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) – 05/05/2022
On Zoom: Federico, Alexandre, Christophe, Christopher, Ruben, Simon M, Daniela, Janusz, Simon, Hideki, Xiaomei, Alexey, Michel, Igor
Apologies: Andrei
Follow-up from previous meetings
- hackathon of April 28th, on 8.0.0a20:
- https://trello.com/b/kSemN0nw/800a20
- issues mostly on WMS and Monitoring. Most of them fixed with PRs already
DIRAC communities roundtable
GridPP:
Daniela+Simon+Janusz
- On pre-prod server tested newest version of DIRACOS2, looks OK
LHCb:
Federico+Christophe+Christopher+Alexandre+Ruben+SimonM+Alexey
- NTR
EGI:
Andrei
- n/a
Belle2
Hideki
- v7r3 in the development environment. Succeded in creating v7r2 client in py3.
- Production still on v7r1
- Q: how to fix the DIRACOS2 version?
- not yet possible, ChrisB will work on it (on to-do list): https://github.com/DIRACGrid/DIRAC/issues/5963
- Q: how to discover which version is running
bash-4.2$ cat $DIRACOS/.diracos_version DIRACOS2 2.21
BES3+Juno
Xiaomei
- very similar situation of Belle2: updated client to v7r2, v7.3 in testbed – but prod still on v7r1.
- Some issues in TS performances, github discussion open
- the settings added in https://github.com/DIRACGrid/DIRAC/discussions/6044 are anyway system specific. Maybe something could be added to https://dirac.readthedocs.io/en/latest/AdministratorGuide/ServerInstallations/InstallingDiracServer.html
NICA
Igor
- NTR
Topics from GitHub/Discussions or Google forum
- from previous meeting Proper way to make a cache persitent in DIRAC
- Andrei maybe try (or get inspired) by what is done in COMDIRAC.
- news?
- How to improve performance of TransformationManager
- BES3 and CTA trying out the given settings
DIRAC releases
- v7r2:
- No new release created, but we should, as few fixes were merged
- v7.3:
- No new release created, but we should – several fixes went in:
- NEW: (#6038) add CloudComputingElement
- NEW: (#6040) bitterly implement multihop
- NEW: (#6015) BringOnline setting for FTS transfers can be SE dependant
- No new release created, but we should – several fixes went in:
- 8.0.0a20:
- CHANGE: (#5944) Divided ComponentMonitoring into AgentMonitoring and ServiceMonitoring
- CHANGE: (#6002) Replaced the flags for monitoring with a new unified flag system
DIRAC projects
DIRAC:
Issues by milestone:
- v7r2:
- PR 5920 was merged but not swept – Chirstopher I will have a look
- v7r3:
- ComputingElement: Path not found warning
- closed
- Added Proxies on worker nodes are only 1024 bits strong
- still valid, needs further changes in PR
- ComputingElement: Path not found warning
- v8.0:
- DIRAC’s Epoch time has time zone dependence and Monitoring generating insane timestamp Bug needs to be fixed in order to have 8.0 out
- v8.1:
- NTR
Other issues:
- https://github.com/DIRACGrid/DIRAC/issues/6042
- mostly affecting
gfal2 - in which contxt should we initialize the way the way we want the external logs?
- mostly affecting
PRs discussed:
- [[rel-v7r2]: WIP : First version of AREX computing elemen(]https://github.com/DIRACGrid/DIRAC/pull/5542)
- In the long term the REST interface is what the ARC developers want us to use
arcis in DIRACOS2 and it’s quite large, would be good to dicth it (in the long term)
- [WIP][v7r3] feat (Resources): define occupancy plugin for pure xroot storages
- single testbed, but looks like working
- [WIP][v7r3] fix (wms): additional fixes for the PushJobAgent
- solution works, but it is not generic
- [v7r3] Fixerrormess
- asked to do differently, ongoing.
- [8.0] Upgrade Stomp version and implement simpler interface
- with the existing system, apart from being (too much) complicated, there are bugs, e.g. you can’t listen to multiple queues
- for v8.0 or v8.1 prob removing the old version – not too quickly anyway
- [integration] Downloading job sandboxes from the web
- approve now, other optimizations in other branches.
- [integration] Hackaton fixes v8.0.0a20 of Monitoring issues
- fixes out of the issues found in the hackathon, and concept of “master” backend
- other issue with Monitoring not loading on the webApp not solved yet
- [Integration] refactor: replacing voimport from Utils by ObjectLoader class
- Correct thing to do, but won’t end up being a small refactor
- [Integration] refactor: removing unused parameter form getSubLogger and improving log consistency
__name__is the name of the module and we don’t want a wrong log. Useself.class.__name__instead.
WebApp:
- https://github.com/DIRACGrid/WebAppDIRAC/pull/646
- need a different solution
Pilot:
- from previous meeting Janusz ditched the Extended Logger, not made a PR yet. Quite some code can be made simpler, both on client and server. No dictionary building on the client side.
- back to the wrapper to change the configuration paramater
- https://github.com/DIRACGrid/DIRAC/pull/5961
- testing is a bit of pain
- https://github.com/DIRACGrid/Pilot/pull/153
- Questions on how to use the
loggingmodule – between Janusz and Alexandre
- Questions on how to use the
DIRACOS2:
- Added celery.
- Fixed versions, mostly HTCondor.
Documentation:
- from previous meeting diracgrid.org still forwards to http://mardirac.in2p3.fr
OAuth2:
- from previous meeting Any news from issue WLCG and Token transition – reminders and requirements ?
- NTR
tornado/HTTPs
- from previous meeting Andrii We can eventually have special classes in Configuration and BundleDelivery to pass binary data - to be discussed.
- NTR
management
- NTR
diraccfg
- NTR
COMDIRAC
- NTR
DB12
- NTR
Rucio
- Mostly “finished”. Multi-VO support for extrating scopes in the making (not on our side).
Developers’ report
Simon Mazenoux
Optimizers in WMS use Executors.
Task Queue: producer - consumer pattern, “scalable”.
Task Queue implementation is implemented in DIRAC as Executor framework. Used for transitioning jobs from Received to Checking to Waiting (but also Staging)
Celery: distributed task queue, written in python. Highly adapted, highly scalable. Can be used to replace the Executor framework
Can remove the executor dependency. Needs RabbitMQ.
Positive feedback, seems doable. Already a PR in the making. It introduces RabbitMQ as dependency. Scalability should not be an issue.
Release planning, tests and certification
-
8.0 release
- Currently missing:
- a few critical issues, but for OAuthZ we don’t want to improve ATM
BLOB->TEXTeverywhere (only 1 DB missing)
- Target: 1 more hackathon.
- Currently missing:
-
Certification machines
- lbcertifdirac70 machine:
- NTR
- lbcertifdiracoauth machine:
- NTR
- lbcertifdirac70 machine:
-
Next hackathon(s)
- Still on lbcertifdirac70
AOB
Next hackathon on May 12th
Next BiLD on May 19th
from previous meeting DIRAC Workshop: next week!
LHCbDIRAC
- Alexey agreed to take care of LHCbDIRAC Bookkeeping.
- developments and maintenance is expected. Program of work will be defined in the next weeks.
- v10r3: deploy board in https://trello.com/b/kzUKdMts/deploy-v10r3
- upgrade of
dfc: Chistophe in contact with DBOD support. Wants to wait and see what happened tolbwmsfirst - HTTPS services: some work is still needed for most of them.
- upgrade of
- LHCbDIRAC hackathon based on 8.0
- pushed to third week of May