BiLD-Dev
Bi-Weekly "Loyal" DIRAC developers meeting. And, following, the LHCbDIRAC developers meeting.
Join Zoom Meeting
https://cern.zoom.us/j/91083694183?pwd=ZkdDY1I1YkJVc2o3UTdBY1BRZE15UT09
Meeting ID: 910 8369 4183
Passcode: 12345678
One tap mobile
+41432107108,,91083694183# Switzerland
+41315280988,,91083694183# Switzerland
Dial by your location
+41 43 210 71 08 Switzerland
+41 31 528 09 88 Switzerland
+41 43 210 70 42 Switzerland
+33 1 8699 5831 France
+33 1 7037 2246 France
+33 1 7037 9729 France
Meeting ID: 910 8369 4183
Find your local number: https://cern.zoom.us/u/artUfAdNB
Join by SIP
91083694183@188.184.89.188
91083694183@188.185.118.153
Join by H.323
188.184.89.188
188.185.118.153
Meeting ID: 910 8369 4183
Passcode: 12345678
BiLD (Bi-weekly DIRAC Development meeting) – 23/02/2023
On Zoom: Federico, Andrei, André, Alexey, Alexandre, Cedric, Christophe, Christopher, Daniela, Ewoud, Lorenzo, Simon, Janusz
Apologies: Ueda, Hideki
Follow-up from previous meetings
- Last BiLD 2 weeks ago
- all updates below
- Hackathon last week on 8.1.0a7 – https://github.com/DIRACGrid/DIRAC/releases/tag/v8.1.0a7
- contained several changes done in the past few months
- changes to DB/config listed in https://indico.cern.ch/event/1252424/
- participation also from members from CTA
- ran using an alpha version of DIRACOS
- several issues discovered and addressed straight away (some discussed below)
DIRAC communities roundtable
LHCb:
Federico+Alexandre+Christophe+Christopher+Alexey
- NTR
EGI
Andrei
- Contemplating how to migrate to 8.0, using a test server especially for the DB updates
- running v7.3/py3, few services on 8.0
ILC/CLIC/FCC/Calice:
André, Lorenzo
- NTR
Belle2
Cedric
- Trying v7r3.
- from previous meeting Production to v7r2, but need to move to v8.0 before May, for HTCondor support, potentially also for Rucio support
GridPP:
Daniela+Simon+Janusz
- Lots of issues with AREX and proxy renewals (classic small VO side effects: our jobs tend to queue a lot longer than LHCb’s) - Simon submitted a bunch of patches. Still on v7.3.34 + hotfixes (see issues wrt v7.3.35)
Topics from GitHub/Discussions or Google forum
only un-answered topics:
- Dynamic RAM requirements for jobs to HTContor CEs
- Andrei The answer is basically “no”
- The Matcher can use the RAM. Maybe would need an update of the PoolCE.
- Andrei In an environmentally-friendly computing we would need to take this into account
- DFC:: list PFNs on SE
- DFC method is implemented only for the
FileManagerPSmanager - Christophe if meant to be implemented for other managers it is easy, because at least the test is already there.
- DFC method is implemented only for the
- DIRAC.Interfaces.API.Job.Job.setOutputData::outputSE - replica level specification with inclusion/exclusion
- Should we turn it into issue?
- How can a user set a priority when submitting jobs through the API
- Maybe worth code, and/or doc? – PR created
- dirac.cfg :: set the value of a non-default name FileCatalog
- Andrei Not a simple fix to be done, I will answer. Semantic of the CLI…
DIRAC releases
- v7r3
- v7.3.35
- NEW: (#6643) Support OpenSSL 3.0.x
- Several fixes for AREX CE
- v7.3.36 should be done soon once existing PRs are merged in.
- v7.3.35
- v8r0
- Last one is v8.0.15
- CHANGE: (#6794) Speed up ReplicateAndRegister operation
- NEW: (#6785) Improve documentation on installing HTTPs services
- NEW: (#6708) SensitiveDataFilter in LogFilters
- Last one is v8.0.15
- v8r1
- Mostly fixes from the hackathon
DIRAC projects
- Federico I have “Archived” from GitHub a few obsoleted projects under https://github.com/DIRACGrid/
- One project was deleted (DIRAC4Android, last commit 10 years ago…)
- Others to be archived: VMDIRAC, BoincDIRAC, DIRACOS (?)
- RESTDirac got an update recently…?
DIRAC:
Issues by milestone:
- v7r3:
- 4 open issues
- v8.0:
- 22 open issues
- Issues with OpenSearch/Elastic based types and dashboards
- Being addressed in https://github.com/DIRACGrid/DIRAC/pull/6845
- Conflicts when updating JobParameters in ElasticJobParametersDB
- Possible solution proposed in https://github.com/DIRACGrid/DIRAC/pull/6802 but unsure + discussions
- Expand integration tests for IaM
- assigned to Alexandre
- Incorrect behavour of BundleDelivery with HTTPS
- Proposal accepted
- Issues with OpenSearch/Elastic based types and dashboards
- 22 open issues
- v8.1:
- 21 open issues
- Pilot.json : add a new format
- Being tested in Pilot PR [devel] Remote Pilot Logger to Tornado new JSON format but of course that would not fully solve this issue
- tested also in Jenkins (test expanded)
- Being tested in Pilot PR [devel] Remote Pilot Logger to Tornado new JSON format but of course that would not fully solve this issue
- (for PilotLogging) find the VO from the proxy (and from the token)
- Also being worked out in Pilot PR [devel] Remote Pilot Logger to Tornado new JSON format
- Pilot.json : add a new format
- 21 open issues
- NTR
PRs discussed:
- PRs created (for this repository and for others in DIRACGrid) with updates to several automated checkers.
- [8.0]Pilot submission with tokens
- Some last minute comments from André
- [8.0] Pilot stamps available in the pilot execution environment
- Merged, tested in hackathon …
WebApp:
- Few PRs open, connected to DIRAC PRs
Pilot:
- Janusz [devel] Remote Pilot Logger to Tornado new JSON format is ready for being reviewed. This is necessary before [integration] Remote Pilot Logger to Tornado could be run.
- Federico tested also in Jenkins
- Jansuz the DIRAC side of things is waiting for this to be merged
DIRACOS2:
- from previous meeting Chris Is anyone using v7r2 client with py3?
- Still, for a few more days (CLIC), a bit longer (May, Belle2)
- Chris we might end up breaking v7r2 support
- will need to update to OpenSSL 3 early next year as OpenSSL 1.1.1 will reach EOL soon: https://github.com/DIRACGrid/DIRACOS2/pull/95
- 9th Feb Seems ready to go
- from previous meeting MySQL 8.0.27 version forced, as it drops support for old TLS
- and VOs using MySQL 5.7.9 or earlier? That would break with the next DIRACOS release
- The PR was merged
- We could potentially use python 3.10, or even 3.11 already
- Python 3.11 seems to be fine
- Propose to update immeidately after OpenSSL 3
- Opened an annoucement about OpenSSL 3
- Seen a couple of sites which don’t work with OpenSSL 3’s more secure defaults
- Included a script to allow others to check if they have affected sites
- Please try it and report the results (and ticket sites if needed)
- Do we need
myproxyin DIRACOS?
Documentation:
- NTR
OAuth2:
- from previous meeting
- WLCG timeline document: https://zenodo.org/record/7014668#.YyLag9JBwQ9
- Meeting note from 8/12/2022: https://demo.hedgedoc.org/cWl2Y5MtTwSn4V5Uic_K5w
- TokenManager WebApp will be added (pretty much like ProxyManager page), plus testing
tornado/HTTPs
- from previous meeting Issue https://github.com/DIRACGrid/DIRAC/issues/6495 keeps track of what can and what will not be moved to https
- BundleDelivery: replaced by a completely different solution? Proposal accepted
- Federico we need to try and propose a coherent update from DIPS to HTTPS services, e.g. for Configuration Services URLs
- who already runs HTTPs services in production?
- apart from LHCb, only very few from EGI.
- who already runs HTTPs services in production?
management
- from previous meeting 3 issues left, still valid
- NTR
diraccfg
- Dropped support for py2 in https://github.com/DIRACGrid/diraccfg/pull/7
- Few PRs followed with related updates
- Added full type hints so
mypy --strictpasses - Chris B Plan to do a little more clean up with the packaging and then make a 1.0.0 release
- Would then be picked up in the following DIRACOS2 version
COMDIRAC
Daniela
- Started implementing the requests in the issue
- Federico Should we archive https://github.com/DIRACGrid/COMDIRAC now?
- yes
DB12
Alexandre
- NTR
Rucio
- NTR
Tests
- Moving to HTTPs services as defaults in PR
- After this, even the “DIPS” tests will use a few HTTPs services, at least the Master CS
- Within this PR we will also:
- Rename CS sections of https services from
Systems/../Service/FileCatalogtoSystems/../Service/TornadoFileCatalogfor added clarity - Update the instructions on how to deploy
- Rename CS sections of https services from
Release planning, tests and certification
Certification machines
- lbcertifdirac70 machine:
- Computing Element of each supported type should be configured (HTCondorCE, 2 ARCS (Alexandre (maybe use RALPP and Glasgow ? I can talk to them if there are issues --Daniela)), 1 Cloud (Daniela - this is already part of the tests), SSH). Consider OAuth tests as well - what is needed as far as certification servers configuration is concerned.
- Daniela: Enabled (and tested, at least with X509) WLCG VO for token testing. Needs update in puppet for /etc/grid-security/vomsdir and /etc/vomses, please.
- Christophe I’ll check and add it
- Daniela Still missing as of 14/12/22 A working config can be found at /cvmfs/grid.cern.ch/etc/grid-security/vomsdir/wlcg/ and /cvmfs/grid.cern.ch/etc/grid-security/vomses/wlcg-voms.cloud.cnaf.infn.it
- Chris https://gitlab.cern.ch/ai/it-puppet-module-voms/-/merge_requests/46
- lbcertifdiracoauth machine:
- NTR
Next hackathon(s)
- March 2nd on lbcertifdirac70
AOB
Next hackathon: March 2nd
Next BiLD: March 9th
LHCbDIRAC
- v11.0: deploy board in https://trello.com/b/Ep0PAkbv/deploy-110
- https://lhcb-auth.web.cern.ch/
- from previous meeting Andrei this is not properly configured yet. Not usable for Pilot submission yet (compute scopes not added).
- Alexandre An admin can allow a user (a client actually) to get access to the
compute.cancel/create/modifyscopes. Chris allowed Andrei and I to get access to the compute scopes, thus we should be able to use our clients to submit pilots.
- Bookkeeping …
- Singularity CE everywhere?
- https://lblogbook.cern.ch/Operations/36258 – set SingularityCE “almost” everywhere?
- yes, ready after ticketing few sites
- Need to start using multicore queues – one for each Site where we have storage. task to be added to Trello
- memory usage of a single core job of MCReco