed:

• feat: enable remote pilot logging system [MISSING AUTH]
  • Federico last comment is “Waiting for the pilot authentication to be ready”, what does it mean exactly? is it expected by who?
    • Robin can work on it --> Adding pilot registrations and authentification
• chore: remove __all__ from non- __init__ files
• fromPreviousMeeting OTel proof of concept PR: https://github.com/DIRACGrid/diracx/pull/379
  • This is a doc PR that is not yet merge-able. Christophe giving instructions to Jorge on how to proceed

DiracX-charts:

• NTR

DiracX-web:

• Lots of “bump” auto PRs

---

Release planning, tests and certification

• Certification machines

  • NTR

• Next hackathon(s)

  • not sure…

• Federico We will tag DIRAC v9 (and diracx, web, etc) in around a week.

Next appointments

• Meetings:

  • BiLD: in 2 weeks

• WS/hackathons/conferences:

  • DiracX hackathon: 5 and 6 May - https://indico.cern.ch/event/1501369/
    • few registered already
  • Dirac Users’ Workshop: 17th-20th September 2205 - https://indico.cern.ch/e/duw11
    • registrations open, Xiaomei added few info for VISA

AOB

• fromPreviousMeeting DIRAC was invited to be an “HSF affiliated project” : https://hepsoftwarefoundation.org/projects/affiliated.html
  • Andrei, André, Federico met with Edoardo and Michel Jouvin for few clarifications. Andrei will call a consortium meeting

---

LHCbDIRAC

• LHCbDiracX packages are now on pypi (“release 0.0.1a3”)

• Alexandre posted update to Moving Job finalization step from the workflow to the JobWrapper: Transition Plan for Enhancing HPC Exploitation in DIRAC/LHCbDIRAC with connected draft PR Draft: feat(wms): New LHCb workflows

• Hardware replacement campaign 2025: OTG – instructions in https://clouddocs.web.cern.ch/using_openstack/resizing_a_vm.html

  • “Resize implies a downtime depending on the amount of data on the disk and the load on the infrastructure during the resize. It may last from 15 minutes to 8 hours.”
  • "On Linux VMs, it’s strongly recommended to execute fstrim -va to discard any unused filesystem blocks, making the resize operation quicker. "
  • Affected machines:

  VM Name	Recommended Action	Flavor	Compatible Flavor
  diracxcertif	RESIZE	m2.large	m4.medium
  lbcertifdirac70	RESIZE	m2.large	m4.meidum
  lbdiraclogstash01	RESIZE	m2.2xlarge	m4.xlarge
  lbdiraclogstash02	RESIZE	m2.2xlarge	m4.xlarge
  lbvobox300	RESIZE	m2.2xlarge	m4.xlarge
  lbvobox401	RECREATE	m2.3xlarge	NONE
  

  • last 4 in DT in next week?

• Update plan: https://codimd.web.cern.ch/dnfwITCRRTSvhopGDjlHSA?both